By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel consensus is that AMD's high forward P/E (63x) is justified by its catch-up potential in the data center market, but the key risk is AMD's dependence on Nvidia's foundry TSMC for production capacity, which could be prioritized for Nvidia's own custom ASICs, potentially limiting AMD's growth.
Risk: TSMC capacity constraints and potential prioritization for Nvidia's custom ASICs
Opportunity: AMD's catch-up potential in the data center market
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
In the rapidly expanding semiconductor ecosystem, Advanced Micro Devices (NASDAQ: AMD) and Nvidia (NASDAQ: NVDA) stand as pivotal players powering the compute backbone of artificial intelligence (AI). Nvidia dominates with its Blackwell and Rubin GPU architectures, which deliver industry-leading performance for large-scale training and inference workloads.
Meanwhile, AMD contributes complementary technology via its Instinct MI-series accelerators and Epyc processors, offering high-performance CPUs and GPUs that hyperscalers integrate for cost-effective AI clusters.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Hyperscalers such as Microsoft, Amazon, Alphabet, and Meta Platforms rely on both companies extensively -- deploying Nvidia GPUs for AI acceleration while leaning on AMD's CPU and GPU offerings for broader data center efficiency and diversification away from single-supplier risk. Yet when the valuations are examined side by side, only one of these AI chip stocks emerges as the clear buy right now.
The chart below illustrates the forward price-to-earnings (P/E) ratios for Nvidia and AMD. The trends reveal a striking divergence: AMD currently trades at a forward P/E of 63, while Nvidia sits at 24. Over the displayed time period spanning mid-2024 through today, Nvidia's forward P/E has largely existed in a band between roughly 20x and 40x, whereas AMD's ratio has climbed sharply in more recent months.
AMD's premium appears counterintuitive. Competitive pressures from custom ASICs developed by hyperscalers -- Google's TPUs, Amazon's Trainium and Inferentia chips, and Microsoft's Maia -- appear to be intensifying primarily against Nvidia, whose GPUs face substitution risk as cloud providers seek to lower infrastructure costs and assert greater control over their chip stack.
Therefore, some investors might actually expect Nvidia's valuation multiple to compress aggressively. However, these same dynamics do not apply to AMD, which commands an elevated valuation, while Nvidia's profile has stabilized to a more modest level.
This discrepancy suggests investors are pricing in catch-up potential for AMD's Instinct accelerators and accompanying software stack, even as the ASIC threat remains acute for Nvidia as the incumbent AI chip designer.
AMD's second-quarter financial results showed solid but comparatively measured expansion. The company reported total revenue of $11.5 billion, a 50% increase year over year. AMD's data center segment -- the primary AI growth driver -- generated $6.7 billion in sales, surging 107% year over year. Meanwhile, free cash flow grew modestly to $1.6 billion.
Nvidia has yet to report second-quarter results. But during the first quarter, Nvidia showed far greater scale and velocity relative to AMD. Total revenue reached $81.6 billion, up 85% year over year, with record data center revenue of $75.2 billion, climbing 92%. Lastly, Nvidia's free cash flow of $48.5 billion nearly doubled from the prior year.
AMD's absolute numbers remain a fraction of Nvidia's, and its overall growth rate lags its larger rival. Despite this slower trajectory and substantially smaller revenue base, AMD's valuation sits at 2.5 times Nvidia's, a valuation gap that seems difficult to justify based on growth and profitability metrics alone.
The trends explored above illustrate a recurring pattern in which both Nvidia and AMD tend to rerate once their forward P/E ratios approach an elevated threshold. In prior cycles, AMD's forward P/E spiked above 50 or 60 only to experience subsequent normalization through multiples compression.
Nvidia has shown similar behavior, with its multiple rising and then settling into more sustainable ranges (which is happening right now). AMD's current levels position the stock for potential normalization in the near term, while Nvidia's relatively modest multiple leaves room for valuation expansion if its growth momentum persists.
