NVIDIA posted 65% fiscal 2026 revenue growth and continues to dominate AI infrastructure, but valuation and margin pressure call for disciplined entry points. The report supports a Buy rating with strong execution and a rich but still defensible setup.
NVIDIA Corporation (NVDA) remains one of the strongest growth businesses in the public markets, but the stock demands disciplined entry points. Fiscal 2026 revenue reached $215.9B, up 65% year over year, while the latest reported quarter produced $96.2B of revenue, $2.46 of GAAP EPS, a 75.0% gross margin, and 106% year-over-year revenue growth. Management also guided to approximately 70% revenue growth in fiscal 2028. That combination supports a Buy rating for a moderate-risk investor with a medium-term horizon, provided valuation and position size remain controlled.
The central strength is NVIDIA's move from selling individual GPUs to selling a full AI infrastructure stack that combines accelerators, CPUs, networking, rack-scale systems, CUDA software, and domain-specific tools. The central risk is that the company has become heavily dependent on a massive AI infrastructure buildout while gross margins are scheduled to fall from 75.0% in the latest quarter to 71% to 72% in fiscal Q4 because of memory costs.
The investment case therefore rests on execution rather than simple enthusiasm. NVIDIA has beaten EPS estimates in all eight quarters listed, carries $51.1B of net cash in the latest core balance-sheet snapshot, and has a PEG ratio of 0.6. Against that, the stock carries a trailing P/E of 29.2, a forward P/E of 25.4, an EV-to-revenue multiple of 18.1, and a free-cash-flow yield of 2.0%. The business is exceptional; the price still requires respect.
Founded in 1993 and headquartered in Santa Clara, California, NVIDIA employs approximately 42,000 people and trades on the NASDAQ under NVDA. The company reports two operating segments: Compute & Networking, which includes data center accelerated computing, networking, AI software, and automotive platforms, and Graphics, which includes GeForce gaming products and professional workstation graphics.
The company's strategy is broader than chip design. NVIDIA's 2026 10-K describes a platform that combines hardware, systems, networking, software, algorithms, libraries, application programming interfaces, and services. That structure lets one programmable architecture serve data center, gaming, professional visualization, and automotive markets. In plain English, NVIDIA is trying to become the operating layer for accelerated computing, not merely the premium chip supplier.
Data Center is the dominant earnings engine. Revenue increased from $115.2B in fiscal 2025 to $193.7B in fiscal 2026, a 68% increase. The latest quarter added $96.2B of company-wide revenue, and management described demand across hyperscalers, AI labs, AI-native companies, enterprises, and sovereign customers. That customer expansion matters because it reduces dependence on one type of buyer even as the largest cloud providers remain essential.
Gaming remains a substantial secondary business. Fiscal 2026 Gaming revenue rose 41% to $16.0B. Blackwell-based GeForce RTX 50 Series products add neural graphics, ray tracing, and DLSS capabilities, giving NVIDIA a way to monetize both traditional gaming demand and local AI workloads on PCs. The smaller scale of Gaming relative to Data Center limits its effect on the total company, but it provides a separate consumer platform and a broad developer base.
Professional Visualization revenue increased 70% to $3.2B in fiscal 2026, while Automotive revenue rose 39% to $2.3B. NVIDIA's automotive opportunity includes DRIVE hardware, operating software, simulation, mapping, and autonomous-driving tools. These businesses are too small to offset a material data center slowdown today, but their growth shows how the CUDA and accelerated-computing platform can extend into specialized markets.
Get AI research on any stock
Instant reports, daily intelligence, and an AI analyst in your pocket.
Blackwell is the current commercial foundation of NVIDIA's AI systems. The company launched Blackwell in fiscal 2025 and scaled Blackwell Ultra in fiscal 2026. The investor presentation attributed the latest year-over-year gross-margin improvement partly to a better product mix from Blackwell Ultra, while management said sustained Blackwell demand drove hyperscale growth.
Vera Rubin is the next major product cycle. Production shipments began in August 2026, and management said every major hyperscaler, AI cloud, and system OEM had placed purchase orders. NVIDIA expects Vera Rubin to account for approximately 20% of Data Center revenue in fiscal Q3 2027 and described it as the fastest product ramp in company history. The platform combines Vera CPUs, Rubin GPUs, NVLink, InfiniBand or Ethernet, and Groq technology.
The product argument is not just faster silicon. NVIDIA says its revenue opportunity per gigawatt rose from roughly $18B with Hopper to $25B with Blackwell and $40B with Vera Rubin. If customers can reuse one system across data preparation, training, post-training, and agentic inference, the platform can retain economic value for longer than a single workload cycle.
