Table of Contents

UNITED STATES

SECURITIES AND EXCHANGE COMMISSION

Washington, D.C. 20549

Form 10-K

(Mark One)

For the fiscal year ended March 31, 2005

OR

For the transition period from                      to                     

Commission file number: 000-23193

APPLIED MICRO CIRCUITS CORPORATION

(Exact name of registrant as specified in its charter)

Registrant’s telephone number, including area code: (858) 450-9333

Securities registered pursuant to Section 12(b) of the Act: None

Securities registered pursuant to Section 12(g) of the Act:

Common Stock, $0.01 par value

Indicate by check mark whether the registrant (1) has filed all reports required to be filed by Section 13 or 15(d) of the Securities Exchange Act of 1934 during the preceding 12 months (or for such shorter period that the registrant was required to file such reports), and (2) has been subject to such filing requirements for the past 90 days.    YES  x    NO  ¨

Indicate by check mark if disclosure of delinquent filers pursuant to Item 405 of Regulation S-K is not contained herein, and will not be contained, to the best of registrant’s knowledge, in definitive proxy or information statements incorporated by reference in Part III of this Form 10-K or any amendment to this Form 10-K.    x

Indicate by check mark whether the Registrant is an accelerated filer (as defined in Rule 12b-2 of Exchange Act).    Yes   x    No  ¨

The aggregate market value of the voting common stock held by non-affiliates of the registrant, based upon the closing sale price of the Registrant’s common stock on September 30, 2004 as reported on the Nasdaq National Market, was approximately $945,081,000. Shares of Common Stock held by each officer and director and by each person who owns 10% or more of the outstanding Common Stock have been excluded in that such persons may be deemed to be affiliates. This determination of affiliate status is not necessarily a conclusive determination for other purposes.

There were 306,520,524 shares of the registrant’s Common Stock issued and outstanding as of May 12, 2005.

DOCUMENTS INCORPORATED BY REFERENCE

Part III incorporates information by reference from the Registrant’s definitive proxy statement to be filed with the Securities and Exchange Commission in connection with the solicitation of proxies for the Registrant’s 2005 Annual Meeting of Stockholders to be held on August 23, 2005.

Table of Contents

APPLIED MICRO CIRCUITS CORPORATION

ANNUAL REPORT ON FORM 10K

FOR THE FISCAL YEAR ENDED

MARCH 31, 2005

TABLE OF CONTENTS

Table of Contents

CAUTIONARY STATEMENT ABOUT FORWARD-LOOKING STATEMENTS

All statements included or incorporated by reference in this report, other than statements or characterizations of historical fact, are forward-looking statements. These forward-looking statements are made as of the date of this report. Any statement that refers to an expectation, projection or other characterization of future events or circumstances, including the underlying assumptions, is a forward-looking statement. We use certain words and their derivatives such as “anticipate”, “believe”, “plan”, “expect”, “estimate”, “predict”, “intend”, “may”, “will”, “should”, “could”, “future”, “potential”, and similar expressions in many of the forward-looking statements. The forward-looking statements are based on our current expectations, estimates and projections about our industry, management’s beliefs, and other assumptions made by us. These statements and the expectations, estimates, projections, beliefs and other assumptions on which they are based are subject to many risks and uncertainties and are inherently subject to change. We describe many of the risks and uncertainties that we face in the “Risk Factors” section in Item 7 and elsewhere in this report. We update our descriptions of the risks and uncertainties facing us in our periodic reports filed with the U.S. Securities and Exchange Commission, known as the SEC, in which we report our financial condition and results for the quarter and fiscal year to date. Our actual results and actual events could differ materially from those anticipated in any forward-looking statement. Readers should not place undue reliance on any forward-looking statement.

PART I

Item 1.    Business.

In this annual report on Form 10-K, “Applied Micro Circuits Corporation”, “AMCC”, the “Company”, “we”, “us” and “our” refer to Applied Micro Circuits Corporation and all of our consolidated subsidiaries.

Applied Micro Circuits Corporation was incorporated and commenced operations in California in 1979. AMCC was reincorporated in Delaware in 1987. Our principal executive offices are located at 6290 Sequence Drive, San Diego, California 92121, and our phone number is 858-450-9333. Our website is located at www.amcc.com. The information that can be accessed on or through our website is not intended to be part of this report. Various documents concerning us that are electronically filed with or furnished to the SEC, including our annual reports on Form 10-K, quarterly reports on Form 10-Q, and current reports on Form 8-K are available, free of charge, on our website. Our common stock trades on the Nasdaq National Market under the symbol “AMCC”.

Overview

We design, develop, market and support high-performance integrated circuit, or IC, products, embedded processors, and storage components for the communications and storage equipment markets. Our products are essential for the transport, processing, switching, routing and storage of information worldwide. In the communications market, we utilize a combination of design expertise coupled with system-level knowledge and multiple technologies to offer IC products, as well as printed circuit board assemblies or PCBAs, for wireline and wireless communications equipment such as wireless base stations, edge switches, routers, and gateways, metro transport platforms and core switches and routers. We generate revenues in the communications market primarily through sales of our IC products, embedded processors and PCBAs to communications equipment manufacturers, such as Alcatel, Ciena, Cisco, Fujitsu, Hitachi, Huawei, JDS Uniphase, Juniper, Lucent, Marconi, NEC, Nortel, Siemens, and Tellabs, who in turn supply their equipment principally to communications service providers. In the storage market, we blend systems and software expertise with high-performance, high-bandwidth silicon integration to deliver high-performance, high capacity Serial ATA, or SATA, storage solutions for emerging storage applications such as disk-to-disk backup, near-line storage, network-attached storage, or NAS, video, and high-performance computing; and various PowerPC control plane processing solutions ideally suited for Redundant Array of Integrated Disks, or RAID, controllers and storage area networking, or SAN, equipment. We generate revenues in the storage market primarily through sales of our SATA storage solutions through our distribution channel partners who in turn sell to enterprises, small and mid-size businesses, value added resellers, or VARs, systems integrators and retail consumers and through sales of our Fibre Channel hardware and software products and embedded processors, to original equipment manufacturers, or OEMs, such as Brocade, Hewlett Packard, Hitachi Data Systems, Network Appliance, StorageTek and Sun Microsystems.

Table of Contents

Industry Background

The Communications Industry

Communications technology has evolved considerably over the last several years due to the substantial growth in the Internet and wireless communications. The emergence of new applications, such as wireless web devices, voice over Internet Protocol, or VOIP, video-on-demand, third generation, or 3G, wireless services, as well as the increase in demand for higher speed, higher bandwidth and remote network access, have increased network bandwidth requirements. The continuing adoption of broadband technology, such as email, instant messaging and e-commerce, and the increasing availability of next-generation wireless devices that incorporate features such as internet browsing, cameras and video recorders is expected to drive additional data traffic through the network infrastructure in the future. The different types of data transmitted at various speeds over the Internet require service providers and enterprises to invest in multi-service equipment that can efficiently manage and transport the varied types of network traffic, regardless of whether it is voice traffic or data traffic. To achieve the performance and functionality required by such systems, OEMs must utilize more complex ICs to address both the cost and functionality of a system. OEMs must address this need for multi-service equipment despite reductions in their development teams. As a result of the pace of new product introductions, the proliferation of standards to be accommodated and the costs and difficulty of designing and producing the required ICs, equipment suppliers have increasingly outsourced these ICs to semiconductor firms with specialized expertise. These trends have created a significant opportunity for IC suppliers that can design cost-effective solutions for the transmission of data. IC suppliers must utilize a variety of skills and technologies to satisfy the requirements of communications OEMs. These OEMs require IC suppliers that possess system-level expertise and can quickly bring to market high-performance, highly reliable, power-efficient ICs. These OEMs seek suppliers with a wide skills base including both analog and digital expertise to provide a more complete solution that enables faster integration into the system design and higher performance.

The increase in volume and complexity of network traffic has led to the development of new technologies for more efficient networks. These technologies provide substantially greater transmission capacity, are less error prone and are easier to maintain than copper networks. These more efficient networks carry high-speed traffic in the form of electrical and optical signals that are transmitted and received by complex networking equipment. To ensure that this equipment and the various networks can communicate with each other, OEMs and makers of semiconductors have developed numerous communications standards and protocols for the industry. For example, the Synchronous Optical Network, or SONET, standard in North America and Japan and the Synchronous Data Hierarchy, or SDH, standard in the rest of the world became the standards for the transmission of signals over optical fiber. The SONET/SDH standards facilitate high data integrity and improved network reliability, while reducing maintenance and other operation costs by standardizing interoperability among equipment from different vendors. With data and video traffic being added in abundance to voice traffic, Asynchronous Transfer Mode, or ATM, emerged as a transmission protocol complementary to SONET/SDH to optimize bandwidth utilization. Many service providers deploy equipment that handles this protocol because it can support voice, data, video and multimedia applications simultaneously with the ability to provide quality-of-service guarantees. With exponential increases in data traffic and very modest increases in voice traffic, data has become the dominant traffic over all networks today. Internet Protocol, or IP, is another transport protocol that maintains network information and routes packets across networks. IP packets are larger and can hold more data than ATM cells, but in some applications may not be able to provide the same quality of service because they are not optimized for time-sensitive signals such as voice and video. Because of the bandwidth growth and cost pressures in today’s datacentric networks, more advanced optical networking technologies, such as Dense Wave Division Multiplexing, or DWDM, have been adopted. DWDM is the optical multiplexing of different wavelengths of light down a single fiber. Each wavelength is the equivalent of an independent optical channel. DWDM greatly increases the capacity of installed fiber. Complementing DWDM transmission capabilities are technologies such as optical Add-Drop Multiplexers, or ADMs, and cross-connects which can more efficiently switch large optical datapaths through the network. Other protocols, such as multi-protocol label switching, or MPLS, have emerged that are better suited for data traffic while providing for the low latency and quality of service needs of voice and video traffic. The SONET/SDH standards have also evolved to more efficiently

2

Table of Contents

handle these new protocols with general framing protocol and virtual concatenation. In addition, emerging technologies such as multi-service provisionary platforms, or MSPPs, and multi-service switches, or MSSs, allow service providers to efficiently manage and transport the varied types of network traffic.

