NVIDIA vs AMD Stock in 2026: AI Growth, Profitability, Valuation and the Real Trade-Off
NVIDIA and AMD are both central to the AI infrastructure buildout, but the investment cases are no longer mirror images.
NVIDIA enters the comparison with vastly greater revenue, data-center scale, margins and free cash flow. AMD enters as the smaller challenger, with rapidly expanding data-center revenue, a new rack-scale AI architecture and a software strategy built around ROCm and open standards. The core question for investors is not simply which company has the faster GPU. It is how much future execution is already embedded in each stock price.
Research cutoff: October 2, 2026. Market prices in this article use the October 1, 2026 U.S. closing prices because the October 2 regular session had not opened at the research cutoff. NVIDIA financial comparisons use fiscal Q2 2027, ended July 26, 2026. AMD comparisons use calendar Q2 2026, ended June 27, 2026. The periods are close in time but are not identical.
Methodology: Reported financial data comes primarily from the companies' official earnings releases. Market-price and valuation-ratio references are separately sourced. BTI fair-value outputs are proprietary analytical estimates, not company guidance, Wall Street consensus or guaranteed future prices.
Executive Summary: NVIDIA vs AMD Stock in 2026
The latest reported numbers show two companies participating in the same AI infrastructure cycle from very different positions. NVIDIA reported $96.221 billion of quarterly revenue, while AMD reported $11.536 billion. NVIDIA's Data Center revenue alone reached $89.0 billion, compared with AMD Data Center segment revenue of $6.718 billion. On those reported quarters, NVIDIA generated more than eight times AMD's total revenue and more than thirteen times AMD's Data Center revenue.
Scale is only one part of the comparison. NVIDIA reported a 75.0% GAAP gross margin and $63.734 billion of GAAP operating income. AMD reported a 54% GAAP gross margin and $1.990 billion of GAAP operating income. Free cash flow was $21.341 billion for NVIDIA and $1.558 billion for AMD. These figures make NVIDIA the much stronger current earnings and cash-generation platform.
AMD's counterargument is growth from a smaller base. Its Q2 revenue increased 50% year over year, while Data Center revenue increased 107%. AMD is also moving beyond individual accelerators toward rack-scale systems through Helios and the Instinct MI400 family. That creates a credible path to participate more deeply in large AI clusters, but it also raises the execution bar because AMD must expand hardware adoption, networking integration, software maturity and customer deployments at the same time.
The valuation contrast is equally important. At the October 1 close, NVIDIA traded at $230.86 with a market capitalization of about $5.57 trillion and a trailing GAAP P/E of 29.19. AMD closed at $615.73 with a market capitalization of about $1.01 trillion and a trailing GAAP P/E of roughly 157.16. A P/E ratio cannot value these businesses by itself, but the gap shows that the market is currently assigning AMD a much larger premium to reported earnings.
Key takeaways
- Scale: NVIDIA's latest quarterly revenue was about 8.3 times AMD's.
- AI infrastructure: NVIDIA Data Center revenue was about 13.2 times AMD Data Center segment revenue in the latest reported quarters.
- Profitability: NVIDIA's 75% GAAP gross margin materially exceeded AMD's 54%.
- Cash generation: NVIDIA generated $21.341 billion of quarterly free cash flow versus AMD's $1.558 billion.
- Growth optionality: AMD's Data Center segment grew 107% year over year, showing that the challenger case is not theoretical.
- Valuation: The October 1 trailing GAAP P/E was about 29.2 for NVIDIA and 157.2 for AMD, implying very different expectation levels.
- BTI model signal: BTI's NVIDIA models are mixed, while BTI's proprietary AMD market-cap model is materially below the current market capitalization. These are model outputs, not facts about future prices.
Latest Reported Quarterly Revenue: NVIDIA vs AMD
96.221 NVIDIA Q2 FY27
11.536 AMD Q2 2026
Source: NVIDIA and AMD official earnings releases
USD billions
Latest reported quarterly revenue. NVIDIA fiscal Q2 2027 and AMD Q2 2026. Source: company earnings releases. BTI original visualization.
