0001045810-26-000021
SEC filingRevenue surged 65% to $215.9B driven by AI demand, but gross margin fell 3.9 pts due to product transition and H20 charge.
NVIDIA Corporation is a data center scale AI infrastructure company that pioneered accelerated computing. Its full-stack platform, anchored by the CUDA development platform, spans hardware (GPUs, CPUs, DPUs, networking), systems, software, and algorithms. The company serves four large markets: Data Center, Gaming, Professional Visualization, and Automotive. Over 7.5 million developers use CUDA, and NVIDIA powers over 78% of TOP500 supercomputers.
NVIDIA reports in two segments: Compute & Networking and Graphics. Compute & Networking includes Data Center accelerated computing and networking platforms, AI software, and Automotive platforms. Graphics includes GeForce GPUs for gaming and Quadro/NVIDIA RTX GPUs for professional workstations. Revenue share by segment is not disclosed in this section.
Key products include the Blackwell architecture (launched fiscal 2025) and Blackwell Ultra (fiscal 2026) for data center-scale AI, the Rubin platform (expected fiscal 2027), GeForce RTX 50 Series GPUs for gaming, and DRIVE Hyperion for autonomous vehicles. Notable platforms include CUDA, NVIDIA AI Enterprise (including NIM, NeMo, AI Blueprints), Omniverse for physical AI, and Clara for healthcare. Networking offerings include NVLink, InfiniBand, Ethernet switches, DPUs, and NVLink Fusion.
NVIDIA sells directly and through a partner network of CSPs, OEMs, ODMs, ISVs, system integrators, and distributors. The sales team includes application engineers and solution architects who provide pre-sales assistance. Customers include all major cloud providers, AI model makers, enterprises, startups, and public sector entities. No single customer concentration is disclosed.
The market is intensely competitive. Named competitors include AMD, Intel, Huawei, Amazon, Microsoft, Google, Baidu, Qualcomm, Broadcom, Samsung, Tesla, and others. Competition spans GPUs, CPUs, DPUs, embedded SoCs, and networking products. Large cloud companies with internal hardware teams are also competitors. NVIDIA notes that export controls have harmed its competitive position, particularly in China, by benefiting competitors.
NVIDIA's five key strategies: (1) advancing the accelerated computing platform with full-stack innovation; (2) extending AI leadership with end-to-end solutions and ecosystem; (3) extending computer graphics leadership with AI-infused rendering; (4) advancing the autonomous vehicle platform (DRIVE Hyperion); and (5) leveraging intellectual property through licensing. The company emphasizes unified architecture across diverse markets.
As of fiscal 2026, NVIDIA had approximately 42,000 employees: 31,000 in R&D and 11,000 in sales, marketing, operations, and administration. Over 80% hold technical roles, and more than half have advanced degrees. Turnover rate was 3.7% in fiscal 2026. The company invests in employee development and aligns compensation with shareholder interests through equity participation.
Fiscal year 2026 was a landmark period for NVIDIA, with revenue soaring 65% to $215.9 billion, driven overwhelmingly by demand for accelerated computing and AI solutions. The Data Center segment, the primary growth engine, saw revenue increase 68% year-over-year, as Blackwell architectures became the majority of Data Center revenue. Despite the top-line strength, gross margin contracted 390 basis points to 71.1% from 75.0% in fiscal 2025. The decline was attributable to two key factors: the business model shift from Hopper HGX systems to Blackwell full-scale datacenter solutions, which carry different cost structures, and a $4.5 billion charge in Q1 FY26 for H20 excess inventory and purchase obligations related to export license restrictions to China. Operating expenses rose 41% to $23.1 billion, reflecting continued investment in headcount and compute infrastructure. Net income improved 65% to $120.1 billion, and diluted EPS grew to $4.90 from $2.94, a 67% increase.
NVIDIA's two operating segments exhibited robust growth. Compute & Networking revenue increased 67% to $193.5 billion, with operating income up 57% to $130.1 billion. Within this segment, Data Center computing grew 59% and Data Center networking surged 142%, fueled by the ramp of NVLink compute fabric for GB200/GB300 systems and expansion of Ethernet and InfiniBand platforms. The Graphics segment revenue rose 57% to $22.5 billion, driven by strong Blackwell demand across gaming, professional visualization (up 70%), and automotive (up 39%). Graphics operating income jumped 80% to $9.2 billion. Notably, a $4.5 billion charge in Compute & Networking partially offset segment profitability, while Graphics benefited from higher revenue without such charges.
Management highlighted both opportunities and headwinds. The company continues to execute a one-year product cadence, with the Rubin platform on the horizon and Blackwell Ultra (GB300) already shipping. However, complexity in product transitions may cause supply-demand mismatches and revenue volatility. Gaming faces near-term supply constraints in Q1 FY27. Geopolitical risks persist, including U.S. export controls on AI chips to China; recent licenses have allowed limited H20 and H200 shipments, but tariffs and regulatory uncertainties remain. NVIDIA is expanding its supply chain into the U.S. and Latin America to add resilience. Capital expenditures are expected to increase in fiscal 2027 to support future growth. The company also continues to invest heavily in the AI ecosystem, with $17.5 billion in private company investments and $3.5 billion in guarantees for datacenter infrastructure in fiscal 2026. Overall, the MD&A portrays a company riding the AI wave while navigating transitional costs and geopolitical complexities.
NVIDIA faces unprecedented export control risks. The USG imposed a series of restrictions on GPUs and AI semiconductors, culminating in a $4.5 billion charge in Q1 FY2026 for H20 excess inventory and purchase obligations. The company is effectively foreclosed from China's data center compute market, and worldwide licensing rules (AI Diffusion IFR) were proposed but rescinded; a replacement rule is expected. The GAIN AI Act in the NDAA could further constrain NVIDIA's export flexibility. Additionally, Chinese antitrust regulators preliminarily found NVIDIA violated terms of the Mellanox acquisition by complying with US export controls, risking penalties or operational restrictions. Competition regulators in the EU, US, UK, and South Korea are investigating NVIDIA's GPU sales, supply allocation, and AI partnerships. AI regulations such as the EU AI Act and multiple US state laws impose compliance costs and may limit deployment of AI models.
NVIDIA depends on long lead times (over 12 months) and third-party foundries in Taiwan and South Korea. Supply constraints are expected to impact Gaming in Q1 FY2027. Product transitions (e.g., Blackwell) have caused low yields and inventory provisions. The company is increasing US-based manufacturing but faces scaling challenges. Geopolitical tensions, climate change (water/energy shortages), and natural disasters threaten operations. Approximately 6,000 employees in Israel are affected by regional conflict.
Customer concentration is high: two direct customers account for 22% and 14% of revenue. Indirect customers also represent significant portions. Competitors with in-house fabs (e.g., Intel, Samsung) or government backing (e.g., Chinese firms) may gain market share, especially as export controls drive customers to seek alternatives. Some customers are developing internal solutions (e.g., cloud providers).
Financial risks include operating result volatility due to demand fluctuations, extended payment terms, and macroeconomic conditions (inflation, tariffs, interest rates). Cybersecurity threats are increasingly sophisticated; data privacy regulations (GDPR, CCPA, China data laws) carry fines up to 4% of revenue. The company faces securities class-action litigation and could be subject to IP infringement claims. Tax risks include changes to U.S. FDDEI deduction and OECD Pillar Two implementation.
No cash flow data available to analyze.