The Storage Industry

The buildout of high-performance computing infrastructure, adoption of higher-bandwidth Local Area Network, or LAN, and Wide Area Network, or WAN, connections, new appliances offering richer content, and increased focus on data redundancy have driven continued strong demand for high-capacity storage within the enterprise and within a wide variety of vertical markets. The volume of data generated, processed, stored and manipulated has grown dramatically over the last decade and managing that data is one of the most difficult challenges facing IT organizations today. The enormous growth in storage capacities is also impacting the type of storage being implemented. IT managers across industries are reassessing their storage needs and implementing topologies and technologies that are appropriate to the criticality of their stored information.

Traditionally, organizations have accessed and stored data using a server-centric architecture known as direct attached storage, or DAS. In this architecture, a single server controls access to each storage device, and stored data is only available to applications running on the server directly connected to the storage device. DAS systems are the lowest cost and easiest to implement and are currently the most widely used storage topology. However, they have limitations in terms of scalability and manageability that are addressed by networked or shared storage solutions such as NAS, and SAN solutions. A NAS device is an array of disk drives connected to the network making centralized stored data available to all connected servers. A SAN is a dedicated storage network of interconnected servers and storage devices that enables data sharing at gigabit speeds. SANs provide a very efficient solution for large enterprises in that they separate processing from storage by placing disk arrays on a separate, high-speed storage-only network. SAN adoption was made possible by the emergence of the Fibre Channel interconnect protocol—a computer communications protocol designed to meet the specific requirements of high performance information transfer.

All high-performance storage systems implement RAID (Redundant Array of Inexpensive Disks) technology, which manages the access and storage of information to the disk drives in the server or storage device. RAID is a technology in which data is stored in a distributed manner across multiple disk drives to improve system performance and to enhance fault tolerance and the ability to survive a hard drive failure. RAID dramatically improves disk access times and provides real time data recovery, with uninterrupted access, when a hard drive fails for increased system uptime and continuous network availability.

Storage spending represents a significant percentage of IT budgets as storage usage is expected to double each year through 2007. New regulations such as the Sarbanes-Oxley Act of 2002, the Health Insurance Portability and Accountability Act of 1996, or HIPAA, and those governing the finance industry are consuming huge amounts of disk capacity driving IT managers to consider less expensive storage technologies to stretch their budgets. New drive interface technologies such as SATA, emerged to provide a much lower cost alternative to traditional enterprise drives. SATA disk drives are rapidly being deployed in secondary storage applications such as back up, archival and near-line storage or the storage of infrequently accessed data.

AMCC Strategy

Our objective is to be the leading supplier of ICs and PCBAs for the transport, switching, routing and storage of information worldwide. Our strategy for achieving this objective includes:

Focus on the Market Leading Systems within the Communications Markets and Increase our Silicon Content in those Systems

We target key OEM product families that hold significant or rapidly growing market share. We have built substantial competencies focused on the specific requirements of these key OEM product families in the areas of semiconductor process technology, mixed-signal and very dense digital design, and substantial expertise in

3

Table of Contents

systems architecture, software and applications support. We believe that the integration of these capabilities enables us to optimize our market opportunities within these product families.

Increase the Number of Products we Provide to Address Specific Protocols and Networking Functions

We focus our new product development efforts on product lines that are complementary to our current product portfolio in order to broaden the number of products we provide to address specific protocols and networking functions. For example, our current product offerings include physical layer products, overhead processor products, and higher layer products for communications applications, and RAID controllers for enterprise storage applications. Both communications and enterprise storage applications may use general purpose processors, like our 400 series PowerPC processors. We believe that we will be able to increase our sales to existing customers and increase our market share in the communications and storage enterprise markets by taking advantage of product synergies and integration opportunities. We strive to create a balanced portfolio of products consisting of our very sophisticated and market-leading switch fabric and network processor solutions together with an array of shorter time-to-market connectivity products, such as (SONET physical layer, PCI and high-speed interconnect devices). We leverage our connectivity intellectual property across all product lines for both stand alone products and capability embedded in our network processor and switch fabric offerings.

Provide a Time-to-Market and Development Cost Advantage to Our Communications Equipment OEMs by Driving the Adoption of Standard Products

Application specific integrated circuits, or ASICs, are custom products that are designed for only a single customer or OEM, and can be sold only to that OEM. Application specific standard products, or ASSPs, are standards based products that are designed for, and can be used by, multiple OEMs. Our customers are looking for ways to accelerate their time-to-market, reduce research and development cost, and ensure interoperability of components in their systems. ASSPs generally can be designed into customers’ systems and brought to market in less time and for less cost. We believe that as more companies realize the development cost and time-to-market benefits that ASSPs provide, they will be more apt to use ASSPs in the future. Most of our products are ASSPs, and we believe that the trend towards greater usage of ASSPs in communications and storage area network systems will continue.

Due to the extended downturn in the communications industry, our OEM customers have become more efficient with their engineering resources and have significantly cut equipment development budgets. Our strategy is to provide our customers with a complete portfolio of IC products. We believe this comprehensive solution strategy provides our customers with guaranteed interoperability, pre-designed subsystems, better cost economics, and system-level expertise. The result for the OEM is faster time-to-market, better performance and lower development cost. To continue these customer benefits in future generations of products, we are pursuing an aggressive product integration strategy to provide greater functionality in fewer ICs.

Focus our Storage Product Roadmap on the High-growth SATA RAID Market

We believe the SATA RAID market offers substantial growth opportunities. By offering innovative performance enhancements and enterprise-class features, our strategy is to provide our customers with highly reliable, cost effective storage solutions that address this growing market. SATA incorporates significant technical enhancements over traditional ATA making it ideal for RAID implementations. Along with several configuration benefits, SATA technology improves data transmission through a point-to-point topology, which eliminates bus sharing, thereby delivering a full 1.5Gb/s bandwidth to each drive. In order to optimize performance in this point-to-point topology, a switch is required. Our products are at the forefront of SATA technology with the only switched fabric architecture—StorSwitch^™. StorSwitch allows the controller to simultaneously communicate with all drives, scaling linearly as more drives are added, resulting in very high storage subsystem performance.

4

Table of Contents

Leverage our Embedded Processor Technology to Increase our Existing Communications and Storage Opportunities and Target other High-growth, High-volume Markets

Our embedded processor products are well suited for our existing storage and communications markets and provide increased opportunity for our silicon content within our current customer base and sales capabilities. The embedded processor technology is also well suited for high-growth, high-volume markets such as printing and imaging, gaming and industrial computing. We currently offer ASSPs targeting these markets. We will focus on new product development efforts on opportunities within these high-growth markets and leverage design re-use and economies of scale between target markets where appropriate.

Products and Customers

Communications Products

Our semiconductor products are used in a wide variety of communications equipment, including routers, optical and digital cross connects, next-generation voice and media gateways, ADMs, MSPPs, MSSs, digital subscriber line access multiplexers, or DSLAMs and wireless base stations and access points. We provide our customers with a complete portfolio of IC products, including physical layer products such as transceivers, overhead processing products such as framers and mappers, and higher layer products such as network processors, traffic managers and switch fabrics. We have different types of communications IC products which are categorized by the order in which they receive and transmit signals and information within communication equipment. These categories are:

Physical Layer:    Our physical layer ICs, or PHYs, transmit and receive signals in a very high-speed serial format and reduce overall system “noise” through the inclusion of highly efficient dispersion compensation methodologies. This low noise capability permits the transmission of signals over greater distances with fewer errors. Our physical layer ICs also convert high-speed serial formats to low-speed parallel formats for the framing layer and vice versa. We introduced our first generation of physical layer products in 1993. We have since developed several generations of these products improving cost, power, functionality, and performance. Examples of our physical layer ICs include clock and data recovery, or CDR, devices, Multiplexers and DeMultiplexers. During fiscal 2005, we introduced the S19233, a dual CDR device with Electronic Dispersion Compensation, or EDC, and the S4850, a dual CDR. Our current customers for physical layer products include Alcatel, Ciena, Cisco, Fujitsu, Hitachi, JDS Uniphase, Juniper, Lucent, Marconi, Nortel, Tellabs, Huawei, and ZTE.

Framing Layer:    Our framing layer ICs transmit and receive signals to and from the physical layer in a parallel format and are used in high-speed transmission equipment, MSPPs, ADMs, digital and optical cross-connects, edge and core routers, and DWDM. These ICs support a number of functions, including framing, overhead processing payload synchronization, performance monitoring, forward error correction, and mapping the data payload to/from the transmission format. The framing layer ICs then pass the data either directly to a switch fabric product, which switches the information to its destination, or to a network processor, which further processes the data prior to forwarding it to a switch fabric product. Framing layer ICs also process signals received from the network processing and switching layers for transmission to the physical layer on their return to the optical network. In fiscal 2005, we introduced the Amur/S1215, a framer and multi-protocol data processing product line that implements all the framing, overhead processing, timing control and standards-compliant SONET/SDH channelization and mapping functions. Our current customers for framing layer products include Ciena, Cisco, Lucent, Marconi, NEC, Nortel, Tellabs, Fujitsu, Huawei, and ZTE.