This Is No Longer Just a GPU Comparison
A consumer comparison often starts with frame rates, power draw and price-to-performance. An investor comparison has to start several layers higher. NVIDIA increasingly sells a complete AI infrastructure stack: accelerators, CPUs, networking, rack-scale systems, software libraries, enterprise software and deployment architecture. AMD also spans CPUs, GPUs, adaptive computing and data-center products, but its AI accelerator platform remains in the share-capture phase.
That distinction matters because AI economics are increasingly determined at the rack and cluster level rather than by one chip in isolation. Training and inference performance depends on compute, memory bandwidth, interconnects, networking, software optimization, power efficiency and system utilization. A supplier that controls more of those layers can capture more economics, but it also assumes more supply-chain complexity and capital commitments.
For investors, the comparison can be reduced to two different questions. With NVIDIA, the question is whether extraordinary scale and margins can remain durable as AI infrastructure spending expands. With AMD, the question is whether strong product execution can translate into enough incremental market share and earnings to support the valuation already assigned to the stock.
Financial Scale: NVIDIA Is Operating in a Different Revenue Class
NVIDIA reported fiscal Q2 2027 revenue of $96.221 billion, up 106% year over year. AMD reported Q2 2026 revenue of $11.536 billion, up 50% year over year. NVIDIA therefore produced roughly 8.3 times AMD's quarterly revenue in the latest reported periods.
The operating-income gap was even wider. NVIDIA generated $63.734 billion of GAAP operating income, while AMD generated $1.990 billion. NVIDIA's GAAP net income was $59.688 billion; AMD's was $2.297 billion.
These figures do not mean AMD lacks operating momentum. AMD's GAAP operating income increased sharply from a loss in the comparable prior-year quarter, and non-GAAP operating margin reached 27%. The important distinction is that NVIDIA is already monetizing the AI cycle at a scale AMD has not yet reached.
NVIDIA official Q2 FY2027 earnings release · AMD official Q2 2026 earnings release
Data Center: The Most Important Gap in the Comparison
NVIDIA's Data Center revenue reached $89.0 billion in fiscal Q2 2027, up 117% year over year. That represented roughly 92% of NVIDIA's total quarterly revenue. AMD's Data Center segment generated $6.718 billion, up 107% year over year, and represented about 58% of AMD's quarterly revenue.
The comparison needs an important caveat: the companies do not define their data-center businesses identically. AMD's segment includes EPYC server CPUs, Instinct accelerators, GPUs, DPUs, AI network interface cards, FPGAs and adaptive products. NVIDIA's Data Center business includes accelerated computing and networking products. Therefore, the revenue numbers are useful for scale but should not be interpreted as a pure GPU-versus-GPU market-share calculation.
Even with that caveat, the gap is strategically meaningful. NVIDIA is already operating at AI-infrastructure scale that supports broad software investment, networking development, supply-chain commitments and rapid platform transitions. AMD has the advantage of a much smaller base: each large hyperscaler or frontier-model deployment can have a more visible percentage impact on AMD's revenue.
Latest Reported Data Center Revenue
89 NVIDIA Data Center
6.718 AMD Data Center Segment
Source: NVIDIA and AMD official earnings releases; segment definitions differ
USD billions
Latest reported Data Center revenue. Segment definitions differ between the companies. Source: NVIDIA and AMD earnings releases. BTI original visualization.
Profitability: Revenue Growth Is Not Equal When the Margins Are Different
NVIDIA reported a 75.0% GAAP gross margin in fiscal Q2 2027. AMD reported a 54% GAAP gross margin in Q2 2026. The 21-percentage-point gap helps explain why NVIDIA converts each additional dollar of revenue into operating profit at a far higher rate today.
NVIDIA's GAAP operating margin was approximately 66% based on reported operating income and revenue. AMD's reported GAAP operating margin was 17%. AMD's non-GAAP operating margin was higher at 27%, but investors should avoid mixing non-GAAP AMD margins with GAAP NVIDIA margins when making a direct comparison.
