Wantstats Reports Company List - Profiles & Intelligence

Advanced Micro Devices, Inc. is a fabless designer of high-performance and adaptive computing silicon and, increasingly, of complete rack-scale AI systems. It is the only company that simultaneously holds a credible x86 server CPU franchise, a merchant data-center GPU franchise, a leading semi-custom console franchise, and the industry's second-largest FPGA/adaptive-SoC franchise. Over the five years to FY2025 the company transformed from a challenger PC and console silicon vendor into a data-center-led platform company: Data Center revenue rose from a minority contributor to 48% of FY2025 revenue and 58% of Q2 2026 revenue. The strategic centre of gravity is now the AMD Instinct accelerator line, the ROCm open software stack and the Helios rack-scale platform, underwritten by multi-gigawatt binding commitments from OpenAI (6 GW) and Anthropic (up to 2 GW). AMD's structural advantages are its TSMC-anchored process leadership, chiplet and advanced-packaging IP, an open-ecosystem software posture positioned against a proprietary incumbent, and a balance sheet in net cash. Its principal vulnerabilities are software-ecosystem depth relative to CUDA, customer concentration, export-control exposure to China, and a valuation that discounts near-flawless execution of the MI450/Helios ramp. FY2025 headcount is per the FY2025 10-K. FY2021–FY2024 figures are as compiled by market-data aggregators from AMD's annual disclosures and should be treated as approximate; the FY2022 step-change reflects the February 2022 Xilinx acquisition, and the FY2025 increase reflects the ZT Systems design-team acquisition plus aggressive AI engineering hiring, which AMD cites as the principal driver of the 25% FY2025 increase in R&D expense. --- ### The company's own characterisation In its FY2025 Form 10-K, AMD describes itself as a driver of innovation in high-performance and AI computing whose technology powers experiences across cloud and AI infrastructure, embedded systems, AI PCs and gaming. It positions its portfolio as a full-stack proposition spanning AI-optimised CPUs, GPUs, networking silicon and software. For data centres it supplies AI accelerators, server CPUs, GPUs, data processing units, AI network interface cards and FPGAs, and it now also supplies data-centre rack-scale platform designs. In client computing it supplies CPUs, APUs and chipsets, and it notes that it was the first company to integrate a dedicated neural processing unit on the same system-on-chip as an x86 CPU. Its embedded portfolio spans CPUs, APUs, FPGAs, system-on-modules and adaptive SoCs for automotive, industrial, healthcare, storage and networking markets. AMD emphasises that it develops the accompanying software stacks — development tools, compilers and drivers — and maintains an open AI software stack across frameworks and models. AMD's stated strategy, in its own framing, rests on the belief that AI will become pervasive across cloud, edge and endpoint devices, and that the breadth of its compute engines, intellectual property and software enablement uniquely positions it to be an end-to-end AI leader. It identifies accelerating Data Center growth as the key 2025 priority and describes an annual cadence for Instinct GPU leadership products. ### Independent characterisation AMD is best understood as three businesses of very different economic character bolted to a single R&D and IP engine: **A hyperscale silicon franchise (Data Center).** This is a design-win business with extreme customer concentration, multi-year qualification cycles, and — since 2024 — a rapidly changing commercial architecture. Historically AMD sold chips to OEMs and ODMs who sold systems to hyperscalers. Following the ZT Systems acquisition, AMD now sells *reference rack designs* and, in the Helios era, integrated rack-scale solutions, moving up the value chain toward Nvidia's system-level model. The economics are shifting from component gross margin (high) toward system-level content (larger dollar TAM, structurally lower percentage margin, higher working-capital intensity). The Q2 2026 balance sheet shows the signature of this shift: accounts payable more than doubled sequentially to 5,359 and property and equipment rose to 3,439 from 2,312 at year-end. **A share-gain franchise (Client and Gaming).