History suggests AMD shares could face selling pressure as its premium unwinds, even as Nvidia could accelerate into a new leg higher. Given Nvidia's superior scale, faster overall growth, significantly higher profitability, and more entrenched position in hyperscale AI ecosystems, I think the risk-reward balance favors investing in Nvidia at the moment. Investors seeking exposure to the ongoing AI semiconductor boom might be better served by buying Nvidia stock over AMD right now while the valuation gap remains this wide.
Before you buy stock in Nvidia, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Nvidia wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $411,427! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,335,252!
Now, it’s worth noting Stock Advisor’s total average return is 965% — a market-crushing outperformance compared to 215% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of August 12, 2026. *
Adam Spatacco has positions in Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Meta Platforms, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"The article's core valuation thesis is reasonable but undermined by factual errors, incomplete ASIC risk analysis for Nvidia, and omission of AMD's potential software catch-up."
The article argues Nvidia (NVDA) at 24x forward P/E is the far better buy than AMD at 63x given NVDA's vastly superior scale ($81.6B Q1 rev vs AMD's $11.5B Q2), faster growth, and higher profitability. Yet the piece contains a glaring error: it claims AMD "outperformed Nvidia throughout 2026"—impossible since we are still in 2024. Historically both stocks have seen multiple compression at high forward P/Es, but the article downplays Nvidia's greater exposure to hyperscaler ASIC substitution (Google TPU, Amazon Trainium, Microsoft Maia) that could cap its growth more than AMD's diversified CPU/GPU mix. AMD's Instinct MI300 ramp and software improvements could narrow the performance gap faster than consensus expects.
If Nvidia continues to dominate AI training and inference with Blackwell/Rubin while AMD's MI-series struggles with ecosystem stickiness, the valuation gap could widen further rather than normalize, making the article's Nvidia preference prescient despite its sloppy 2026 reference.
"The valuation spread between AMD and NVDA is less about growth potential and more about the market pricing in the high probability of NVDA's margin compression versus AMD's operational leverage."
The article’s reliance on a 63x forward P/E for AMD vs. 24x for NVDA is a classic 'value trap' setup that ignores the composition of earnings. AMD’s multiple is inflated by heavy R&D reinvestment and a cyclical trough in its gaming/embedded segments, while NVDA’s 24x multiple reflects a peak-margin environment that may face mean reversion as hyperscalers aggressively deploy custom ASICs. While NVDA’s cash flow is undeniably superior, the market is currently pricing AMD as a high-beta recovery play. The real risk isn't just valuation compression; it's the software moat. If ROCm (AMD’s software stack) fails to achieve parity with CUDA, that 63x multiple will collapse, regardless of hardware performance.
If hyperscalers prioritize supply chain diversification to break NVDA's pricing power, AMD could capture significant market share regardless of software friction, justifying a premium for its role as the 'only alternative.'
"AMD's 63x multiple is elevated but defensible if it captures 20%+ share of incremental AI accelerator demand; the article's 'history suggests reversion' argument ignores that structural demand dynamics have shifted since prior cycles."
The article's valuation argument is mechanically sound—63x forward P/E vs. 24x is a real gap—but it conflates mean reversion with investment merit. AMD's premium isn't irrational if hyperscalers genuinely believe MI-series can displace 30-40% of Nvidia's TAM over 3-5 years. The article acknowledges this ('catch-up potential') then dismisses it. More critically: Nvidia's 24x assumes current $75B data center revenue sustains at 92% growth. If custom ASICs compress that to 50% growth, Nvidia's multiple compresses faster than AMD's normalizes. The article also omits that AMD benefits from supply-chain diversification mandates—a structural tailwind, not a cyclical one.
If Nvidia's moat is as durable as the article claims, why are hyperscalers spending billions on custom silicon? And if AMD's valuation is purely cyclical compression waiting to happen, why hasn't it compressed despite MI-series shipping in volume for 18+ months?