The investment implication is not that AMD cannot close part of the profitability gap. Higher accelerator revenue, stronger data-center mix and greater software/platform scale could improve its economics. The point is that current valuation should be judged against current profitability and the amount of improvement required to reach future assumptions.
GAAP Gross Margin: Latest Reported Quarter
75 NVIDIA
54 AMD
Source: NVIDIA and AMD official earnings releases
percent
GAAP gross margin comparison from the latest reported quarters. Source: company earnings releases. BTI original visualization.
Free Cash Flow: NVIDIA Is Stronger, but Its Cash Conversion Deserves Attention
NVIDIA generated $21.341 billion of free cash flow in fiscal Q2 2027. AMD generated $1.558 billion in Q2 2026. On reported quarterly revenue, that is a free-cash-flow margin of roughly 22% for NVIDIA and 14% for AMD.
NVIDIA's number requires a second look because free cash flow fell from $48.554 billion in the prior quarter even as revenue and net income increased. The cash-flow statement shows substantial working-capital absorption: accounts receivable increased by $22.346 billion during the quarter, inventory consumed $5.784 billion and prepaid expenses and other assets consumed another $5.497 billion.
That does not automatically indicate deteriorating earnings quality. A company growing more than 100% year over year can consume significant working capital as receivables, inventory and supply commitments scale. But it changes what investors should monitor. If working-capital demands continue rising faster than operating cash generation, the gap between accounting earnings and cash conversion would become more important.
AMD's free cash flow also fell sequentially, from $2.566 billion in Q1 to $1.558 billion in Q2. AMD spent $808 million on capital expenditures during the quarter. Its balance sheet remained liquid, with $13.111 billion of cash, cash equivalents and short-term investments against $3.226 billion of total debt.
Quarterly Free Cash Flow
21.341 NVIDIA Q2 FY27
1.558 AMD Q2 2026
Source: NVIDIA and AMD official earnings releases; company definitions may differ
USD billions
Quarterly free cash flow in the latest reported periods. Company definitions of non-GAAP free cash flow may differ. BTI original visualization.
CUDA vs ROCm: Why Software Still Matters to the Stock Thesis
NVIDIA's most important competitive advantage is not simply that it sells fast accelerators. CUDA, CUDA-X libraries, networking, systems software and a large developer ecosystem make the platform difficult to evaluate as a commodity chip business. Customers building large AI systems care about time to deployment, model compatibility, orchestration, reliability and utilization as much as peak hardware specifications.
AMD's strategic response is ROCm. The company positions the stack as an open-source AI and HPC foundation and is building compatibility across widely used frameworks. The MI400 platform is also designed around open standards including OCP, UALink and UEC. That approach can be attractive to customers that want more supplier choice or less dependence on a proprietary infrastructure stack.
The key investor question is not whether ROCm can technically run modern AI workloads. It can. The more important question is whether software maturity, deployment tooling and ecosystem support become strong enough that large customers view AMD as a repeatable default platform rather than a negotiated alternative for selected workloads.
Vera Rubin vs Instinct MI400 and Helios: The Rack Is Becoming the Product
NVIDIA's Vera Rubin platform is already ramping into full production. The platform integrates Rubin GPUs, Vera CPUs, NVLink, ConnectX networking, BlueField infrastructure processing and Spectrum networking. NVIDIA is explicitly designing around AI factories and multi-rack systems rather than selling an isolated accelerator.
AMD is pursuing the same strategic direction through the Instinct MI400 family and the Helios rack-scale reference design. AMD says Helios uses 72 MI455X GPUs and is built around ROCm and open rack standards. Volume deployments are expected in the second half of 2026. The important detail for investors is that Helios is a reference design, not a single AMD-branded server sold directly to every customer; OEM and ODM partners can build systems around the architecture.
