By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel is divided on Nvidia's CPU push. Bulls see it as a strategic move to diversify revenue and open a new AI data-center TAM, while bears caution about execution risks, unproven technology, and potential competitive responses from Intel and AMD.
Risk: Execution risk in integrating a new CPU architecture and achieving seamless compatibility with legacy x86 workloads.
Opportunity: Potential to capture new AI-native workloads in AI clusters where x86 has zero installed base.
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 →
Nvidia is known for its dominance in the graphics processing unit (GPU) sector.
But now the company is getting involved in another critical area of artificial intelligence.
Intel and AMD are already deeply involved in this area, and the emergence of Nvidia could potentially threaten their market share or runway.
As has become the norm, the artificial intelligence (AI) chip giant Nvidia (NASDAQ: NVDA) recently reported strong earnings for the first quarter of its fiscal 2027.
The company reported adjusted earnings and revenue ahead of Wall Street estimates and also provided a current-quarter revenue outlook well ahead of analysts' expectations.
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
However, the unexpected part of the quarter came when Nvidia Chief Executive Officer Jensen Huang, on theearnings conference call announced that the company is making a big move in central processing units (CPU) and is expecting a strong year of revenue from the division, as well as a new total addressable market.
Did Nvidia just say "checkmate" to Intel and Advanced Micro Devices?
In recent years, one of the big parts of the AI story, particularly as it relates to Nvidia, has been graphics processing units (GPUs), chips that are vital for training large language models (LLMs) and which are deployed in data centers specifically designed to run AI solutions.
GPUs have parallel processing capabilities and can therefore process much more data and consider many more solutions to a problem than CPUs. Nobody makes GPUs better than Nvidia, which has at least 80% of the market.
Furthermore, Nvidia's software layer for its GPUs, called CUDA (compute unified device architecture), was released in 2006, and Nvidia has built a strong ecosystem many companies and developers have become accustomed to. This makes switching away from this operating system difficult and is part of the reason Nvidia currently has such a strong competitive moat.
CPUs are the chips that are usually used in older, more common devices like desk top computers and mobile phones, and, until recently, were not viewed as a big part of the AI story.
However, the rise of agentic AI, autonomous systems that can complete a range of tasks without human intervention, has led to soaring CPU demand. Companies have found that CPUs are the most efficient at orchestrating the specific workflows that enable AI agents to perform tasks.
One example Huang gave on the company'searnings callis that if an agent wanted to do a task that involved fetching data from an internet browser, that task would run on a CPU.
Nvidia has long specialized in GPUs, and dominance in this market has led the company to a more than $5.2 trillion market cap, making it the largest publicly traded company by market cap in the world.
However, surging demand for CPUs for agentic AI has driven tremendous gains this year alone for some of the leaders in the industry, such as Intel and AMD.
The opportunity only seems to be growing, too, according to Intel CEO Lip-Bu Tan.
"On the inference side, in terms of orchestration, control plane, and also managing all the different agents with data, CPU is much more efficient," Tan said on Intel's most recentearnings conference call "The ratio of CPU to GPUs used to be 1-to-8, and now it is 1-to-4, and I think it could move toward parity or even better."
On Nvidia's conference call, management surprised Wall Street analysts by announcing that it is now in the data center CPU market, specifically making ones for agentic AI use.
Nvidia Chief Financial Officer Collette Kress said the company is rolling out its new Vera CPU, which is reportedly up to 1.5 times faster than comparable alternatives.
Kress also said that the Vera CPU creates a brand-new $200 billion opportunity for Nvidia and that the company expects nearly $20 billion in total CPU revenue in the current fiscal year alone.
In the first quarter of 2026, Intel delivered more than $5 billion of revenue in its data center and AI division, which includes its CPU product business. Annualized, that's slightly more than Nvidia's $20 billion CPU projection.
AMD reported nearly $5.8 billion in data center revenue in the first quarter of this year. However, this includes both GPU and CPU sales, which the company does not break out.
So if Nvidia were to hit $20 billion in CPU sales this year, it could very well be in line with Intel and larger than AMD's business right away.
Is it an immediate checkmate?
I suspect there are likely enough opportunities for Intel, AMD, and Nvidia to all do well in the agentic AI CPU market. But make no mistake, Nvidia's move likely makes it an immediate leader in the industry, and a big challenger to Intel and AMD.
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 $477,813! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,320,088!
Now, it’s worth noting Stock Advisor’s total average return is 986% — a market-crushing outperformance compared to 208% 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 May 26, 2026. *
Bram Berkowitz has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Advanced Micro Devices, Intel, 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
"Nvidia's CPU entry faces steeper ecosystem and qualification barriers than the earnings hype implies, capping near-term displacement of Intel and AMD."
Nvidia's Vera CPU announcement and $20B FY revenue target target the agentic AI orchestration layer where CPUs handle workflow control better than GPUs. Yet the article glosses over Nvidia's limited CPU track record—Grace has seen narrow adoption—and the x86 compatibility moat Intel and AMD retain in enterprise data centers. Manufacturing scale, driver maturity, and the 1-to-4 CPU/GPU ratio cited by Intel's CEO suggest Nvidia faces multi-year qualification cycles rather than instant share capture. The new $200B TAM claim assumes seamless integration that prior Nvidia CPU efforts never achieved.
Nvidia could bundle Vera with its dominant GPUs and CUDA stack to create an integrated platform that forces rapid customer migration, bypassing traditional CPU evaluation timelines.
"Nvidia's $20B CPU revenue projection is aspirational guidance, not evidence of market capture—execution risk and unproven customer adoption make this a 2027 story, not a 2026 certainty."