** Client is a merchant CPU business competing against a single structurally impaired incumbent. AMD's revenue growth here has been driven far more by mix and pricing than by unit volume expansion in a flat-to-declining PC TAM: FY2025 Client revenue rose 51% on a 31% unit increase and a 15% ASP increase. Gaming has two distinct components — a contractual, console-cycle-dependent semi-custom business (Sony PlayStation 5, Microsoft Xbox Series S/X, and the Valve Steam Machine) with predictable decay and step-function renewal, and a merchant discrete GPU business where AMD is a structural number two. **An industrial-cycle franchise (Embedded).** Overwhelmingly the legacy Xilinx business. Long product lifecycles, high gross margins, sticky design wins in aerospace/defence, test and measurement and communications, but exposed to distributor inventory cycles — as the FY2025 3% revenue decline demonstrates. ### Revenue model AMD's revenue is almost entirely product revenue from the sale of semiconductors and, increasingly, semiconductor-based systems. There is no material subscription revenue. Three secondary revenue mechanisms exist: **Non-recurring engineering (NRE) fees:** charged to semi-custom and custom-silicon customers for design and development services, recognised alongside subsequent product sales. - **IP licensing**, which AMD states it may undertake "from time to time." Effective Q1 2025 licensing gains were reclassified against Marketing, General and Administrative expense because the amounts were immaterial — a useful indication that licensing is not a meaningful profit centre. - **Rack-scale systems revenue**, a new and growing category post-ZT, where the unit of sale is a populated rack rather than a chip. ### Value-chain position AMD is fabless. It occupies the design, IP and system-architecture layer and outsources every physical manufacturing step: **Wafer fabrication:** TSMC for all HPC, FPGA and adaptive-SoC wafers, and for all microprocessor and GPU products at 7nm and below; GLOBALFOUNDRIES for microprocessor and GPU products at nodes larger than 7nm, specifically 12nm and 14nm, under a wafer supply agreement providing minimum annual capacity allocation and set pricing through 2026; TSMC, United Microelectronics Corporation and Samsung Electronics for programmable logic devices. **Assembly, test, mark and packaging (ATMP):** two joint ventures with affiliates of Tongfu Microelectronics Co., Ltd., which provide the majority of ATMP services, plus Siliconware Precision Industries (SPIL) and King Yuan Electronics (KYEC). **Board-level graphics manufacturing:** outsourced to third-party manufacturers and add-in-board partners. This model means AMD's cost of goods is largely a pass-through of foundry and packaging pricing, and its principal operating leverage comes from ASP and mix rather than from manufacturing efficiency. It also means AMD carries no fab capex burden — FY2025 capital expenditure was 974, or 2.8% of revenue, versus the tens of billions spent annually by integrated device manufacturers. ### Customer types and routes to market | Segment | Customer types | Route to market | |---|---|---| | Data Center | Hyperscale cloud providers and hyperscale private data centres; OEMs; ODMs; system integrators; independent distributors | Direct sales force; OEM/ODM channel; distributors; increasingly direct rack-scale supply | | Client and Gaming | PC OEMs; independent distributors; ODMs manufacturing motherboards; add-in-board manufacturers; independent game-console and portable-gaming-device manufacturers; contract manufacturers for AMD-branded graphics cards | OEM partnerships; distribution; AIB partners; direct semi-custom contracts | | Embedded | Aerospace and defence, automotive, industrial, vision and healthcare, communications infrastructure, test and measurement, emulation and prototyping, audio/video/broadcast, data centre | Direct sales plus a network of distributors, OEM partners, value-added resellers and integrated solution vendors (notably for Alveo) | AMD explicitly warns in its risk factors that it depends on a small number of customers for a substantial portion of its business and expects this to continue. It also notes the emerging structural risk that its own hyperscale customers are internally developing competing data-centre microprocessors and accelerators. ### End markets served Cloud and hyperscale AI infrastructure; enterprise IT and on-premise servers; supercomputing and exascale HPC; sovereign AI; consumer and commercial PCs (desktop, notebook, workstation, handheld); gaming consoles and cloud gaming; professional visualisation and content creation; aerospace and defence; automotive (infotainment, ADAS-adjacent adaptive compute); industrial automation and robotics; medical imaging and healthcare; wired and wireless communications infrastructure; test and measurement; broadcast and professional AV; capital-markets low-latency networking (Solarflare).