This is where the competitive battle becomes more complex. NVIDIA's advantage is integration and an installed software ecosystem. AMD's opportunity is to offer credible rack-scale economics while giving customers more architectural openness. If hyperscalers want a genuine second source at scale, AMD does not need to displace NVIDIA everywhere to produce significant incremental revenue. But each layer added to the platform also creates another area where execution has to be proven.
NVIDIA Vera Rubin production announcement · AMD Instinct MI400 and Helios platform
Guidance Shows Both Companies Still Expect Strong AI Demand
NVIDIA guided fiscal Q3 2027 revenue to $108.0 billion, plus or minus 2%. Importantly, that outlook assumes no Data Center compute revenue from China. The company guided GAAP and non-GAAP gross margin to approximately 74.0%, plus or minus 50 basis points.
AMD guided Q3 2026 revenue to approximately $13.0 billion, plus or minus $300 million. The midpoint implies about 41% year-over-year growth and roughly 13% sequential growth. AMD guided non-GAAP gross margin to approximately 56%.
The guidance reinforces the same contrast seen in the reported results. NVIDIA is trying to sustain extraordinary growth on a much larger revenue base. AMD is attempting to accelerate from a smaller base while widening adoption of EPYC, Instinct and Helios. Both can grow at the same time because the addressable AI infrastructure market is expanding, but the stocks can still produce very different returns if valuation multiples move differently.
Valuation: The Market Is Pricing Two Very Different Expectation Bars
At the October 1, 2026 close, NVIDIA traded at $230.86, giving it a market capitalization of approximately $5.57 trillion. AMD closed at $615.73, with a market capitalization of approximately $1.01 trillion.
On trailing GAAP earnings, NVIDIA's P/E was approximately 29.19. AMD's was approximately 157.16. AMD's much higher multiple does not automatically make it overvalued; investors may be discounting rapid future earnings growth as Instinct and Helios scale. But it does mean the stock is more sensitive to any gap between expected growth and delivered growth.
NVIDIA presents the opposite valuation paradox. It has the larger market capitalization, but current earnings have expanded so rapidly that the trailing multiple is materially lower than AMD's. The risk is that today's extraordinary NVIDIA earnings represent a period of unusually high AI infrastructure spending and margins that cannot be extrapolated indefinitely.
Trailing GAAP P/E at October 1, 2026 Close
29.19 NVIDIA
157.16 AMD
Source: StockAnalysis market snapshot, October 1, 2026
times
Trailing GAAP P/E ratios as of the October 1, 2026 close. Source: StockAnalysis. Multiples change with price and earnings. BTI original visualization.
NVIDIA October 1 market-price reference · AMD October 1 market-price reference
BTI Valuation Models: Why the Signals Are Not Perfectly Consistent
BTI's proprietary NVIDIA valuation work does not produce a single clean answer. Three high-suitability models in the supplied BTI research snapshot produced fair-value estimates of $222.70, $196.43 and $222.70 per share. Against the $230.86 October 1 close, those outputs imply that NVIDIA was modestly to moderately above those model values.
However, a separate BTI market-cap fair-value framework indicated NVIDIA was approximately 14.79% below its estimated fair value. The disagreement is analytically useful. It shows that the conclusion is highly dependent on model architecture, normalized earnings, exit multiples and the treatment of future AI growth.
For AMD, the supplied BTI snapshot estimated a proprietary fair market capitalization of approximately $343.75 billion versus an October 1 market capitalization near $1.01 trillion. On that specific BTI framework, the market value is roughly 193% above the model estimate. That is not a statement that AMD's share price must fall. It is a statement that the model requires substantially more evidence before it can justify the market's current capitalization.
BTI NVIDIA model outputs versus the October 1 closing price. These are analytical estimates, not company guidance or Wall Street targets.
Investors can test their own assumptions with the BTI Advanced Valuation tool and review the broader framework through the BTI Stock Analysis Tool.
What Wall Street Is Pricing
As of the October 1 market snapshot, StockAnalysis listed an average NVIDIA analyst price target of approximately $327.70, compared with the $230.86 closing price. For AMD, the listed average target was approximately $618.51, close to the $615.73 closing price.