Network Processing and Traffic Manager Layer:    Our network processor ICs are software programmable processors that receive and transmit signals from and to the framing layer and perform the processing of packet and cell headers, including such functions as real-time parsing, matching and table look-up, as well as bit stream manipulations, such as adding, deleting, substituting, appending and pre-pending. They can perform intelligent packet classification for policy-based network services. Our traffic managers interface with the network processors and perform the queuing and buffering functions required on packets and cells. Traffic managers

5

Table of Contents

usually interface with network processors on one side and switch fabric devices on the other. During fiscal 2005, we introduced the nP3705 family of integrated network processors and traffic managers. The architecture of this product family integrates multiple network-optimized programmable co-processors architected to deliver wire-speed performance while processing complex nested protocol stacks and mixed ATM/Packet payloads. Our current customers for network processors and traffic manager devices include Alcatel, Cisco, Fujitsu, Nortel, Lucent, Huawei and Juniper.

Switching Layer:    Our switch fabric ICs switch information in the proper priority and to the proper destinations. Our switch fabric product portfolio includes our PRS, or packet routing switch, fabric devices such as the PRS 5G, PRS 20G, PRS 28.4G, PRS 64G, 64Gu, Q-64G, PRS 80G / C48X and C192X and queuing managers like UDASL, C48 and C192. Our switch fabric portfolio also includes our Cyclone Terabit Switch Fabric series, a highly integrated, low-power, area-efficient chip set implementing a switching fabric that seamlessly handles many combinations of packet, cell, and TDM traffic. In fiscal 2005, we introduced a new PRS Q-80G switch device delivering higher overspeed for demanding Ethernet and Packet over SONET applications. Our current customers for switching layer products include Alcatel, CNT/Inrange, Fujitsu, Huawei, Lucent, Maranti, Marconi, Mitsubishi, Motorola, Nortel, Siemens and Tellabs.

Control Plane:    Sitting outside the datapath, our embedded processors or control plane processors control overall system functioning. These processors are high performance devices enabling high-speed computations that help identify, optimize and control the flow of data within the network. Our PowerPC product line of embedded processors enable complex applications such as deep content switching, routing and load balancing to be performed at wire speed.

Storage Products

Our current storage products include serial and parallel ATA RAID controllers and embedded storage processors. In fiscal 2005, we announced that we would discontinue development of our Fibre Channel Host Bus Adapter product line but will continue to sell and support our current line of HBA products.

RAID Controllers:    Through our acquisition of 3ware, we design, manufacture and sell a line of serial and parallel ATA RAID controllers. A SATA RAID controller installs in the PCI slot on a standard motherboard and delivers high performance, highly reliable storage for servers and network attached storage devices. A single controller manages up to 12 hard disk drives or six terabytes of data for Linux and Windows operating environments. Our proprietary packet-switched RAID architecture, StorSwitch^TM delivers best of class performance that we believe has secured AMCC a leadership position in the SATA RAID marketplace. In August 2002, 3ware introduced the 8000 series, the industry’s first SATA RAID controller which led the market in performance and capacity. In April 2004, AMCC introduced the 9000 series which increased performance and provided enterprise class features to expand the market for SATA.

Embedded Processors:    The PowerPC 440SP and 440 SPe Processors, members of the PPC440 embedded processor family, offer exceptional performance, high bandwidth, design flexibility, and robust features geared to demanding embedded storage and networking applications. The PowerPC 440SP and 440 SPe processors are ideally suited for RAID controllers and SAN equipment.

We are not currently developing new host bus adapters for the storage market, but we continue to sell Fibre Channel HBA products:

Fibre Channel HBAs:    An HBA is a PCBA that fits standard sockets on motherboards for servers and workstations that enable high-speed data transfer within the SAN. Communication between the HBA and the operating system is regulated by device driver software that is included with the HBA. The device driver software also provides a high-reliability data path from a user’s application to a storage device across a SAN. Working in conjunction with our device driver software, our HBAs work in many SAN topologies, interoperate with major operating systems and can be used with the PCI, PCI-X, cPCI, PCI express and SBus interfaces. Our HBAs are deployed in demanding SAN environments by Global 1000 enterprises.

6

Table of Contents

HCA Modules:    A host channel adapter, or HCA, is an interface that resides within a server and communicates directly with the server’s memory and processor as well as with the InfiniBand network fabric. When placed on a PCBA similar to a HBA, it is referred to as an HCA module under the naming guidelines set forth by the InfiniBand Architecture specification. An HCA module guarantees delivery of data, performs advance memory access and can recover from transmission errors.

We also sell fibre channel IC’s separately to OEMs for products such as SAN controllers for disk arrays, tape libraries, switches and other SAN devices.

Embedded Processor Products

Through our acquisition of the Embedded Processor Business from IBM in May 2004, we have enhanced our standard products portfolio for the communications and storage markets, and have also diversified into other markets which make use of embedded processors. Our embedded processor products are comprised of approximately 150 ICs which utilize IBM’s PowerPC 4xx processor cores in various speed grades, together with many different functions, to perform numerous tasks in products sold by our customers. Our customers in the communications market utilize these products in applications including wireless base stations, access points, networking hubs, edge routers and switches. Our customers in the storage market utilize these products in controllers, switches, adapters, servers, RAID systems and work stations. Our customers in the pervasive computing market utilize these products in printers, internet access and gaming devices.

Automated Test Equipment, Military and High-Speed Computing Products

We are not currently developing new products for the Automated Test Equipment, or ATE, or military markets, but we continue to sell ASIC products to customers such as Northrop Grumman, Raytheon, and Teradyne. The majority of these products were manufactured in our internal wafer manufacturing facility, which closed in fiscal 2003. During fiscal 2005, we continued to fill last-time-buy orders for these products. Our high-speed computing products were not manufactured in our internal wafer manufacturing facility, and we will continue to sell these products for the foreseeable future. The revenue from such products is expected to be modest.

Technology

We utilize our technological and design competencies to solve the problems of high-speed analog, digital and mixed-signal circuit designs for optical communications systems and provide the essential products for the transport, switching, routing and storage of information worldwide. We blend systems and software expertise with high-performance, high-bandwidth silicon integration to deliver communications ICs and software for global communication networks and hardware and software solutions for high-growth storage markets such as SATA RAID. Our embedded processor product line delivers performance and a rich mix of features for Internet, communication, data storage, consumer and imaging applications.

Knowledge of Communications and Storage Systems

Our systems architects, design engineers and technical marketing and applications engineers have a thorough understanding of the fiber optic communications and enterprise storage systems for which we design and build ASSPs. Using this systems expertise, we develop semiconductor and storage connectivity devices to meet the OEMs’ high-bandwidth requirements. By understanding the systems into which our products are designed, we believe that we are better able to anticipate and develop solutions optimized for the various cost, power and performance trade-offs faced by our customers. We believe that our systems knowledge also enables us to develop more comprehensive, interoperable solutions. This allows us to develop products that fulfill customers’ system needs from fiber-through-switch fabrics, enabling faster integration into their products.

7

Table of Contents

Design of Communications ICs and Storage Solutions

We have developed multiple generations of products that integrate both analog and digital elements on the same IC, while balancing the difficult trade-offs of speed, power and timing inherent in very dense high-speed applications. We were one of the first companies to embed analog phase locked loops in bipolar chips with digital logic for high-speed data transmission and receiver applications. Since the introduction of our first on-chip clock recovery and clock synthesis products in 1993, we have refined these products and have successfully integrated multiple analog functions and multiple channels on the same IC. The mixing of digital and analog signals poses difficult challenges for IC designers, particularly at high frequencies. We have gained significant expertise in mixed-signal IC designs through the development of multiple product generations. We will continue to apply these competencies in the development of more complex products in the future.

We have developed storage connectivity products that interoperate with all SAN topologies and major operating systems and interfaces. We intend to continue working closely with leaders in the storage, networking and computing industries to design and develop new and enhanced storage connectivity products. We believe that establishing strategic relationships with technology partners is essential to ensure that we continue to design and develop competitive products that integrate well with solutions from other leading participants in the storage markets.

Research and Development

Our research and development expertise and efforts are focused on the development of high-performance analog, digital and mixed-signal IC’s for the communications and storage markets, and PCBA’s and software solutions for storage markets such as SATA RAID. We also develop high-performance libraries and design methodologies that are optimized for these applications. Our primary research and development facilities are located in San Diego and Sunnyvale, California, Raleigh, North Carolina and Andover, Massachusetts in the United States; Essonnes and Sophia, France; and Bangalore, India. During the fiscal years ended March 31, 2005, 2004, and 2003, we expended $118.6 million, $112.6 million, and $131.9 million, respectively, on research and development activities.

Our IC product development is focused on building high-performance, high-gate-count digital and analog-intensive designs that are incorporated into well-documented blocks that can be reused for multiple products. We have made, and will continue to make, significant investments in advanced design tools to leverage our engineering staff. Our product development is driven by the imperatives of reducing design cycle time, increasing first-time design correctness, adhering to disciplined, well documented design processes, and continuing to be responsive to customer needs. We are also developing high-performance final assembly packages for our products in collaboration with our packaging suppliers and our customers.

Our PCBA product development efforts are focused on building high-performance SATA RAID adaptors, and related software drivers, tools and products and AdvancedTCA, or ATCA, boards and software for our switch fabric, network processors and embedded PowerPC processors for the communications market. Before a new product is developed, our research and development engineers work with marketing managers and customers to develop a comprehensive requirements specification. After the product is designed and commercially released, our engineers continue to work with customers on early design-in efforts to understand requirements for future generations and upgrades.

Manufacturing

Manufacturing of Integrated Circuits

The manufacturing of ICs requires a combination of competencies in advanced silicon technologies, package design and manufacturing, and high speed test and characterization. We have obtained access to advanced CMOS and SiGe processes through foundry relationships. We have substantial experience in the development and use of plastic and ceramic packages for high-performance applications. The selection of the

8

Table of Contents

optimal package solution is a vital element of the delivery of high-performance products and involves balancing cost, size, thermal management, and technical performance. We purchase our ceramic packages from several vendors including IBM, Kyocera America, Motorola and NTK Ceramics and our plastic packaging from Amkor, ASE, ASAT and IBM.