Lenovo occupies a position no other technology company holds: undisputed global leader in personal computing, with roughly one in four PCs shipped worldwide carrying its brands, married to a fast-scaling enterprise infrastructure business that has just crossed US$19 billion of annual revenue and turned profitable, and a US$10 billion services arm earning better than 22 percent operating margins. The strategic thesis management calls "Hybrid AI" — one personal AI across many devices, plus enterprise AI that keeps customer data close to the customer — is the connective tissue. Financially, Lenovo remains a low-margin, high-turnover, working-capital-financed business: 15.4 percent gross margin and 2.3 percent net margin on US$83.1 billion of FY2025/26 revenue. The investment question is whether the AI infrastructure cycle permanently re-rates the mix, or whether Lenovo remains, at core, a superbly executed but structurally thin-margin PC company that has rented a seat at the AI table. Lenovo Group Limited is a Hong Kong–incorporated, Beijing- and Morrisville-headquartered global technology company that is simultaneously the world's largest personal computer vendor by shipments and, since FY2025/26, one of the world's three largest server vendors by revenue. It is the operating heart of a corporate lineage that began in 1984 inside the Chinese Academy of Sciences and now spans 180 markets, more than 30 manufacturing sites in eleven markets, and a workforce of roughly 73,500. In the fiscal year ended March 31, 2026 the Group crossed US$80 billion of revenue for the first time and, in the June 2026 quarter, posted the largest quarterly revenue and the highest adjusted quarterly profit in its history. --- ### What the company does Lenovo designs, develops, manufactures, markets, sells and services a full-stack portfolio of computing products. In the Group's own words from the FY2025/26 corporate description, Lenovo is "a US$83 billion revenue global technology powerhouse … serving millions of customers every day in 180 markets," guided by a vision of "Smarter Technology for All" and "executing a Hybrid AI strategy that spans Personal AI – one personal AI, multiple devices; and Enterprise AI – helping customers turn data into insights and value." The stated delivery mechanism is "a full-stack AI portfolio, including devices (PCs, workstations, smartphones, tablets, accessories), infrastructure solutions (server, storage, edge, high performance computing and software defined infrastructure), as well as software, solutions, and services." Independent characterization: Lenovo is best understood as three businesses of very different economic character bolted onto a single, exceptionally efficient global supply chain and go-to-market apparatus. The first business, Intelligent Devices Group, is a scale-and-mix machine. It generated US$58.9 billion of FY2025/26 revenue at a 7.2 percent segment operating margin. Its competitive advantage is not technology in the narrow sense but cost position, channel breadth, and a manufacturing model — in-house plants plus "ODM+" hybrid outsourcing — that lets Lenovo flex capacity and geography faster than peers. Margin here is won through premium mix (premium PCs reached 50 percent of Q4 shipments), adjacency attach (tablets, monitors, accessories, smartphones), and component procurement leverage. The second business, Infrastructure Solutions Group, is a capital-intensive, cyclical, contract-driven business that until FY2025/26 destroyed operating profit. It reached US$19.2 billion of revenue and US$73 million of full-year operating profit in FY2025/26 — a genuine inflection — then in the June 2026 quarter delivered US$8.5 billion of revenue and US$777 million of operating profit at a 9.1 percent margin. The economics have two very different halves: hyperscaler and cloud-service-provider business, which is high volume and structurally low margin, and Enterprise/SMB, which carries far richer economics. The Infinidat acquisition (completed April 2026) is an explicit attempt to shift the mix toward the latter. The third business, Solutions and Services Group, is the profit-quality engine. On US$10.0 billion of FY2025/26 revenue it produced US$2.2 billion of segment operating profit — a 22.4 