By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panelists generally agree that Nvidia's $1T data-center capex projection for 2027 is ambitious and relies on sustained returns on AI spend. They highlight potential risks such as energy constraints, custom ASIC adoption, and the possibility that AI ROI may not justify the spend. TSM and Micron are seen as beneficiaries but also face risks like geopolitical issues and memory-cycle volatility. The market is pricing in rapid growth, leaving little room for disappointment.
Risk: The energy wall and shift towards lower-power, specialized ASICs, which could cannibalize general-purpose GPU spend and blunt NVDA's capex upside.
Opportunity: Continued hyperscaler infrastructure spending, driving demand for NVDA, TSM, and MU products.
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's growth will last into 2027 if this projection is correct.
Taiwan Semiconductor is making the most of Nvidia's chips.
Micron will be a huge beneficiary of increased memory chip demand.
Nvidia (NASDAQ: NVDA) dropped a major bombshell during its Q1 earnings. Management expects data center capital expenditures to reach $1 trillion in 2027, which keeps the entire industry on track to achieve a massive $3 trillion to $4 trillion annual spend by 2030. That's a huge growth prediction, but Nvidia is already probably receiving orders for 2027 products that allow it to back this projection. That's major news, but Nvidia isn't the only one affected by it.
I've got three stocks that are no-brainer buys today if Nvidia is right on its $1 trillion call, and investors should consider scooping them up right now.
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 »
If $1 trillion in data center capital expenditures occurs in 2026, then Nvidia is clearly a buy in its own right. Nvidia is the leading artificial intelligence computing unit provider, and with the launch of its new Rubin architecture platform later this year, it has a built-in growth lever for generational change. Furthermore, as data centers are built out, there will be fewer costs devoted to land acquisition and construction, and more to the computing equipment that goes in the data center. That should boost Nvidia's revenue growth rate overall, even if Wall Street analysts project only 39% growth for the next fiscal year.
Analysts are not optimistic about underprojecting Nvidia's growth a year out. If Nvidia is right about its capital expenditure trajectory, its growth rate will likely be far higher, leading to impressive gains in 2026 and 2027.
Nvidia only designs the chips that go into its devices; it doesn't make them. Taiwan Semiconductor (NYSE: TSM) handles most of the fabrication work, as it does with several of Nvidia's competitors. Taiwan Semiconductor is the ultimate neutral investment in the AI space, as it is only a manufacturer. At the same time, it won't see the peak gains of AI's best performer, but it will benefit from the overall rising tide of increased chip demand.
Taiwan Semiconductor expects its AI chip business to grow at nearly a 60% compounded annual growth ratefrom 2024 to 2029, and with data center capital expenditure expected to rise significantly again next year, that bodes well for the business.
At 26 times forward earnings, Taiwan Semiconductor isn't the cheapest stock around, but it's also not overly expensive considering its positioning in the AI build-out.
Given the major impending AI growth, I think Taiwan Semiconductor is still a strong buy and a cornerstone for investors building an AI prototype.
Taiwan Semiconductor makes logic chips, which are a major component of Nvidia's GPUs. However, those devices also require memory chips, and Micron (NASDAQ: MU) is a key manufacturer. The memory chip market has been overwhelmed by demand from AI, and has practically sold out of all production capacity in 2026. In fact, Micron told investors that it can only meet half to two-thirds of medium-term demand. This low-supply and high-demand market has caused memory chip prices to skyrocket, leading to huge growth for Micron.
If data center capital expenditures take a major step forward in 2027, that will only increase the strain on the memory chip supply chain, as many of the efforts by memory chip producers to increase output won't be ready until later in 2027. That will keep memory chip prices elevated, allowing Micron to capitalize on the major growth. Right now, it's expected to triple its revenue. Next year, Wall Street analysts expect nearly a 60% growth rate -- all of which will allow investors to cash in on major memory chip demand via a Micron investment.
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 $463,900! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,294,401!
Now, it’s worth noting Stock Advisor’s total average return is 978% — a market-crushing outperformance compared to 211% 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 31, 2026. *
Keithen Drury has positions in Nvidia and Taiwan Semiconductor Manufacturing. The Motley Fool has positions in and recommends Micron Technology, Nvidia, and Taiwan Semiconductor Manufacturing. 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 $1T capex forecast is plausible but far from assured, leaving NVDA exposed to downside if ROI evidence or power availability disappoints."
Nvidia's $1T data-center capex call for 2027 implies continued 40%+ revenue growth for NVDA and outsized gains for TSM and MU, but the projection rests on unproven sustained returns on AI spend. Energy constraints, custom-ASIC adoption by hyperscalers, and potential 2026 digestion after the current build-out could cap actual orders. TSM's 60% AI CAGR guidance already prices in much of this upside at 26x forward earnings, while MU's memory pricing spike may reverse once new capacity comes online late 2027. The article underplays execution risk and valuation compression if growth merely meets rather than exceeds consensus.
If hyperscalers confirm the $1T trajectory with actual 2025-26 budgets, NVDA's Rubin ramp and MU's sold-out HBM capacity could drive multiples higher than today's already elevated levels.
"Micron is the only pick with genuine optionality—memory scarcity is structural through mid-2027, but TSM and NVDA already price in the bull case."