CUDA remains NVIDIA's most durable competitive asset. The 10-K cites more than 7.5 million developers using CUDA and related tools. More than half of NVIDIA's engineers work on software, and the company supports approximately 6,000 applications. That installed base raises switching costs because customers have built code, workflows, and employee expertise around NVIDIA's libraries and development environment.
NVIDIA has also built an unusually broad systems moat. Its platform now spans GPUs, CPUs, DPUs, NVLink, InfiniBand, Ethernet, rack-scale systems, AI models, and enterprise software. The 10-K says NVIDIA infrastructure powers more than 78% of the world's TOP500 supercomputers, including nine of the top ten Green500 systems. Those figures support the view that NVIDIA's advantage is measured across the stack, not only by benchmark scores for an individual accelerator.
Research spending has compounded that position. NVIDIA says it has invested more than $76.7B in research and development since inception, with major milestones including CUDA in 2006, Tensor Cores in 2017, Mellanox networking in 2020, Grace CPUs, Blackwell, and Vera Rubin. The cadence creates a powerful commercial loop: a large developer ecosystem supports product adoption, product adoption funds research, and new architectures reinforce the ecosystem.
NVIDIA uses a fabless manufacturing model. TSMC and Samsung produce semiconductor wafers, while SK Hynix, Micron, and Samsung supply memory. CoWoS advanced packaging supports the company's large AI systems, and Hon Hai, Wistron, and Fabrinet participate in assembly, testing, and packaging. This model limits the fixed-cost burden of owning fabs but leaves NVIDIA exposed to foundry, packaging, memory, and logistics capacity.
Management described supply as the binding constraint on the fiscal 2028 outlook. NVIDIA expects supply to remain a bottleneck through at least the end of fiscal 2028, while memory prices are rising sharply. Inventory increased to $32B in preparation for the Vera Rubin launch, and days sales outstanding rose to 60 days because certain investment-grade customers received extended payment terms for purchases shipping over multiple quarters.
The near-term margin tradeoff is explicit. NVIDIA expects fiscal Q3 GAAP and non-GAAP gross margin of 74.0% plus or minus 50 basis points, followed by a 71% to 72% range in fiscal Q4 and 72% to 73% in fiscal 2028. The decline reflects memory costs, although management said price increases are scheduled to take effect in fiscal Q1. Supply constraints are therefore both a growth limiter and a margin risk.
The semiconductor market is expanding, but AI is driving a disproportionate share of the growth. Gartner forecasts worldwide semiconductor revenue of $717B in 2025, $909.8B in 2026, and more than $1T in 2028. Gartner also forecasts GPUs and AI accelerators rising from $80B in 2024 to more than $280B by 2029. These figures provide a favorable industry backdrop for NVIDIA's accelerator and networking portfolio.
NVIDIA is targeting a larger market than conventional semiconductor revenue. Its investor materials frame the AI infrastructure opportunity at $3T to $4T by 2030, while management says each gigawatt of Vera Rubin infrastructure represents a $40B revenue opportunity for NVIDIA's platform. That framing includes systems, networking, CPUs, software, and services, which explains why NVIDIA's revenue growth has outpaced the broader semiconductor market.
Power, cooling, advanced packaging, and high-bandwidth memory are becoming as important as chip architecture. Management said customers must secure land, power, and data-center shells years in advance. The industry is consequently moving toward integrated AI factories, where the value of a complete platform can exceed the value of an individual GPU.
Like what you're reading?
Get full access to AI-powered research reports, market analysis, and portfolio tools.
NVIDIA serves major cloud providers, AI model makers, enterprises, startups, public-sector organizations, system vendors, original equipment manufacturers, distributors, and automotive companies. The latest management commentary divided demand between hyperscalers and ACIE, which includes AI clouds, industrial customers, enterprises, sovereign customers, and edge deployments. That mix gives the company exposure to both large infrastructure budgets and a growing long tail of specialized users.
The spending pipeline remains large. Management cited cloud-industry backlog above $2T and projected capital expenditures of nearly $800B from the five largest hyperscalers in 2026 and $1.3T in 2027. NVIDIA also said AI venture funding exceeded $400B in the first half of 2026, with roughly 70% directed toward compute. Those figures support continued demand for training and inference infrastructure, although the concentration of spending among large customers creates bargaining and timing risk.
The customer base is also expanding geographically. NVIDIA described sovereign AI revenue as growing 35% sequentially and more than tripling year over year in the latest quarter. Partnerships with regional cloud providers in Armenia, Africa, Taiwan, India, Australia, Malaysia, Japan, South Korea, and Europe show how the company is using local infrastructure partners to extend its reach beyond the largest U.S. cloud platforms.