percent margin in the fourth quarter, and a business whose operating profit has more than doubled over five years at a 21 percent CAGR. Importantly, 62 percent of SSG revenue now comes from Managed Services and Project & Solutions rather than warranty-attached support, which is the shift that makes the margin durable. ### Revenue model and mix | Revenue by recognition timing (USD M) | FY2025 | FY2026 | |---|---|---| | Point in time (product sale) | 66029 | 79205 | | Over time (services, subscriptions) | 2779 | 3433 | | Lease revenue (TruScale, DaaS leasing) | 269 | 437 | | Total | 69077 | 83075 | Lenovo's revenue recognition disclosure is the cleanest window into the business model: Point-in-time product sales remain 95.3 percent of revenue in FY2025/26. Recurring and consumption-based revenue — over-time plus lease — is 4.7 percent, up from 4.4 percent, and growing faster than the base (over-time revenue +23.5 percent; lease revenue +62.2 percent). The strategic aspiration is to grow the recurring share; the financial reality is that Lenovo remains overwhelmingly a hardware transaction business. Deferred revenue on the balance sheet, a proxy for contracted future services, stood at US$3,957 million at March 31, 2026 (current US$1,988 million plus non-current US$1,969 million), up from US$3,194 million a year earlier. There is no meaningful licensing revenue stream disclosed. Lenovo is a net payer, not a net receiver, in standard-essential patent licensing. ### Value chain position and customer types Lenovo sits at the systems-integration and brand layer of the computing value chain. It does not fabricate semiconductors, memory, or displays; it is a very large buyer of them, and its FY2025/26 gross margin compression (16.1 percent to 15.4 percent) is direct evidence of that dependency during the DRAM and NAND shortage. Upstream it depends on Intel, AMD, Qualcomm, NVIDIA, Samsung, SK hynix, Micron, and Taiwanese ODMs. Downstream it reaches customers through a hybrid of direct enterprise sales, a very large two-tier distribution and reseller channel, telecom carriers (for Motorola), retail, and e-commerce. Customer types span, in rough order of revenue contribution: consumers and small business through retail and e-commerce; commercial mid-market and enterprise IT buyers; public sector and education; cloud service providers and hyperscalers (ISG's CSP business); telecommunications carriers; and, increasingly, "neocloud" and sovereign AI buyers. End-markets served include general commercial IT, education, healthcare, financial services, retail, manufacturing, telecommunications, public sector, media and entertainment, sports, and scientific/high-performance computing. ### The Global/Local operating model Management's most-repeated differentiator is what it calls the "Global/Local" or "global resources, local delivery" model, and the associated "China+N" manufacturing footprint. The FY2024/25 results announcement described "an end-to-end integrated global operations (design, demand forecasting, procurement, manufacturing, marketing, sales, and services), ODM+ manufacturing model, and global resources/local delivery model," supported by "30+ manufacturing sites (either in-house or outsourced) in 11 different markets." The FY2025/26 announcement attributed the Group's resilience through tariffs and component inflation directly to this footprint. This is a real and quantifiable advantage: Lenovo ranked fifth in the Gartner Supply Chain Top 25 for 2026, its highest-ever placement, after eighth in 2025, tenth in 2024 and eighth in 2023.

Deep Intelligence Across Industries

ADTRAN, Inc.

ADVA Optical Networking

Advanced Micro Devices, Inc.

What our clients say

Aten International Co., Ltd.

Zytronic PLC

ROHM Company Ltd

ST Microelectronics NV

Seqirus

Cypress Semiconductor Corporation

ALLIED TELESIS, INC.

HMD Global

Kalray Corporation

CHIESI Farmaceutici SpA

Beximco Pharmaceuticals Ltd

Boehringer Ingelheim International

Cipla Inc.

ASICS CORPORATION

Fujitsu Ltd

UNDER ARMOUR, INC.

LENOVO GROUP LIMITED

TomTom International BV

Fresenius

Booz Allen Hamilton Inc.

DXC Technology Company

Iron Mountain, Inc

News Corporation

Micro Focus

FireEye, Inc.

Hitachi Systems Security Inc

Symantec Corporation