By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel consensus is bearish on Nvidia's $15T valuation target by 2029, citing unsustainable growth rates, margin pressure from custom silicon competition, and potential TAM compression due to geopolitical risks and hardware commoditization.
Risk: Hardware commoditization and shift to custom silicon for inference, potentially leading to margin pressure and TAM compression.
Opportunity: None identified as a consensus opportunity.
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 the largest company in the world right now, but don't be surprised if it gets even bigger.
Nvidia's latest results provide a clear indication that its AI chip dominance isn't going to fade.
Nvidia (NASDAQ: NVDA) became the first company in the world to achieve a $5 trillion market capitalization in October 2025, fueled by the artificial intelligence (AI)-powered growth in its revenue and earnings in recent years.
The good news for Nvidia stock investors is that the semiconductor giant's growth isn't showing any signs of slowing down. It continues to dominate the lucrative AI chip market, and more importantly, Nvidia continues to hunt for new opportunities to sustain its phenomenal growth.
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 »
Nvidia's performance in the first quarter of fiscal 2027 (which ended on April 26) clearly shows that the intensifying competition in the AI chip market isn't affecting its growth. In fact, I won't be surprised if it becomes the world's first $15 trillion company in the next three years. Let's see why that's likely to be the case.
Nvidia has dominated the AI chip market in recent years thanks to its graphics processing units (GPUs), which offer massive parallel computing power, making them ideal for training large language models (LLMs). However, there is enough evidence suggesting that hyperscalers and AI companies prefer custom chips in the inference era to reduce compute costs.
That's not surprising, as inference workloads require much less computational power than the training phase, which is why GPUs are considered overkill for inference applications. However, Nvidia's latest results make it clear that its GPUs remain relevant in the AI inference era.
The company reported an 85% year-over-year increase in revenue in fiscal Q1 to $81.6 billion. That was a significant improvement over the 69% revenue growth it reported in the same quarter last year. The semiconductor specialist's non-GAAP earnings jumped by a whopping 140% year over year to $1.87 per share, again exceeding the 33% growth it clocked in the year-ago period.
Nvidia's guidance clearly suggests that its growth is poised to accelerate. The company anticipates $91 billion in revenue in the current quarter, a 95% increase over the year-ago period. The company's ability to accelerate growth despite achieving a massive revenue base is commendable, indicating that Nvidia is now in a robust position to capitalize on the next phase of the AI computing cycle.
According to Deloitte, AI inference will account for two-thirds of compute power in data centers this year. However, the consulting giant adds that instead of inference-focused chips, the majority of the computing will be performed by powerful chips such as GPUs. Given that Nvidia is designing its server racks to deliver higher inference performance at lower costs, it is easy to see why hyperscalers, sovereign customers, and cloud computing providers continue to line up for its chips.
Nvidia management noted on the latest earnings call that its next-generation Vera Rubin server racks can "deliver up to 35x higher inference throughput and up to 10x greater AI factory revenue compared with Blackwell." So, it is easy to see why Nvidia management is confident of achieving $1 trillion in revenue from its Blackwell and Rubin chips in 2026 and 2027.
What's more, the company believes that inference and agentic AI applications will significantly boost AI infrastructure spending from an estimated $1 trillion in 2026 to a range of $3 trillion to $4 trillion by the end of the decade. Nvidia reported $75.2 billion in data center revenue in fiscal Q1, translating into an annual run rate of $300 billion. The massive AI infrastructure spending the company expects by 2030 suggests it still has significant room for growth in this market.
That's the reason why analysts have become more bullish about its prospects, paving the way for Nvidia to cross the $15 trillion market cap milestone within the next three years.
Analysts have revised their growth expectations following Nvidia's latest results. As the following chart shows us, Nvidia's earnings estimates for the next three fiscal years have crept up.
Nvidia's expectation of a significant acceleration in AI infrastructure spending over the coming years could help it sustain its remarkable growth. However, even if its earnings increase to $15.51 per share in fiscal 2029 (which will end in January 2029) and it trades at 43 times earnings at that time (in line with the tech-focused Nasdaq Composite index's earnings multiple), its stock price could jump to $667.
That's just over triple its current stock price, which should be enough for this AI stock to breach the $15 trillion market cap milestone, given its current $5.2 trillion market cap. So, it would make sense for investors to continue loading up on Nvidia shares, as it can soar higher over the next three years on account of its outstanding growth potential.
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 23, 2026. *
Harsh Chauhan has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends 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 reaching $15T by 2029 is unlikely because custom-chip substitution and multiple compression will cap both earnings growth and valuation far below the article's assumptions."
The article's $15T target by FY2029 rests on Nvidia sustaining 80-95% revenue growth into a $300B+ annual run-rate while holding a 43x multiple. This ignores that inference workloads are shifting toward lower-cost custom ASICs from hyperscalers, which already account for rising shares of capex. Nvidia's own Blackwell-to-Rubin transition and rack-level pricing power may not offset margin pressure once AI infrastructure spend plateaus below the projected $3-4T. Historical mega-cap multiples have compressed sharply once growth slows from triple digits. The $667 share-price math assumes no competitive or cyclical erosion.
Even modest share gains in inference plus continued sovereign and cloud demand could keep Nvidia's growth above 40% long enough to justify elevated multiples, as the article's $1T Blackwell/Rubin revenue target for 2026-27 already implies.
"Nvidia reaching $15T requires both flawless execution on a tripling of earnings AND a multiple that doesn't compress—a two-variable bet where either one breaking materially changes the outcome."