Wafer Fabrication

During fiscal 2003, we closed our internal wafer fabrication facility in San Diego. As a result, we are a fabless company, meaning we do not own or operate foundries for the production of silicon wafers from which our products are made. We will continue to use external foundries such as IBM, Taiwan Semiconductor Manufacturing Corporation, or TSMC, and United Microelectronics Corporation, or UMC, for a majority of our production of silicon wafers. Subcontracting our manufacturing requirements eliminates the high fixed cost of owning and operating a semiconductor wafer fabrication facility and enables us to focus our resources on design and test applications where we believe we have greater core competencies and competitive advantages.

Assembly and Testing

Our wafer probe and other product testing is conducted at our internal testing facility as well as at independent test subcontractors. After testing is complete, the majority of our products are sent to multiple subcontractors located in Asia and the United States for assembly. Following assembly, some of the devices are tested at the subcontractors and returned to us ready for shipment to our customers; or to us for final testing and marking prior to shipment to customers. Certain of these services are available from a limited number of sources and lead times are occasionally extended.

Manufacturing of Printed Circuit Board Assemblies

We believe most component parts used in our RAID adaptors and ATCA evaluation boards are standard off-the-shelf items that can be purchased from two or more sources, other than our proprietary ASICs and certain ICs. We select suppliers on the basis of functionality, manufacturing capacity, quality and cost. Whenever possible and practicable, we strive to have at least two manufacturing locations for each product. Our contract manufacturers generally purchase the components for our products, and assemble them to our specifications.

Sales and Marketing

Our sales and marketing strategy is to develop strong, engineering-intensive relationships with the design teams of the market leading platforms at our customers. We maintain close working relationships with these customers so our marketing team can focus on identifying and developing new products that will meet their needs in the future, involving us in the early stages of our customers’ plans to design new equipment. We sell our products both directly and through a network of independent manufacturers’ representatives and distributors. Our direct sales force is technically trained. Expert technical support is critical to our customers’ success and we provide such support through our field applications engineers, technical marketing team and engineering staff, as well as through our extranet technical support web site.

We augment this strategic account sales approach with domestic and foreign distributors that service primarily smaller accounts purchasing standard IC and Board products. Typically, these distributors handle a wide variety of products, including those that compete with our products, and fill orders for many customers. For our RAID products we exclusively use the distribution model and spend a good deal of our sales time supporting their efforts. We use MARCOM expenditures to augment the distribution effort to reach many of our customers. Most of our sales to distributors are made under agreements allowing for price protection and right of return on stipulated quantities of unsold merchandise. Our sales headquarters is located in San Diego, California. We maintain sales offices throughout the world. Net revenues generated from each category of our products is summarized in Item 7, “Management’s Discussion and Analysis of Financial Condition and Results of

9

Table of Contents

Operations.” Information regarding net revenue generated from each of our significant customers, as well as domestic and foreign net revenues from the sale of our products, is provided in Note 11 to the Consolidated Financial Statements.

Backlog

Our sales are made primarily pursuant to standard purchase orders for the delivery of products. Quantities of our products to be delivered and delivery schedules are frequently revised to reflect changes in customers’ needs; customer orders generally can be cancelled or rescheduled without significant penalty to the customer. For these reasons, our backlog as of any particular date is not representative of actual sales for any succeeding period, and therefore, we believe that backlog is not necessarily a good indicator of future revenue.

Competition

In the communications IC markets, we compete primarily against companies such as Agere, Broadcom, Intel, Mindspeed, PMC-Sierra, and Vitesse. Our principal competitors in the PCBA market are Adaptec, Emulex, Qlogic, Agilent, Hewlett-Packard and LSI Logic. In the embedded processor market, we compete with other large technology companies such as Freescale and IBM. In addition, certain of our customers and potential customers have internal IC or storage design or manufacturing capability with which we compete.

The communications IC and storage markets are highly competitive and are subject to rapid technological change, evolving standards, short product life cycles, and price erosion. We typically face competition at the design stage when our customers are selecting which components to use in their next generation equipment. In the storage market, our products can be qualified at any time, and are generally distinguished through a combination of pricing, features and reliability. We believe that the principal factors of competition for the markets we serve include: product performance, quality, reliability, integration, price, and time-to-market, as well as our reputation and level of customer support. Our ability to successfully compete in these markets depends on our ability to design and subcontract the manufacture of new products that implement new technologies and gain end market acceptance in a time efficient and cost effective manner.

Proprietary Rights

We rely in part on patents to protect our intellectual property. We have been issued approximately 172 patents, which principally cover certain aspects of the design and architecture of our IC and enterprise storage products. In addition, we have over 140 inventions in various stages of the patenting process in the United States and abroad. There can be no assurance that our pending patent applications or any future applications will be approved, or that any issued patents will provide us with competitive advantages or will not be challenged by third parties or that if challenged, will be found to be valid or enforceable, or that the patents of others will not have an adverse effect on our ability to do business. There can be no assurance that others will not independently develop similar products or processes, duplicate our products or processes or design around any patents that may be issued to us.

To protect our intellectual property, we also rely on a combination of mask work protection under the Federal Semiconductor Chip Protection Act of 1984, trademarks, copyrights, trade secret laws, employee and third-party nondisclosure agreements and licensing arrangements.

As a general matter, the semiconductor and enterprise storage industries are characterized by substantial litigation regarding patent and other intellectual property rights. In the past we have been, and in the future may be, notified that we may be infringing on the intellectual property rights of third parties. We have certain indemnification obligations to customers with respect to the infringement of third party intellectual property rights by our products. There can be no assurance that infringement claims by third parties or claims for

10

Table of Contents

indemnification by customers or end users of our products resulting from infringement claims will not be asserted in the future or that such assertions, if proven to be true, will not materially adversely affect our business, financial condition or operating results. In the event of any adverse ruling in any such matter, we could be required to pay substantial damages, which could include treble damages, cease the manufacture, use and sale of infringing products, discontinue the use of certain processes or obtain a license under the intellectual property rights of the third party claiming infringement. There can be no assurance that a license would be available on reasonable terms or at all. Any limitations on our ability to market our products, any delays and costs associated with redesigning our products or payments of license fees to third parties or any failure by us to develop or license a substitute technology on commercially reasonable terms could have a material adverse effect on our business, financial condition and operating results.

Environmental Matters

We are subject to a variety of federal, state and local governmental regulations related to the use, storage, discharge and disposal of toxic, volatile or otherwise hazardous chemicals that were used in our manufacturing process. Any failure to comply with present or future regulations could result in the imposition of fines, the suspension of production or a cessation of operations. Such regulations could require us to acquire costly equipment or incur other significant expenses to comply with environmental regulations or clean up prior discharges. Since 1993, we have been named as a potentially responsible party, or PRP, along with a large number of other companies that used Omega Chemical Corporation in Whittier, California to handle and dispose of certain hazardous waste material. We are a member of a large group of PRPs that has agreed to fund certain remediation efforts at the Omega Chemical site, which efforts are ongoing. As of 2003, we closed our wafer fabrication facility in San Diego, and the property has been returned to the landlord.

Employees

As of March 31, 2005, we had 755 full-time employees: 77 in administration, 437 in research and development, 78 in operations, and 163 in marketing and sales. Our ability to attract and retain qualified personnel is essential to our continued success. None of our employees are covered by a collective bargaining agreement, nor have we ever experienced any work stoppage.

Executive Officers of the Registrant

Our executive officers and their ages as of May 31, 2005, are as follows:

Kambiz Hooshmand joined AMCC as President and Chief Executive Officer and as a member of our Board of Directors in March 2005. Prior to AMCC he was with Cisco Systems, where he most recently served as Vice President and General Manager of Cisco’s Optical and Broadband Transport Technology group. At Cisco, Mr.

11

Table of Contents

Hooshmand held several executive-level positions in Multi-Service Switching, DSL, Carrier Core and Multi-Service, and Optical and Broadband Transport business units. He joined Cisco as a director of engineering as part of the StrataCom acquisition in 1996. Mr. Hooshmand was one of the early leaders at StrataCom that developed the nascent Frame Relay market into what is now a multi-billion dollar worldwide services market. Mr. Hooshmand’s work with both domestic and international carriers led to the deployment of many of the largest and most high-availability data networks in the world. Mr. Hooshmand has over two decades of experience in core routing, VoIP, ATM, access and transport technologies. Mr. Hooshmand holds a Master of Science degree in Engineering Management from Stanford University and a Bachelor of Science degree in Electrical Engineering from California State University at Chico.

Thomas L. Tullie joined AMCC in 1996 as Vice President of Sales. Prior to his current position as Chief Operating Officer, he has held executive-level positions in sales and operations, and has led AMCC’s Communications Group as General Manager. Mr. Tullie has been instrumental in managing the organization through a complex transition from a core transport company to a broad-based provider of networking solutions. Mr. Tullie came to AMCC from S-MOS Systems, where he held several strategic sales management positions. Prior to joining S-MOS, Tullie served as a design engineer with Digital Equipment Corporation. Mr. Tullie earned a Master of Business degree from Clark University, and a Bachelor of Science degree from the University of Massachusetts.

Timothy M. Heenan joined us in October 2000 through the acquisition of MMC Networks. Mr. Heenan was promoted to Vice President of Operations in August 2001 and to Senior Vice President in September 2003. Prior to joining MMC Networks, Mr. Heenan was the Director of Test Operations at Cirrus Logic, Inc., where he was responsible for worldwide manufacturing test operations. Before his tenure at Cirrus Logic, Mr. Heenan held various engineering positions at Signetics Corporation. Mr. Heenan holds a B.S. degree in Materials Engineering from Rensselaer Polytechnic Institute and a M.S. degree in Engineering Management from Santa Clara University.