Nvidia's $1T capex projection for 2027 is credible given their order visibility, but the article conflates a supply-side forecast with demand certainty. The real risk: this assumes AI ROI justifies the spend. If LLM productivity gains plateau or prove unprofitable at scale, capex collapses faster than it rose. TSM at 26x forward P/E prices in most of this growth already. Micron is the only name with genuine upside asymmetry—memory pricing power persists only if supply remains constrained through 2027, which is NOT guaranteed given announced fab expansions by SK Hynix and Samsung.
The article assumes capex scales linearly with AI adoption, but hyperscalers are already optimizing inference costs and consolidating workloads; utilization rates may not justify incremental spending at the 2027 run rate, causing a sharp deceleration.
"The sustainability of the $1 trillion CapEx target depends entirely on the transition from infrastructure build-out to measurable, high-margin software revenue for the hyperscalers."
Nvidia’s $1 trillion CapEx projection for 2027 is a massive bet on continued hyperscaler infrastructure spending. While TSM and Micron are logical downstream beneficiaries, the market is currently pricing in a 'perfect execution' scenario. We are seeing a shift from 'AI experimentation' to 'AI monetization,' but the ROI for these hyperscalers remains unproven. If enterprise software adoption doesn't accelerate to justify this hardware spend, we face a classic cyclical glut. TSM is the most resilient play due to its foundry monopoly, but Micron’s HBM (High Bandwidth Memory) margins are highly sensitive to supply-side discipline. Investors must watch for signs of 'CapEx fatigue' among the big tech players before assuming 2027 growth is linear.
The thesis assumes that AI infrastructure spending will continue to scale linearly, ignoring the risk of a massive 'air pocket' if hyperscalers realize their current GPU clusters are underutilized and pause orders to focus on software-driven revenue.
"Durable AI compute demand is necessary for the capex ramp, and Nvidia's moat (software, ecosystem, and architecture) will determine whether the trillion-dollar story translates into sustained earnings growth."
The headline is that Nvidia's Q1 signal points to a data-center capex surge to $1 trillion by 2027 and a broader AI build-out. The strongest counter: capex is a demand signal, not revenue; it can be episodic, delayed, or pulled forward, and not all capex converts into Nvidia revenue or margin expansion. Even with higher capex, customers may diversify with non-GPU accelerators, software/hardware efficiency, and budget constraints. TSM's capacity, geopolitical risk, and Micron's memory-cycle volatility add timing risk. The stock's lofty multiple already prices in rapid, durable growth; a misstep in capex timing or AI uptake could punish valuations.
Against this bullish read: the $1T capex target is a forecast, not a guarantee. If AI demand cools or GPU share is squeezed by competitors or efficiency gains, Nvidia may not translate capex into proportional earnings.
"TSM packaging sold-out status through 2026 may force capex continuity despite ROI doubts raised by the panel."
Gemini highlights CapEx fatigue risk but misses how TSM's CoWoS advanced packaging remains sold out into 2026, creating a hard supply choke that could compel sustained hyperscaler orders even if near-term AI ROI disappoints. Grok's energy constraints would amplify this bottleneck if new capacity lags. This dynamic could support NVDA and MU revenues longer than the digestion scenario others assume, though it raises TSM's own execution exposure.
"CoWoS scarcity is real but not insurmountable; packaging diversification could undermine the supply-lock thesis by 2026."
Grok's CoWoS bottleneck argument is mechanically sound, but it assumes hyperscalers can't route around it. Samsung's advanced packaging and Intel's foundry push are real alternatives maturing 2025–26. If even 20% of orders divert, TSM's sold-out narrative collapses and NVDA's margin profile compresses despite capex growth. The article ignores competitive packaging capacity entirely—a material blind spot.
"Power density constraints and specialized ASIC adoption pose a greater threat to GPU-centric capex than competitive packaging alternatives."
Claude is overly optimistic about Intel and Samsung's packaging capabilities. Intel Foundry is struggling with yield and cost discipline, while Samsung’s CoWoS-equivalent yields remain unproven at scale. TSM’s moat is not just capacity; it is the ecosystem integration with NVDA. If we are looking for a real risk, it is not competitive packaging, but the 'energy wall.' Hyperscalers are hitting power density limits that will force a shift toward lower-power, specialized ASICs, cannibalizing general-purpose GPU spend.
"Energy-density constraints could push hyperscalers toward lower-power ASICs, undermining GPU-driven capex and NVDA's revenue/multiples growth by 2027."
Responding to Gemini: the energy-density hurdle you flagged could meaningfully blunt NVDA’s capex upside, not just delay it. If hyperscalers accelerate toward lower-power, custom ASICs and software-driven optimization, the ROI of additional GPU-based infrastructure may shrink—cannibalizing NVDA's growth even with a $1T 2027 capex target. CoWoS bottlenecks matter, but a demand mix shift away from general-purpose GPUs would be the bigger, underappreciated risk to revenue and multiples.
The panelists generally agree that Nvidia's $1T data-center capex projection for 2027 is ambitious and relies on sustained returns on AI spend. They highlight potential risks such as energy constraints, custom ASIC adoption, and the possibility that AI ROI may not justify the spend. TSM and Micron are seen as beneficiaries but also face risks like geopolitical issues and memory-cycle volatility. The market is pricing in rapid growth, leaving little room for disappointment.
Continued hyperscaler infrastructure spending, driving demand for NVDA, TSM, and MU products.
The energy wall and shift towards lower-power, specialized ASICs, which could cannibalize general-purpose GPU spend and blunt NVDA's capex upside.