The $15T thesis rests on Nvidia hitting $15.51 EPS by fiscal 2029 while maintaining a 43x P/E multiple. That's aggressive: it assumes revenue grows from ~$300B run rate (Q1 FY27) to ~$600B+ by FY29, with margins staying fat despite inevitable competitive pressure from custom silicon (AMD, Intel, and hyperscaler chips). The Deloitte stat claiming 'majority of inference still uses GPUs' contradicts the article's own premise that hyperscalers prefer cheaper custom chips. The math works only if Nvidia's moat is unbreakable—but we're already seeing TSMC capacity constraints ease and AMD gaining traction. A 43x multiple on a $600B revenue company assumes no multiple compression even as growth decelerates from 95% to perhaps 30-40% by 2029.
If custom inference chips capture even 40% of the inference workload (vs. the article's implied <30%), Nvidia's addressable market shrinks materially, and a 43x multiple becomes indefensible. Regulatory risk around chip export restrictions to China also isn't priced into this optimistic scenario.
"The transition from training to inference will inevitably commoditize AI hardware, leading to a significant compression of Nvidia's current valuation multiples."
The $15 trillion valuation target by 2029 is a mathematical hallucination that ignores the law of large numbers and hardware commoditization. While NVDA's Q1 fiscal 2027 revenue growth of 85% is impressive, maintaining that velocity on a $300B+ run rate is historically unprecedented. The article assumes a 43x P/E multiple remains sustainable as the company transitions from a high-growth disruptor to a cyclical infrastructure utility. If hyperscalers like Google, Amazon, and Microsoft successfully shift to custom silicon for inference to protect their own margins, Nvidia’s pricing power will erode. We are looking at a potential 'peak hardware' scenario where the infrastructure build-out outpaces actual enterprise software revenue realization.
If Nvidia effectively pivots to a software-and-services model via CUDA and AI factories, they may sustain premium margins that decouple them from traditional semiconductor cyclicality.
"Nvidia’s long-term upside hinges on an uninterrupted AI capex boom and a extreme valuation expansion that may not survive cyclical or regulatory headwinds."
The piece is aggressively bullish, tying Nvidia’s upside to a fresh wave of AI infrastructure spend and a potential $15 trillion market cap by 2029. The strongest countercase: the math requires a sustained, outsized multiple and perpetual demand for AI compute. Even if earnings stay robust, a 43x forward earnings multiple by 2029 implies an extreme re-rating that may not survive cyclical normalization, margin compression, or a shift toward cheaper inference hardware. Nvidia’s revenue is highly concentrated in data-center GPUs; any deceleration in AI capex, faster competition from custom chips, or regulatory/geopolitical frictions could cap upside. Valuation risk dwarfs near-term catalysts.
Bear case: AI demand could plateau or shift to cheaper, bespoke accelerators, narrowing Nvidia's advantages. Even if that proves less likely, a policy shock or supply-chain constraint could trigger a rapid mark-to-market off a stretched multiple.
"Export controls may hasten Chinese chip alternatives and shrink Nvidia's global TAM more than current ASIC or cyclical risks imply."
Claude flags the Deloitte inference contradiction but misses how US export curbs on advanced chips could accelerate Huawei Ascend adoption across China and Belt-Road markets, cutting Nvidia's addressable data-center TAM by 15-20% faster than domestic ASIC competition alone. This compounds ChatGPT's geopolitical point into a structural rather than cyclical risk, making the 43x multiple even harder to defend by FY2029 regardless of Rubin ramp success.
"Export restrictions are a tail risk; the core vulnerability is whether total AI infrastructure capex actually reaches the $3-4T the article implies by 2029."
Grok's geopolitical TAM compression is material but assumes Huawei Ascend scales faster than evidence suggests. More pressing: nobody has quantified what 'plateau' actually means for Nvidia's base case. If AI capex stays $500B annually (vs. article's $3-4T projection), Nvidia still grows 25-30% on inference alone. The $15T thesis dies on *demand assumptions*, not just competition. That's the real stress test.
"A $500B capex floor forces a transition to utility-like multiples, rendering a 43x P/E unsustainable as hardware commoditizes."
Claude, your focus on the $500B capex floor is the right stress test. If hyperscalers shift to custom silicon for inference, they aren't just cutting Nvidia's TAM; they are reclaiming the margins Nvidia currently captures. The $15T valuation assumes Nvidia remains the sole infrastructure provider, yet the shift toward 'AI factories' implies a commoditization of the hardware layer. If capex plateaus at $500B, Nvidia’s growth profile shifts from hyper-growth to a cyclical semiconductor utility, making a 43x multiple mathematically impossible.
"The missing link is how much Nvidia can monetize CUDA and services at scale if hardware growth slows; a durable software moat may not be enough to sustain 43x by FY2029."
Grok's China-export angle is plausible, but the real hole is assuming Nvidia can outgrow a hardware-price squeeze with software monetization alone. If hyperscalers move to in-house inference ASICs and export curbs accelerate, the TAM compression could dwarf the 15T thesis. The missing link is how much Nvidia can monetize CUDA and services at scale if hardware growth slows; a durable software moat may not be enough to sustain 43x by FY2029.
The panel consensus is bearish on Nvidia's $15T valuation target by 2029, citing unsustainable growth rates, margin pressure from custom silicon competition, and potential TAM compression due to geopolitical risks and hardware commoditization.
None identified as a consensus opportunity.
Hardware commoditization and shift to custom silicon for inference, potentially leading to margin pressure and TAM compression.