Candace H. Kilburn joined us in September 1996 as Director of Human Resources and was promoted to Vice President in August 1999 and to Senior Vice President in December 2003. From 1990 to 1996, Ms. Kilburn served as Director of Human Resources with Buck Knives Inc., where she was responsible for international human resources. She has also held positions at Handyman Corporation and Rohr Industries. Ms. Kilburn earned a B.S. in Business Administration from United States International University, and a M.B.A. from Chapman University. She is designated as a Senior Professional in Human Resources, a certified Employee Benefits Specialist, and has two certificates in Human Resources Management.

Faye Pairman joined AMCC in 2004 when AMCC acquired 3ware. At 3ware, she served as President and Chief Executive Officer. Ms. Pairman has more than 15 years experience in marketing, sales and executive management. Prior to joining 3ware, Ms. Pairman served eight years in numerous management and executive positions with Adaptec, including VP and General Manager of Storage Networking Solutions Group, Host Interface Solutions Group, and Distribution Products Group. She has also held marketing management positions at SuperMac Technology and the Eastman Kodak Company. Ms. Pairman earned a B.A. from the University of the Pacific and a Masters in Business Administration degree from the Harvard Graduate School of Business Administration.

Jeffery Blazevich, Vice President, Controller and Acting Chief Financial Officer, joined AMCC in 1999 as Assistant Controller and has served as Controller since April 2003. In March 2004, Mr. Blazevich was named Vice President. Mr. Blazevich was named Acting Chief Financial Officer effective January 1, 2005. Prior to AMCC, Mr. Blazevich was an Audit Manager at Ernst & Young, a professional services and registered public accounting firm. He received a Bachelor’s Degree of Accountancy from the University of San Diego. Mr. Blazevich is a certified public accountant and member of the American Institute of Certified Public Accountants.

12

Table of Contents

Brian Wilkie joined us in 2004 to lead the company’s Embedded Products Group. Prior to AMCC, Mr. Wilkie spent 25 years with Motorola Semiconductor, having held a number of PowerPC-related executive positions in semiconductor IC development, product development and operations, including leadership roles in the Microcontroller, Automotive Powertrain and PowerPC Computing Platform Divisions. Mr. Wilkie was instrumental in achieving numerous significant design wins and mass-market acceptance for PowerPC in the networking and automotive arenas. He has also been a consultant to Chartered Semiconductor, and worked as a mentor with STARTech Early Ventures. Mr. Wilkie holds a BSEE degree from Glasgow University in Scotland, in addition to several patents in the field of microcontroller architecture and timers.

Daryn Lau joined us in May 2005 as Vice President and General Manager, Communications Business Unit. Before joining AMCC, Mr. Lau was Vice President and General Manager for the IDT serial switching division. In October 1999, Mr. Lau co-founded ZettaCom and served as president and CEO. ZettaCom was acquired by IDT in May 2004. Prior to his executive role at ZettaCom, Mr. Lau spent seven years at Cisco Systems where he held various technical positions in both enterprise and service provider business units. Prior to Cisco, Mr. Lau led the ASIC development group NET, a telecom equipment vendor and, at Amdahl, the mainframe division of Fujitsu. Mr. Lau holds a Bachelor of Science degree in Electrical Engineering from University of California, Berkeley.

13

Table of Contents

Item 2.    Properties.

Our corporate headquarters are located in San Diego, California. Below is a summary of principal properties leased on May 31, 2005 (net of subleases):

In an effort to improve the efficiency of the workforce and reduce our cost structure, we implemented several plans to consolidate our workforce into certain designated facilities. As a result, approximately 30,000 square feet of unoccupied properties with non-cancelable lease commitments expiring through fiscal 2011 are included in the above summary. On April 29, 2005, the Company negotiated an assignment of the excess lease liability assumed in the JNI acquisition. We believe that the facilities under lease by us will be adequate for at least the next 12 months.

During fiscal year 2005, we purchased two facilities. In June, we purchased a 50,000 square foot building and the adjacent land in Andover, Massachusetts. Additionally, in November, we purchased a 150,000 square foot building in Sunnyvale, California. The Sunnyvale property will consolidate two engineering, sales and marketing offices.

Our foreign locations consist of the following: Kanata, Canada; Manchester and Cheshire, United Kingdom; Munich, Germany; Essonnes and Sophia, France; Tokyo, Japan; Beijing, Shenzhen and Shanghai, People’s Republic of China; and Bangalore, India.

For additional information regarding our obligations under property leases, see Note 9 of Notes to Consolidated Financial Statements, included in Part IV, Item 15 of this Report.

Item 3.    Legal Proceedings.

The information set forth under Note 12 of Notes to Consolidated Financial Statements, included in Part IV, Item 15 of this Report, is incorporated herein by reference.

Item 4.    Submission of Matters to a Vote of Security Holders.

No matters were submitted to a vote of the Company’s stockholders during the fourth quarter of the fiscal year ended March 31, 2005.

14

Table of Contents

PART II

Item 5.    Market for Registrant’s Common Equity, Related Stockholder Matters and Issuer Purchases of Equity Securities.

Our common stock is traded on the Nasdaq National Market under the symbol AMCC. The following table sets forth the high and low sales prices of our common stock as reported by the Nasdaq National Market for the periods indicated.

At May 12, 2005, there were approximately 775 holders of record of our common stock.

Dividend Policy

We have never declared or paid cash dividends on shares of our common stock. We currently intend to retain all of our earnings, if any, for use in our business, for the purchases of our common stock or for the acquisitions of other businesses, assets, products or technologies. We do not anticipate paying any cash dividends in the foreseeable future.

Recent Sales of Unregistered Securities

There were no sales of equity securities by us that were not registered under the Securities Act of 1933 during fiscal 2005.

Securities Authorized for Issuance under Equity Compensation Plans

The information included in Part III, Item 12 of this report, is hereby incorporated herein by reference. For additional information on our stock incentive plans and activity, see Note 5 of Notes to Consolidated Financial Statements, included in Part IV, Item 15 of this Report.

Issuer Purchases of Equity Securities

In August 2004, our Board of Directors authorized a program to repurchase shares of our common stock. The Board approved the repurchase of shares having an aggregate value of up to $200 million, depending on market conditions. These repurchases are made in open market or privately negotiated transactions in compliance with SEC Rule 10b-18, subject to market conditions, applicable legal requirements and other factors. This program does not obligate us to acquire any particular amount of common stock and may be suspended at any time at our discretion.

15

Table of Contents

Below is a summary of stock repurchases for the quarter ended March 31, 2005 (in thousands, except average price per share). See Note 5 of our Notes to Consolidated Financial Statements for additional information regarding our stock repurchase program.

During fiscal 2005, we repurchased on the open market 5.4 million shares of our common stock for approximately $16.9 million. In addition, during fiscal 2005, we entered into structured stock repurchase agreements totaling $59.5 million. These agreements settle in cash or stock depending on the closing market price of our common stock on the expiration date of the agreements. These agreements had various expiration dates through May 31, 2005. Upon expiration of each agreement, if the closing market price of the Company’s common stock is at or above the pre-determined price, we will have our investment returned with a premium. If the closing market price is below the pre-determined price, we will receive a predetermined number of shares. During the year ended March 31, 2005 upon settlement of the underlying agreements, we received 2.5 million shares of our common stock at an effective purchase price of $3.03 per share from the settlements resulting in the delivery of shares and received cash totaling $48.3 million from the settlements resulting in the return of our investment with a premium.

At March 31, 2005, we had four open structured stock repurchase agreements totaling $9.5 million that have varying maturities through May 31, 2005. Under the remaining agreements, we could receive up to $10.3 million of cash, or the delivery of up to 3.0 million shares of our common stock.

16

Table of Contents

Item 6.    Selected Financial Data.

The following table sets forth selected financial data for each of our last five fiscal years ended March 31, 2005. This information includes the results of operations of acquisitions accounted for using the purchase method of accounting commencing as of their respective acquisition dates (Note 4 to the Consolidated Financial Statements). You should read this data together with the Consolidated Financial Statements and related Notes, as well as “Management’s Discussion and Analysis of Financial Condition and Results of Operations”, contained elsewhere in this report.

17

Table of Contents

Item 7.    Management’s Discussion and Analysis of Financial Condition and Results of Operations.

Management’s discussion and analysis of financial condition and results of operations, or MD&A, is provided as a supplement to the accompanying consolidated financial statements and footnotes to help provide an understanding of our financial condition, changes in our financial condition and results of our operations. The MD&A is organized as follows:

CAUTION CONCERNING FORWARD-LOOKING STATEMENTS

This section should be read in conjunction with the consolidated financial statements and notes thereto included elsewhere in this report. This discussion contains forward-looking statements. These forward-looking statements are made as of the date of this report. Any statement that refers to an expectation, projection or other characterization of future events or circumstances, including the underlying assumptions, is a forward-looking statement. We use certain words and their derivatives such as “anticipate”, “believe”, “plan”, “expect”, “estimate”, “predict”, “intend”, “may”, “will”, “should”, “could”, “future”, “potential”, and similar expressions in many of the forward-looking statements. The forward-looking statements are based on our current expectations, estimates and projections about our industry, management’s beliefs, and other assumptions made by us. These statements and the expectations, estimates, projections, beliefs and other assumptions on which they are based are subject to many risks and uncertainties and are inherently subject to change. We describe many of the risks and uncertainties that we face in the “Risk Factors” section of MD&A. We update our descriptions of the risks and uncertainties facing us in our periodic reports filed with the SEC in which we report our financial condition and results for the quarter and fiscal year-to-date. Our actual results and actual events could differ materially from those anticipated in any forward-looking statement. Readers should not place undue reliance on any forward-looking statement.