Price targets are not intrinsic-value facts. They are aggregated analyst estimates that can change rapidly after earnings, product launches or changes in AI infrastructure spending. Their usefulness in this comparison is therefore not as a forecast of where either stock will trade, but as a measure of how different the current expectation setup is: analyst targets imply much more percentage distance for NVIDIA than for AMD at the stated market prices.
Risk 1: AI Capital Spending Can Slow Even If AI Adoption Keeps Growing
Both companies are exposed to the same macro question: how long can hyperscalers, cloud providers, sovereign buyers and AI labs keep expanding compute infrastructure at current rates? AI usage can continue growing while infrastructure returns decline if utilization, token pricing or customer monetization disappoint.
NVIDIA has more absolute exposure because Data Center dominates its revenue. AMD has more valuation sensitivity because a large part of its current multiple depends on future earnings expansion. A broad AI-capex slowdown would therefore hurt both companies through different channels.
Risk 2: Export Controls and China Remain Material
NVIDIA's Q3 outlook explicitly assumes no Data Center compute revenue from China, showing that export restrictions remain embedded in near-term planning. AMD has also been affected by U.S. export controls; its prior-year comparison included an $800 million inventory and related charge associated with Instinct MI308 restrictions.
The relevant investor lesson is that China should not be treated as a simple upside option. Policy can change product eligibility, customer demand, inventory planning and the economics of products designed to comply with export restrictions.
Risk 3: NVIDIA Must Protect Cash Conversion While Scaling Supply
NVIDIA's earnings growth is exceptional, but the Q2 free-cash-flow decline shows that scale creates working-capital demands. Receivables and inventory can rise rapidly when customers, systems partners and supply chains expand simultaneously. The risk is not the existence of working capital itself; it is whether cash conversion weakens persistently even after revenue growth normalizes.
Risk 4: AMD Must Prove That Product Momentum Becomes Durable Platform Adoption
AMD's Data Center growth demonstrates real demand. The harder test is durability. Investors should watch whether MI400 and Helios deployments broaden beyond selected strategic customers, whether ROCm reduces software friction, whether gross margins expand with AI mix and whether the company can scale accelerator revenue without sacrificing pricing discipline.
Because AMD's trailing GAAP P/E is far above NVIDIA's at the current snapshot, execution misses can have an amplified valuation impact. The same mechanism can work in the other direction if earnings compound faster than current expectations.
What Could Change the NVIDIA Thesis?
- Vera Rubin ramps faster or slower than expected across major cloud and AI customers.
- Data Center growth remains above expectations or begins normalizing more quickly.
- Free cash flow reconverges toward net income as working-capital growth moderates, or the cash-conversion gap persists.
- Gross margins hold near the mid-70% range or compress as system mix, memory costs and competition change.
- China restrictions become more or less restrictive than currently assumed.
- Alternative accelerators, custom silicon or competing rack-scale systems reduce NVIDIA's pricing power.
What Could Change the AMD Thesis?
- MI400 and Helios move from announced deployments into broad production revenue.
- ROCm adoption accelerates among major model developers and cloud platforms.
- Data Center revenue maintains triple-digit or very high growth as the comparison base becomes larger.
- Gross and operating margins expand enough to close part of the profitability gap with NVIDIA.
- Large customer wins translate into repeat deployments rather than one-time strategic programs.
- The earnings growth required by the current valuation fails to materialize, creating multiple-compression risk.
NVIDIA vs AMD: A Practical Investor Framework
Business scale: NVIDIA is substantially larger. This is visible in total revenue, Data Center revenue, operating income and free cash flow.
Current profitability: NVIDIA has the stronger reported gross margin and operating economics.
Growth from a smaller base: AMD has more room for percentage share gains to materially change its revenue profile.
Software and ecosystem: NVIDIA benefits from CUDA and a highly integrated full-stack platform. AMD is building around ROCm and open standards, offering customers an alternative architecture.