AMD is NVIDIA's clearest direct competitor in data-center accelerators and discrete graphics. Intel competes in CPUs and AI accelerators, while Broadcom participates in custom silicon and networking. NVIDIA's 10-K also identifies Huawei, Qualcomm, Marvell, Cisco, Arista Networks, and several other companies across processors, networking, automotive systems, and embedded products.
The more important competitive threat may come from customers themselves. Amazon, Alphabet, Microsoft, Meta, and other large cloud companies are designing internal chips to optimize cost, supply, and workload performance. Custom silicon can pressure NVIDIA's pricing over time, particularly for predictable inference workloads. NVIDIA's response is to make the full platform more valuable than an isolated accelerator through CUDA, networking, system design, and software.
NVIDIA still holds the strongest position in the supplied competitive evidence. AMD's filing identifies NVIDIA as the discrete GPU market-share leader, while NVIDIA's 10-K cites broad deployment across cloud providers, supercomputers, enterprises, and AI model makers. The moat is substantial, but its durability depends on maintaining software compatibility, product cadence, supply volume, and a clear total-cost advantage.
Export controls are the most direct geopolitical risk. NVIDIA's fiscal 2026 filing records a $4.5B charge tied to H20 excess inventory and purchase obligations after U.S. restrictions reduced demand. The latest fiscal Q3 outlook assumes no Data Center compute revenue from China, and management said less than 1% of total Data Center revenue in the latest quarter came from Hopper 200 shipments to China under U.S. licenses.
The 10-K also describes a 25% tariff on H200 products imported into the United States under a specific licensing program. Restrictions can reduce NVIDIA's addressable market, encourage Chinese competitors to build stronger ecosystems, and increase product-transition costs. China is therefore a two-sided issue: lost near-term revenue and a potential long-term competitive threat.
Power availability and infrastructure construction are another macro constraint. Management said AI customers must coordinate land, power, cooling, labor, and data-center shells, sometimes two to three years ahead. That supports the durability of the infrastructure cycle, but it also means project delays, electricity constraints, and financing conditions can affect the pace at which booked demand becomes recognized revenue.
$51.1B of net cash gives NVIDIA a strong cushion, even as the report flags valuation discipline and a business increasingly tied to AI infrastructure spending.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Fiscal 2026 revenue jumped to $215.9B and the latest quarter delivered $96.2B of sales with a 75.0% gross margin, underscoring exceptional operating momentum.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Management is guiding to roughly 70% revenue growth in fiscal 2028, while the company has beaten EPS estimates in all eight quarters listed.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →A 29.2 trailing P/E, 25.4 forward P/E, 18.1 EV/revenue multiple, and 2.0% free-cash-flow yield show a premium stock that still needs a careful entry point.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →The report’s Buy call is anchored by strong AI growth and execution, but the $327.13 fair value leaves room for valuation risk if margin pressure or demand cools.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →NVIDIA's investment profile is unusually clear. The company owns the leading accelerated-computing ecosystem, has converted CUDA into a major switching-cost advantage, and is expanding from GPUs into CPUs, networking, systems, and AI software. Fiscal 2026 revenue of $215.9B, latest-quarter revenue of $96.2B, and a 55.6% annual net margin show that the opportunity is already producing extraordinary financial results rather than existing only as a technology narrative.
The prudent stance is Buy, not speculation without limits. Supply constraints, memory inflation, China restrictions, customer concentration, custom silicon, and a 25.4x forward P/E create real risk. For a medium-term investor, staged exposure around the Buy and Strong Buy levels offers a better balance than chasing every product announcement. NVIDIA has the business quality to lead the AI infrastructure cycle, but even the best company cannot repeal valuation.
Get AI-powered research reports, daily market intelligence, and a personal analyst in your pocket.
Get Full Access →Not ready to subscribe? ·
Cancel anytime
NVIDIA Corporation (NVDA) rises after fiscal Q2 2027 results as investors focus on stronger AI infrastructure demand and upbeat forward guidance. Despite an EPS miss, hyperscale revenue surged and the market is rewarding the company’s longer growth runway, though valuation and volatility remain elevated.
Nvidia's Aug. 27 report is not merely a beat-or-miss event: it is a test of whether the customers powering AI can keep financing the buildout. The roughly $194 billion gap between projected hyperscaler capex growth and added operating cash flow makes demand quality more important than another headline surprise.
NVIDIA Corporation (NVDA) drops sharply after a Reuters report said it will invest $1 billion in newly issued Naver shares tied to a South Korea data center partnership. The move comes amid broader semiconductor weakness, even as NVIDIA’s revenue growth and earnings streak remain strong.