Overview

AMCC provides semiconductor and board level products that are the essential building blocks for the processing, moving and storing of information worldwide. The company blends systems and software expertise with high-performance, high-bandwidth silicon integration to deliver silicon, hardware and software solutions for global wide area networks (WAN), embedded applications such as PowerPC and programmable SOC architectures, storage area networks (SAN), and high-growth storage markets such as Serial ATA (SATA) RAID. AMCC’s corporate headquarters are located in San Diego, California. Sales and engineering offices are located throughout the world.

18

Table of Contents

Over the last several years, we have undertaken significant restructuring activities in an effort to reduce operating costs. In addition, in an effort to diversify our customer base and markets that we serve, we have also made several acquisitions. In September 2003 and January 2004, we purchased assets and licensed intellectual property associated with IBM’s PowerPRS Switch Fabric product line, or the PRS Business, for approximately $51 million in cash to complement our existing communications products portfolio. In October 2003, we completed the acquisition of all outstanding shares of JNI Corporation, a provider of Fibre Channel hardware and software products that are critical elements of storage networks, for approximately $196.4 million in cash. In April 2004, we completed the acquisition of 3ware, Inc., a provider of high-performance, high-capacity SATA storage solutions, for a purchase price of approximately $145 million in cash. In May 2004 and December 2004, we acquired intellectual property and a portfolio of assets associated with IBM’s 400 series of embedded PowerPC^® standard products, or the Embedded Products Business, for approximately $232 million in cash. The PowerPC 400 series product line targets Internet, communication, data storage, consumer and imaging applications.

CRITICAL ACCOUNTING POLICIES

The preparation of financial statements in accordance with accounting principles generally accepted in the United States requires us to make estimates and assumptions that affect the reported amounts of assets and liabilities at the date of the financial statements and the reported amounts of net revenue and expenses in the reporting period. We regularly evaluate our estimates and assumptions related to inventory valuation and warranty liabilities, which affects our cost of sales and gross margin; the valuation of purchased intangibles and goodwill, which affects our amortization and impairments of goodwill and other intangibles; the valuation of restructuring liabilities, which affects the amount and timing of restructuring charges; and the valuation of deferred income taxes, which affects our income tax expense and benefit. We also have other key accounting policies, such as our policies for revenue recognition, including the deferral of a portion of revenues on sales to distributors, and allowance for bad debts. The methods, estimates and judgments we use in applying these most critical accounting policies have a significant impact on the results we report in our financial statements. We base our estimates and assumptions on historical experience and on various other factors that we believe to be reasonable under the circumstances, the results of which form the basis for making judgments about the carrying values of assets and liabilities that are not readily apparent from other sources. The actual results experienced by us may differ materially and adversely from management’s estimates. To the extent there are material differences between our estimates and the actual results, our future results of operations will be affected.

We believe the following critical accounting policies require us to make significant judgments and estimates in the preparation of our consolidated financial statements.

Inventory Valuation and Warranty Liabilities

Our policy is to value inventories at the lower of cost or market on a part-by-part basis. This policy requires us to make estimates regarding the market value of our inventories, including an assessment of excess or obsolete inventories. We determine excess and obsolete inventories based on an estimate of the future demand for our products within a specified time horizon, generally 12 months. The estimates we use for future demand are also used for near-term capacity planning and inventory purchasing and are consistent with our revenue forecasts. If our demand forecast is greater than our actual demand we may be required to take additional excess inventory charges, which would decrease gross margin and net operating results in the future. Our products typically carry a one to three year warranty. We establish reserves for estimated product warranty costs at the time revenue is recognized. Although we engage in extensive product quality programs and processes, our warranty obligation is affected by product failure rates, use of materials and service delivery costs incurred in correcting any product failure. Should actual product failure rates, use of materials or service delivery costs differ from our estimates, additional warranty reserves could be required, which could reduce our gross margins.

19

Table of Contents

Goodwill and Intangible Asset Valuation

The purchase method of accounting for acquisitions requires extensive use of accounting estimates and judgments to allocate the purchase price to the fair value of the net tangible and intangible assets acquired, including in-process research and development, or IPR&D. Goodwill and intangible assets deemed to have indefinite lives are not amortized, but are subject to annual impairment tests. The amounts and useful lives assigned to other intangible assets impact future amortization, and the amount assigned to IPR&D is expensed immediately. Determining the fair values and useful lives of intangible assets especially requires the use of estimates and the exercise of judgment. While there are a number of different generally accepted valuation methods to estimate the value of intangible assets acquired, we primarily use the discounted cash flow method and the market comparison approach. These methods require significant management judgment to forecast the future operating results used in the analysis. In addition, other significant estimates are required such as residual growth rates and discount factors. The estimates we use to value and amortize intangible assets are consistent with the plans and estimates that we use to manage our business and are based on available historical information and industry estimates and averages. These judgments can significantly affect our net operating results.

We are required to assess goodwill impairment annually using the methodology prescribed by Statement of Financial Accounting Standards, or SFAS, No. 142, “Goodwill and Other Intangible Assets.” SFAS No. 142 requires that goodwill be tested for impairment at the reporting unit level on an annual basis and between annual tests in certain circumstances. Application of the goodwill impairment test requires judgment, including the identification of reporting units, assigning assets and liabilities to reporting units, assigning goodwill to reporting units and determining the fair value of each reporting unit. In fiscal 2005, in accordance with SFAS No. 142, management determined that there were three reporting units to be tested. The goodwill impairment test compares the implied fair value of the reporting unit with the carrying value of the reporting unit. The implied fair value of goodwill is determined in the same manner as in a business combination. Determining the fair value of the implied goodwill is judgmental in nature and often involves the use of significant estimates and assumptions. These estimates and assumptions could have a significant impact on whether or not an impairment charge is recognized and also the magnitude of any such charge. Estimates of fair value are primarily determined using discounted cash flows and market comparisons. These approaches use significant estimates and assumptions, including projection and timing of future cash flows, discount rates reflecting the risk inherent in future cash flows, perpetual growth rates, determination of appropriate market comparables, and determination of whether a premium or discount should be applied to comparables. It is reasonably possible that the plans and estimates used to value these assets may be incorrect. If our actual results, or the plans and estimates used in future impairment analyses, are lower than the original estimates used to assess the recoverability of these assets, we could incur additional impairment charges.

For fiscal 2005, the discounted cash flows for each reporting unit were based on discrete five-year financial forecasts developed by management for planning purposes. Cash flows beyond the five-year discrete forecasts were estimated using terminal value calculations. The sales compound annual growth rates ranged from 15% to 33% for the reporting units during the discrete forecast period and the future cash flows were discounted to present value using a discount rate of 17% and terminal growth rates of 7.5%. We did not recognize any goodwill impairment as a result of performing this annual test. A variance in the discount rate or the estimated revenue growth rate could have a significant impact on the estimated fair value of the reporting unit and consequently the amount of identified goodwill impairment. For example, a 1% - 2% increase in the discount rate would have resulted in an indication of possible impairment that would have led us to further quantify the impairment and potentially record a charge to write-down these assets.

Restructuring Charges

Over the last three years we have undertaken significant restructuring initiatives, which have required us to develop formalized plans for exiting certain business activities and reducing spending levels. We have had to record estimated expenses for employee severance, long-term asset write downs, lease cancellations, facilities consolidation costs, and other restructuring costs. Given the significance, and the timing of the execution, of such

20

Table of Contents

activities, this process is complex and involves periodic reassessments of estimates made at the time the original decisions were made. Prior to 2003, the liability for certain exit costs was recognized on the date that management committed to a plan. In 2003, new accounting guidance was issued requiring us to recognize costs associated with our exit and disposal activities at fair value when a liability is incurred. In calculating the charges for our excess facilities, we have to estimate the timing of exiting certain facilities and then estimate the future lease and operating costs to be paid until the lease is terminated and the amount of any sublease income. To form our estimates for these costs, we performed an assessment of the affected facilities and considered the current market conditions for each site. Our assumptions for the operating costs until termination or the offsetting sublease revenues may turn out to be incorrect, and our actual costs may be materially different from our estimates, which could result in the need to record additional costs or to reverse previously recorded liabilities. Our policies require us to periodically evaluate, at least semiannually, the adequacy of the remaining liabilities under our restructuring initiatives.

Valuation of Deferred Income Taxes

We record valuation allowances to reduce our deferred tax assets to an amount that we believe is more likely than not to be realized. We consider estimated future taxable income and ongoing prudent and feasible tax planning strategies, including reversals of deferred tax liabilities, in assessing the need for a valuation allowance. If we were to determine that we will not realize all or part of our deferred tax assets in the future, we would make an adjustment to the carrying value of the deferred tax asset, which would be reflected as income tax expense. Conversely, if we were to determine that we will realize a deferred tax asset, which currently has a valuation allowance, we would reverse the valuation allowance which would be reflected as an income tax benefit or as an adjustment to stockholders’ equity, for tax assets related to stock options, or goodwill, for tax assets related to acquired businesses.