Rack-scale roadmap: NVIDIA's Vera Rubin is in production ramp. AMD's MI400 and Helios platform is expected to move through volume deployment in the second half of 2026.
Valuation sensitivity: AMD's current trailing GAAP multiple is much higher, so future earnings growth carries more of the valuation burden. NVIDIA's lower multiple does not eliminate risk because the business is already generating unusually high margins and growth on a massive base.
BTI model signal: NVIDIA's proprietary BTI models disagree with one another enough that the stock should not be reduced to a single fair-value number. AMD's BTI market-cap model sits far below the current capitalization, highlighting the amount of future execution embedded in the market price.
Frequently Asked Questions
Is NVIDIA financially stronger than AMD today?
On the latest reported quarters, yes on the measurable operating metrics covered here. NVIDIA generated much more revenue, operating income and free cash flow, while also reporting a higher GAAP gross margin. That describes current financial strength, not future stock performance.
Is AMD growing faster than NVIDIA?
It depends on the metric. AMD's total Q2 revenue grew 50% year over year and Data Center revenue grew 107%. NVIDIA's total fiscal Q2 revenue grew 106% and Data Center revenue grew 117%. On those reported growth rates, NVIDIA was still growing faster despite its much larger base.
Which company has the stronger AI software ecosystem?
NVIDIA currently has the more established integrated ecosystem through CUDA, CUDA-X and its broader infrastructure stack. AMD's ROCm ecosystem is expanding and emphasizes open standards. The investment question is how quickly AMD can reduce the deployment and software advantages that support NVIDIA's platform economics.
Which stock is cheaper on trailing earnings?
At the October 1, 2026 close, NVIDIA's trailing GAAP P/E was about 29.2 versus about 157.2 for AMD. That makes NVIDIA cheaper on this single metric. It does not by itself establish intrinsic value because future growth, margins and capital requirements differ.
Does BTI consider NVIDIA undervalued or overvalued?
BTI's supplied models are mixed. Three high-suitability per-share models produced values below the October 1 price, while a separate market-cap framework indicated a discount to fair value. The disagreement means the valuation conclusion depends heavily on assumptions and methodology.
Does BTI consider AMD overvalued?
BTI's supplied proprietary market-cap model estimated fair value materially below AMD's current market capitalization. That is a model output, not a market fact or prediction. Investors should test the revenue, margin and multiple assumptions that create the gap.
What is the biggest thing to watch for NVIDIA?
The combination of Vera Rubin deployment, Data Center growth, gross-margin durability and free-cash-flow conversion. NVIDIA can continue growing while still disappointing the stock market if results fall below the expectations embedded in a multi-trillion-dollar valuation.
What is the biggest thing to watch for AMD?
Whether MI400, Helios and ROCm adoption convert announced customer momentum into recurring, high-margin Data Center revenue at sufficient scale to support the current valuation.
BTI Bottom Line
NVIDIA and AMD are not simply two interchangeable ways to buy exposure to AI chips. NVIDIA is the scaled full-stack infrastructure leader, with current economics that are dramatically stronger across revenue, margins and cash generation. AMD is the high-growth challenger, with a credible rack-scale roadmap and a strategic opening created by customers' desire for additional suppliers and open infrastructure.
The valuation setup makes the comparison more demanding. NVIDIA's lower trailing earnings multiple is supported by extraordinary present profitability, but its size means the market expects AI infrastructure spending to remain enormous. AMD's higher multiple gives it more room to create value through successful share gains, but less room for execution disappointment.
For research purposes, the cleanest framework is therefore not “Which stock wins?” but which set of assumptions is more defensible: continued durability of NVIDIA's platform economics, or enough AMD share capture and margin expansion to justify a much richer multiple on current earnings.
Disclosure: This article is for informational and educational purposes only. It is not personalized investment advice, a recommendation to buy or sell a security, or a guarantee of future returns. Financial data can change after publication. BTI valuation outputs are analytical estimates and should be reviewed alongside primary-source filings and independent assumptions.