"AMD could deliver meaningful upside by gaining share in hyperscaler AI workloads through its Instinct MI lineup and software stack, potentially narrowing the gap with Nvidia and justifying a tighter multiple over time."
The article frames Nvidia as the AI compute king and AMD as a distant challenger, but the risk is distraction by optics: Nvidia’s scale and software moat could sustain a premium even if AI capex moderates, while AMD’s huge forward multiple implies outsized skepticism about its growth trajectory. AMD’s Q2 data-center growth of 107% on $6.7B rev, vs Nvidia’s $75.2B data-center rev in Q1 and $48.5B FCF, highlights that AMD can deliver rapid expansion on a smaller base. The missing piece is whether AMD can monetize its Instinct/MI stack and software gains fast enough to justify the valuation, especially if hyperscalers tilt back toward Nvidia CUDA and ecosystem lock-in.
NVIDIA's platform moat and superior profitability justify the premium; AMD already faces a path to slower growth and margin pressure as capex normalizes, making a re-rating unlikely in the near term.
"Hyperscaler ASIC builds tighten CoWoS supply, indirectly supporting higher margins for both NVDA and AMD rather than compressing Nvidia's growth."
Claude's point on ASIC spend undermining Nvidia's assumed growth is valid, but nobody has flagged the second-order winner: TSMC. Both NVDA and AMD are 100% reliant on TSMC's CoWoS capacity; any ASIC ramp from Google/Amazon simply starves GPU supply further, creating an artificial scarcity premium that benefits the duopoly rather than eroding Nvidia's pricing power.
"TSMC supply constraints will force hyperscalers to prioritize their own custom ASICs over AMD, disproportionately eroding Nvidia's high-margin marginal revenue."
Grok, your focus on TSMC capacity is the missing link, but you’re ignoring the 'Nvidia tax.' If hyperscalers face CoWoS bottlenecks, they will prioritize their own custom ASICs over AMD’s MI-series to maximize ROI per wafer. This makes Nvidia’s moat even more vulnerable to internal cannibalization by its own customers. Claude is right about the 30-40% TAM displacement risk; if TSMC capacity is the constraint, Nvidia loses the marginal unit, which is where the highest margins reside.
"TSMC capacity constraints hurt AMD more than Nvidia because hyperscalers will starve AMD allocation to maximize their own ASIC ROI."
Gemini and Grok are both right about TSMC capacity, but they're missing the asymmetry: Nvidia's custom ASIC customers (Google, Amazon, Meta) have zero incentive to share wafer wins with AMD. They'll hoard CoWoS for their own silicon. AMD doesn't benefit from scarcity—it gets rationed out. The real question: does AMD's MI-series even reach meaningful volume if TSMC prioritizes Nvidia and hyperscaler ASICs? That 107% growth rate assumes supply availability that may not materialize.
"CoWoS bottlenecks could accelerate a shift to bespoke AI accelerators, allowing AMD to win incremental workloads and eroding Nvidia's exclusivity."
Responding to Claude's asymmetry point: the real wildcard is whether hyperscalers’ CoWoS bottlenecks accelerate a broader shift away from general GPUs toward bespoke AI accelerators. If the 'NVIDIA tax' persists but capacity constraints drive customers to own silicon, Nvidia's share gains may stall even with pricing power. In that case, AMD's MI/ROCm could capture incremental workloads, not just software parity, due to silicon diversification reducing Nvidia's exclusivity.
The panel consensus is that AMD's high forward P/E (63x) is justified by its catch-up potential in the data center market, but the key risk is AMD's dependence on Nvidia's foundry TSMC for production capacity, which could be prioritized for Nvidia's own custom ASICs, potentially limiting AMD's growth.
AMD's catch-up potential in the data center market
TSMC capacity constraints and potential prioritization for Nvidia's custom ASICs