Revenue Recognition

We recognize revenue in accordance with SEC Staff Accounting Bulletin No. 101 “Revenue Recognition in Financial Statements”, or SAB 101 as well as SAB 104, “Revenue Recognition”. These pronouncements require that four basic criteria be met before revenue can be recognized: 1) there is evidence that an arrangement exists; 2) delivery has occurred; 3) the fee is fixed or determinable; and 4) collectibility is reasonably assured. We recognize revenue upon determination that all criteria for revenue recognition have been met. In addition, we do not recognize revenue until all customers’ acceptance criteria have been met. The criteria are usually met at the time of product shipment, except for shipments to distributors with rights of return. Revenue from shipments to distributors subject to rights of return is deferred until all return or cancellation privileges lapse. Revenue from shipments to distributors without return rights is recognized upon shipment. In addition, we record reductions to revenue for estimated allowances such as returns, competitive pricing programs and rebates. These estimates are based on our experience with product returns and the contractual terms of the competitive pricing and rebate programs. Shipping terms are generally FCA shipping point. If actual returns or pricing adjustments exceed our estimates, additional reductions to revenue would result.

Allowance for Bad Debt

We maintain an allowance for doubtful accounts for estimated losses resulting from the inability of our customers to make required payments. Our allowance for doubtful accounts is based on our assessment of the collectibility of specific customer accounts, the aging of accounts receivable, our history of bad debts, and the general condition of the industry. If a major customer’s credit worthiness deteriorates, or our customers’ actual defaults exceed our historical experience, our estimates could change and impact our reported results.

21

Table of Contents

RESULTS OF OPERATIONS

Comparison of the Year Ended March 31, 2005 to the Year Ended March 31, 2004

Net Revenues.    Net revenues for the year ended March 31, 2005 were approximately $253.8 million, representing an increase of 93.4% from the net revenues of approximately $131.2 million for the year ended March 31, 2004. The increase in total net revenues was primarily attributable to the revenues generated from our fiscal 2004 and 2005 acquisitions, as well as increases in revenue from our existing communications product portfolio, offset by decreases in other revenue. See the following tables (dollars in thousands):

Communication revenue consists of our historical communications business and the PRS Business acquisition. Storage revenue is derived from of our JNI acquisition and our 3ware acquisition. The Embedded Products revenue consists of our Embedded Products Business acquisition and other revenue consists of our legacy non-communications products, such as computer, ATE and military products.

Based on direct shipments, net revenues to customers that exceeded 10% of total net revenues in any of the three years ended March 31, 2005 were as follows:

Looking through product shipments to distributors and subcontractors to the end customers, net revenues to end customers that exceeded 10% of total net revenues in any of the three years ended March 31, 2005 were as follows:

The decline in revenues attributable to Harris Corporation was due to the fulfillment of certain last time buy orders of non-communications products in fiscal 2003.

Revenues based on direct shipments outside the United States of America accounted for 52% of net revenues for the year ended March 31, 2005 compared to 46% for the year ended March 31, 2004.

22

Table of Contents

Gross Profit.    The following table presents net revenues, cost of revenues and gross profit for fiscal years ended March 31, 2005 and March 31, 2004 (dollars in thousands):

The increase in gross profit for the year ended March 31, 2005 was primarily attributable to the increase in revenues, offset by an increase of $14.4 million and $825,000 of amortization of developed technology and purchased inventory fair value adjustment included in our cost of revenues, respectively.

The amortization of purchased intangible assets included in cost of revenues during the year ended March 31, 2005 was $23.3 million compared to $9.0 million for the year ended March 31, 2004. The increase is a result of our fiscal 2005 acquisitions. Based on the amount of capitalized purchased intangibles on the balance sheet as of March 31, 2005, we expect amortization expense for purchased intangibles charged to cost of revenues to be $17.6 million and $14.5 million for each of the fiscal years ending March 31, 2006, and 2007 and $50.4 million for fiscal periods thereafter, respectively. Future acquisitions of businesses may result in substantial additional charges which would impact the gross margin in future periods.

Research and Development and Selling, General and Administrative Expenses.    The following table presents research and development and selling, general and administrative expenses for fiscal years ended March 31, 2005 and March 31, 2004 (in thousands):

Research and Development.    Research and development, or R&D, expenses consist primarily of salaries and related costs of employees engaged in research, design and development activities, costs related to engineering design tools, subcontracting costs and facilities expenses. The increase in R&D of 5.4% for the year ended March 31, 2005, respectively, was primarily due to higher payroll, the related benefits expense, and facilities costs of approximately $18.8 million, resulting from our fiscal 2004 and fiscal 2005 acquisitions, offset by lower design costs and software and equipment depreciation costs of approximately $7.0 million and $5.7 million, respectively, resulting from our restructuring initiatives. We believe that a continued commitment to R&D is vital to our goal of maintaining a leadership position with innovative communications, storage and embedded products. Currently, R&D expenses are focused on the development of communications, storage, and embedded products and we expect to continue this focus. Future acquisitions of businesses may result in substantial additional on-going costs.

Since the start of fiscal 2003, we have invested a total of approximately $363.1 million in the research and development of new products, including higher-speed, lower-power and lower-cost products, products that combine the functions of multiple existing products into single highly integrated products, and other products to complete our portfolio of communications and storage products. For most products developed by us, due to their complexity and the complexity of our OEM customers’ equipment, it often takes several years to complete development and qualification. We have not yet generated significant revenues from many of these new products for two additional reasons. First, the dramatic and extended downturn in the telecommunications market has

23

Table of Contents

severely impacted our customers and has resulted in significantly less demand for the quantity of these products than expected when some of the developments commenced. Second, as a result of restructuring activities, we have discontinued development of several new products and slowed down development of other new products as we realized that demand for these products would not materialize as originally anticipated.

Selling, General and Administrative.    Selling, general and administrative, or SG&A, expenses consist primarily of personnel-related expenses, professional and legal fees, corporate branding and facilities expenses. The increase in SG&A expenses of 32.6% for the year ended March 31, 2005 was primarily due to the effect of higher payroll and related benefits, corporate branding, commissions, and contracted services of approximately $16.8 million, as a result of our fiscal 2004 and fiscal 2005 acquisitions, offset by lower professional services and insurance expense of $2.1 million. Future acquisitions of businesses may result in substantial additional on-going costs.

Stock-Based Compensation.    The following table presents stock-based compensation expense for employees engaged in R&D and S,G&A activities for the fiscal years ended March 31, 2005 and March 31, 2004, all of which was excluded from those operating expenses (in thousands):

Stock-based compensation expense represents the amortization of deferred compensation related to acquisitions. Deferred compensation is the difference between the fair value of our common stock at the date of each acquisition and the exercise price of the unvested stock options assumed in the acquisition. In fiscal 2005, we recorded approximately $19.0 million of deferred compensation in connection with stock options assumed in our acquisition of 3ware. Stock-based compensation charges, including amounts charged to cost of revenues, were $9.3 million for the year ended March 31, 2005, compared to $21.2 million for the year ended March 31, 2004. We currently expect to record amortization of deferred compensation with respect to these assumed options of approximately $6.1 million in fiscal 2006, $2.8 million in fiscal 2007, and $200,000 in fiscal 2008. These charges could be reduced as a result of employee turnover. Acquisitions of businesses may result in substantial additional on-going costs. Such charges may cause fluctuations in our interim or annual operating results.

Acquired In-process Research and Development.    For the fiscal year ended March 31, 2005, we recorded $13.4 million of acquired IPR&D resulting from the acquisitions of 3ware and the Embedded Products Business. These amounts were expensed on the acquisition dates because the acquired technology had not yet reached technological feasibility and had no future alternative uses. The IPR&D charge related to the 3ware acquisition was made up of two projects that were 42% and 25% complete at the date of acquisition. The estimated cost to complete these projects was $650,000 and $2.3 million, respectively, and the discount rate applied to calculate the IPR&D charge was 30% and 35%, respectively. The IPR&D charge related to the Embedded Products Business acquisition was made up of three projects, which were between 42% and 69% complete at the date of acquisition. The estimated aggregate cost to complete these projects was $9.1 million. The discount rate applied to calculate the IPR&D charge ranged from 25% to 30%.

For the year ended March 31, 2004, we recorded $21.8 million of acquired IPR&D resulting from the acquisition of JNI Corporation and the PRS business. These amounts were expensed on the acquisition dates because the acquired technology had not yet reached technological feasibility and had no future alternative uses. The IPR&D charge related to the PRS acquisition was made up of five projects which were between 38% and

24

Table of Contents

68% complete at the date of acquisition. The estimated aggregate cost to complete these projects was $5.3 million. The discount rate applied to calculate the IPR&D charge ranged from 20% to 30%. The IPR&D charge related to the JNI Corporation acquisition was made up of six projects, which were between 33% and 88% complete at the date of acquisition. The estimated aggregate cost to complete these projects was $2.3 million. The discount rate applied to calculate the IPR&D charge ranged from 22% to 35%. There can be no assurance that acquisitions of businesses, products or technologies by us in the future will not result in substantial charges for acquired IPR&D that may cause fluctuations in our interim or annual operating results.

Goodwill and Purchased Intangible Asset Impairment Charges.    To coincide with our annual long-range planning process, we assess goodwill for impairment annually in the fourth quarter, or more frequently if the indicators of impairment are present. As a result of the November 2004 restructuring program, we decided to reduce our future investment in certain products acquired in the JNI transaction. During the third quarter of fiscal 2005, we performed an impairment test as required by SFAS 144 which resulted in a non-cash charge of $27.3 million for the impairment of purchased intangibles. This charge is recorded in operating expenses in the consolidated statement of operations for the year ended March 31, 2005.

Based upon an analysis performed in the fourth quarter of fiscal 2005, which included a discounted cash flow analysis, as well as market comparables, no impairment of goodwill or other purchased intangibles was evident.

Restructuring Charges.    The following table presents restructuring charges for the fiscal years ended March 31, 2005 and March 31, 2004 (in thousands):

In July 2001, we announced the first of our restructuring programs. The July 2001 restructuring plan was in response to the sharp downturn in business at the end of the Company’s fiscal 2001 and included reducing our overall cost structure and aligning manufacturing capacity with the then current demand. The restructuring plan consisted of the elimination of approximately 50 employees, or 5% of our workforce, the consolidation of excess facilities and the write-off of certain property and equipment. As a result of the July 2001 restructuring, we recorded a charge of $11.6 million consisting of $900,000 for employee severances, $2.0 million of non-cancelable lease commitments, and $8.7 million for the disposal of excess manufacturing equipment, the abandonment of certain leasehold improvements, and the write off of software licenses.