The article conflates announcement with execution. Nvidia claiming $20B CPU revenue in FY2027 is a projection, not a result. Vera CPUs are unproven in production at scale—1.5x faster than 'comparable alternatives' is vague (comparable to what? AMD EPYC? Intel Xeon?). The real risk: CPU design is fundamentally different from GPU dominance. Intel and AMD have 20+ years of CPU ecosystem, OEM relationships, and validation. Nvidia's CUDA moat doesn't transfer to CPUs. The article also ignores that agentic AI CPU demand is speculative—no customer has yet committed to massive Vera orders. Finally, $20B projection assumes zero execution risk and zero competitive response from Intel/AMD in a 12-month window.
If Nvidia actually ships Vera at scale with 1.5x performance and lands hyperscaler design wins, $20B is conservative—the TAM could justify it. CUDA's ecosystem advantage might transfer via software integration with GPU clusters.
"Nvidia’s move into CPUs is less about revenue diversification and more about creating a closed-loop hardware ecosystem that makes Intel and AMD redundant in the AI-agent era."
Nvidia’s pivot into CPUs via the 'Vera' architecture is a strategic masterstroke, effectively verticalizing the data center stack. By bundling GPUs with proprietary CPUs, Nvidia forces a 'lock-in' that threatens Intel’s x86 dominance and AMD’s EPYC market share. If Nvidia captures $20B in CPU revenue, they aren't just competing; they are commoditizing the host processor. However, investors must be wary of execution risk. Integrating a new CPU architecture into existing hyperscaler supply chains is vastly different from selling high-margin GPUs. If Vera fails to achieve seamless compatibility with legacy x86 workloads, Nvidia’s 'checkmate' could quickly become a costly distraction that dilutes their focus on the GPU moat.
Nvidia’s entry into the CPU space may trigger an antitrust 'anti-bundling' backlash from regulators and hyperscalers who are already desperate to reduce their dependency on a single-vendor ecosystem.
"Nvidia could unlock a meaningful CPU-driven growth leg, but hitting $20B in annual CPU revenue remains highly uncertain and contingent on rapid adoption and software-stack compatibility."
Nvidia’s CPU push could diversify revenue beyond GPUs and open a new AI data-center TAM (Vera CPU, ~$200B) with an implied ~$20B CPU revenue target. If demand materializes, it could meaningfully broaden Nvidia’s growth runway and potentially compress the tech giants’ AI hardware cycle into a more balanced mix. Yet the article’s optimism rests on an unproven Vera SKU, untested software integration, and an apples-to-oranges comparison to Intel/AMD’s existing data-center revenues. Adoption timelines, margins, and the risk of cannibalization or stalling partnerships are real headwinds that could erase this story’s near-term upside.
The move may be a classic early-stage bet that never pays off: data-center buyers cling to established x86 ecosystems and CUDA-centric workflows, and Nvidia lacks a proven CPU software moat, so Vera’s adoption could be slow and financially underwhelming.
"Prolonged qualification cycles pose a larger risk to the $20B target than regulatory pushback."
Gemini overstates the antitrust threat while underplaying how Intel and AMD could accelerate their own AI-optimized CPUs in response. The $20B target hinges on hyperscalers accepting Arm-based Vera despite x86 software inertia that Grok highlighted. If qualification cycles stretch beyond 2027 as Intel's CEO implied, Nvidia's vertical integration play collapses into margin dilution from unsold inventory rather than lock-in success.
"Vera's addressable market is AI orchestration, not x86 replacement—a distinction that materially raises adoption probability versus the panel's framing."
Claude and Grok both assume x86 inertia is insurmountable, but they're conflating two different buyer behaviors. Hyperscalers (Meta, Google, MSFT) already run custom silicon for inference; they're not locked into x86 like enterprise IT is. Vera's real TAM isn't replacing Xeon in legacy workloads—it's capturing orchestration layers in AI clusters where x86 has zero installed base. The $20B assumes Nvidia wins *new* AI-native workloads, not x86 displacement. That's a different (and higher-probability) thesis than the panel is debating.
"Nvidia's attempt to verticalize the CPU stack risks cannibalizing their relationship with hyperscalers who are already investing heavily in proprietary, non-Nvidia custom silicon."
Claude is right about the hyperscaler pivot, but misses the capital expenditure reality. Hyperscalers are increasingly designing their own custom silicon—Google’s Axion or Microsoft’s Maia—to avoid vendor lock-in. If Nvidia forces a 'Vera-only' stack, they risk alienating the very customers currently funding their GPU dominance. This isn't just about x86 vs. Arm; it's about Nvidia competing directly with their largest customers' internal roadmap, which is a massive strategic overreach that could trigger supply chain friction.
"Software ecosystem parity is the real gate to Vera's adoption; without production-grade toolchains and libraries, the 20B TAM cannot materialize."
Claude makes a plausible case that TAM hinges on AI-native workloads, but the real hinge is software ecosystem parity. Vera will only gain share if Nvidia ships production-grade toolchains, compilers, libraries (BLAS-like, CUDA‑translation), and robust virtualization—all without derailing GPU workflows. Absent those, hyperscalers either ignore Vera or keep running legacy x86 orchestration, so the '20B' claim looks optimistic regardless of hardware talent.
The panel is divided on Nvidia's CPU push. Bulls see it as a strategic move to diversify revenue and open a new AI data-center TAM, while bears caution about execution risks, unproven technology, and potential competitive responses from Intel and AMD.
Potential to capture new AI-native workloads in AI clusters where x86 has zero installed base.
Execution risk in integrating a new CPU architecture and achieving seamless compatibility with legacy x86 workloads.