As a result of our July 2001 restructuring activities, we realized approximately $4 million of annual savings relating to fixed cost of sales overhead and approximately $2 million of annual savings relating to operating expenses. During fiscal 2005, we have completed the restructuring activities contemplated by the July 2001 restructuring plan and no further payments or expenses are anticipated under this program.

In July 2002, we announced our second restructuring program, as a result of the prolonged downturn in the telecommunications industry. The July 2002 restructuring program consisted of the closure of our wafer manufacturing facility in San Diego and a workforce reduction of approximately 165 employees or 25% of our workforce. During fiscal 2003, we recognized a total charge of $7.0 million consisting of $4.0 million for the estimated facility restoration costs and the severance packages for approximately 70 manufacturing employees, and of $3.0 million consisting of employee severances, related to the closure of a United States design center and the disposal of its related assets.

During the third quarter of fiscal 2004, we completed the activities contemplated by the plan. As a result, we recorded an adjustment to the restructuring liability for the excess accrued severance and facilities restoration

25

Table of Contents

costs, and recognized a restructuring benefit of approximately $537,000. We do not expect any future charges or benefits related to the closure of the wafer manufacturing facility. As a result of the closure of our internal wafer manufacturing facility and workforce reduction, we realized annual savings totaling approximately $14 million relating to fixed cost of sales overhead and approximately $16 million of annual savings relating to operating expenses in fiscal 2004.

In April 2003, we announced our third restructuring program. As the downturn in the telecommunications industry continued, it became evident that further cost reductions were necessary. The April 2003 restructuring program consisted of a workforce reduction of 185 employees, further consolidation of excess facilities and additional fixed asset disposals. In June 2002, the FASB issued SFAS 146 requiring that costs associated with exit or disposal activities be recognized when they are incurred rather than at the date of a commitment to an exit or disposal plan. Accordingly, approximately $281,000 was charged for severance packages communicated to employees in March 2003. The remaining restructuring costs of $23.5 million were recognized in the first quarter of fiscal 2004 and consisted of approximately $5.4 million for employee severances, $7.2 million representing the discounted cash flow of lease payments on exited facilities, $3.4 million for the disposal of certain software licenses, and $7.5 million for the write off of leasehold improvements and property and equipment.

As a result of our April 2003 restructuring activities, we realized approximately $4 million of annual savings relating to fixed cost of sales overhead and approximately $34 million of annual savings relating to operating expenses in fiscal 2004. However, in November 2003 we elected to reoccupy a portion of the 58,000 square foot building in San Diego. This decision was based on the acquisition of JNI Corporation and the need to integrate the operations of the two companies in order to achieve the planned cost savings. As a result of this decision to reoccupy the San Diego building, we reversed a portion of the prior accrual for the excess lease commitment and reinstated the book value of the leasehold improvements, which were previously abandoned. We recorded a net restructuring benefit of approximately $2.4 million related to this activity. In addition, we recorded an adjustment to the amount of accrued severance of approximately $200,000 because we overestimated the amount of severance that would be paid.

In November 2003, we implemented a fourth workforce reduction and restructuring. The November 2003 workforce reduction was implemented as a means to achieve certain cost savings anticipated in connection with the fiscal 2004 acquisitions. The restructuring consisted of the elimination of approximately 50 employees and the abandonment of certain leased property. As a result of the November 2003 restructuring, we recorded a charge of approximately $2.8 million, consisting of $1.2 million for employee severances and $1.6 million for excess facilities costs. During fiscal 2005, we completed the restructuring activities contemplated by the November 2003 workforce reduction program and no further payments or expenses are anticipated under this program. We achieved annual operating expense savings of approximately $8 million in fiscal 2005, as a result of our November 2003 workforce reduction.

In November 2004, we implemented a fifth workforce reduction and realignment. The November 2004 workforce reduction was implemented as a means to reduce ongoing operating expenses by restructuring our operations, consolidating our facilities and reducing our workforce. The restructuring consisted of the elimination of approximately 150 employees, or 20 percent of our workforce, the closure of our Israel facility and consolidating other locations. As a result of the November 2004 restructuring, we recorded a charge of approximately $9.1 million, consisting of $4.4 million for employee severance, $4.2 million for property and equipment write-offs, and $500,000 for the closure of its Israel facility and abandonment of certain leased property. We estimate that as a result of the November 2004 workforce reduction, we will achieve annual operating expense savings of approximately $24 million to $32 million annually.

26

Table of Contents

Interest and Other Income and Expenses.    The following table presents interest and other income and expenses for the fiscal years ended March 31, 2005 and March 31, 2004 (dollars in thousands):

Interest Income, net.    Net interest income reflects interest earned on cash and cash equivalents and short-term investment balances, as well as realized gains and losses from the sale of short-term investments, less interest expense on our debt and capital lease obligations. The decrease for the year ended March 31, 2005 is primarily due to lower cash and short-term investment balances as well as lower net realized gains on the portfolio.

Other Income, net.    Other income, net includes recorded gains on strategic equity investments as well as net gains from the sale of real estate, and property and equipment. The decrease is due to lower gains on strategic equity investments, on the sales of real estate, and on the sale of property and equipment recognized in fiscal 2004 and not recognized in fiscal 2005.

Income Taxes.    The following table presents our income tax expense (benefit) for the fiscal years ending March 31, 2005 and March 31, 2004 (dollars in thousands):

The federal statutory income tax rate was 35% for the fiscal years ending March 31, 2005 and March 31, 2004. Our income tax expense (benefit) in 2005 and 2004 primarily represents taxes on our foreign operations and in fiscal year 2004, the realization of a tax benefit resulting from the reversal of certain prior period tax accruals related to the completion of IRS audits for the fiscal years through 2001. In addition, as a result of our cumulative losses and the full utilization of our loss carry backs, we provided a full valuation allowance against our net deferred tax assets in 2005 and 2004.

Comparison of the Year Ended March 31, 2004 to the Year Ended March 31, 2003

Net Revenues.    Net revenues for the year ended March 31, 2004 were approximately $131.2 million, representing an increase of 29% from the net revenues of approximately $101.6 million for the year ended March 31, 2003. The increase in total net revenues was attributable to an increase in communications revenue, including revenues generated by the PRS business of $13.2 million from the date of acquisition, and the revenue generated in the storage business from the acquisition of JNI Corporation of $13.0 million from the date of acquisition, offset by decreases in other revenue of 61% as a result of higher shipments of last time buy products in the prior year. See the following table (dollars in thousands):

27

Table of Contents

Revenues based on direct shipments outside the United States of America accounted for 46% of net revenues for the year ended March 31, 2004, compared to 41% for the year ended March 31, 2003.

Gross Profit.    The following table presents net revenues, cost of revenues and gross profit for the fiscal years ended March 31, 2004 and March 31, 2003 (dollars in thousands):

The increase in gross profit in 2004 was primarily attributable to the reduced fixed cost of manufacturing overhead of approximately $16.0 million as a result of the permanent closure of our internal wafer fabrication facility in March 2003 and the effects of the workforce reductions, as well as decreased stock-based compensation charges included in cost of revenues of $2.0 million for the year ended March 31, 2004. In addition, during the year ended March 31, 2004, we recognized a benefit of approximately $1.1 million related to the sales of inventory, which had been previously reserved. Partially offsetting the increases in gross margin was an increase in the amortization of purchased intangibles of $4.1 million as a result of our fiscal 2004 acquisitions.

Research and Development and Selling, General and Administrative Expenses.    The following table presents R&D and S,G&A expenses for fiscal years ended March 31, 2004 and March 31, 2003 (dollars in thousands):

Research and Development.    The decrease in R&D expenses for the year ended March 31, 2004 was primarily due to lower payroll and related benefits expense of approximately $9.6 million and lower software and equipment depreciation costs of approximately $8.7 million, the results of our restructuring initiatives. These decreases were partially offset by increases in payroll and related benefits resulting from the acquisition of JNI Corporation on October 28, 2003 and our acquisition of the PRS business on September 30, 2003.

Selling, General and Administrative.    The decrease in SG&A expenses for the year ended March 31, 2004 was primarily due to the effect of lower payroll and related benefits expense of approximately $6.1 million following workforce reductions as well as lower legal and professional fees and commissions of approximately $3.3 million and $4.0 million, respectively.

Stock-Based Compensation.    The following table presents stock-based compensation expense for employees engaged in R&D and S,G&A activities expenses for the fiscal years ended March 31, 2004 and March 31, 2003, all of which was excluded from those operating expenses (dollars in thousands):

28

Table of Contents

In the third quarter of fiscal 2004, we recorded approximately $4.2 million of deferred compensation in connection with stock options assumed in our purchase acquisition of JNI Corporation. Stock-based compensation charges, including amounts charged to cost of revenues, were $21.2 million for the year ended March 31, 2004, compared to $131.9 million for the year ended March 31, 2003.

Goodwill and Purchased Intangible Asset Impairment Charges.    To coincide with our annual long-range planning process, we assess goodwill for impairment annually in the fourth quarter, or more frequently if the indicators of impairment are present. Based on the analysis performed in the fourth quarter of fiscal 2004, which included a discounted cash flow analysis, as well as an evaluation of market comparables, no impairment of goodwill or other purchased intangibles was evident. The following table presents goodwill and purchased intangible asset impairment charges for the fiscal years ended March 31, 2004 and March 31, 2003 (dollars in thousands):