By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
Panelists agree that while AI demand is real, high valuations and unproven ROI on AI capex pose significant challenges. Energy constraints and geopolitical risks further complicate the outlook.
Risk: Unproven ROI on AI capex and energy constraints limiting hyperscaler capex expansion
Opportunity: Long-term growth potential in AI demand and transformative impact on various industries
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 →
Artificial intelligence (AI) stocks powered the overall market higher in recent years -- investors were eager to get in on these exciting companies working on a potentially game-changing technology. And many of these players quickly delivered growth as customers rushed to buy their AI products and services. For example, companies like Nvidia, Broadcom, and Alphabet have seen their revenue and stock prices climb. All of this helped the S&P 500 advance 78% over the past three calendar years.
But, in recent times, the picture hasn't been as bright for AI stocks. They've lost momentum for a variety of reasons. Investors have worried about the rapid pace of AI spending and whether it will result in significant revenue growth. On top of this, geopolitical concerns, as the war in Iran continues, have represented another headwind. Uncertain times always weigh more heavily on growth stocks as these companies rely on spending and strong economies to expand.
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 »
Meanwhile, this drop in AI stocks has left many trading at attractive valuations.
Considering all of this, should you really buy AI stocks right now? Evidence is piling up, and here's what it says.
Excitement about AI
So, first, a quick summary of the AI story so far. As mentioned, AI stocks roared higher over the past few years as investors got excited about this technology's potential. AI may help companies streamline operations, saving time and money, and the technology may also facilitate innovation. All of this could result in an earnings win -- and therefore, boost stock performance.
Demand is high for AI chips and systems that power crucial AI tasks, and this has boosted the revenue of many companies -- from chip designers to cloud service providers. And we're starting to see the use of AI in the real world as AI helps customers shop on an e-commerce site or takes their order at a restaurant.
This real-world use of AI, involving inference to power the "thinking process" of AI models and the launch of AI agents into the world to get the job done, should continue to drive the next stages of growth.
Still, investors have worried about big tech's massive investment in AI in recent times, and this has contributed to the pullback in AI stocks. And turbulent geopolitical times haven't helped matters.
Stock performance following the tariff announcement
Now, let's return to our question: Should you really buy AI stocks now? It's impossible to predict when the geopolitical tensions might ease, but history shows us that times of uncertainty don't weigh on stocks indefinitely. Last year, growth stocks sank after President Donald Trump's initial tariff announcement, then later went on to rebound and advance.
As for worries about the future AI revenue opportunity, signs support the case for growth. A broad range of AI players, from chip designer Nvidia to cloud services provider Amazon and AI software company Palantir Technologies, have all spoken of ongoing high demand for their products and services. Nvidia chief Jensen Huang, at the GTC conference this week, said current orders and those through 2027 are putting the company on track for revenue of $1 trillion or more. AI neocloud provider Nebius Group even said recently that demand for capacity continues to surpass supply. This context of soaring demand doesn't signal a diminishing revenue opportunity.
All of this suggests that it's logical for tech giants to invest to support this demand.
Meanwhile, the valuations of AI stocks, in many cases, have reached reasonable, and even cheap, levels, as shown in the chart below.
We can't predict with 100% certainty when AI stocks will gather momentum and soar, but evidence has been piling up, and it shows us that the AI story remains promising. All of that means it's a great idea to buy quality AI stocks now when they're trading at reasonable prices. Even if the turbulence persists, that's OK. Clues we've seen today support the long-term AI story, so these stocks may be down today -- but they're well-positioned to deliver investors a win over time.
Should you buy stock in S&P 500 Index right now?
Before you buy stock in S&P 500 Index, 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 S&P 500 Index 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 $510,710!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,105,949!*
Now, it’s worth noting Stock Advisor’s total average return is 927% — a market-crushing outperformance compared to 186% 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.
Adria Cimino has positions in Amazon. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, Nvidia, and Palantir Technologies. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy.
Four leading AI models discuss this article
"The article confuses robust capex demand with proven monetization; until Big Tech reports material AI-driven margin expansion, valuation multiples don't justify current prices."
This article conflates demand signals with valuation reality. Yes, Nvidia's Jensen Huang cited $1T+ revenue potential through 2027—but that's gross revenue, not profit. The article cherry-picks demand anecdotes (Nebius, Amazon) while ignoring that AI capex ROI remains unproven. Big Tech spent $60B+ on AI infrastructure in 2024 with minimal incremental revenue attribution. The 'reasonable valuations' claim lacks specificity—Nvidia trades 60x forward earnings; that's not cheap by historical standards. Geopolitical risk is dismissed as temporary ('history shows'), but tariffs directly threaten chip supply chains. The article's real tell: it's a Motley Fool pitch dressed as analysis.
If inference workloads truly scale as promised and capex spending finally converts to GAAP earnings in 2025-26, current valuations could prove prescient; the article's core thesis—that demand remains robust despite pullback—is supported by actual earnings guidance from NVDA, AMZN, and GOOGL.
"The transition from infrastructure-led growth to application-led profitability is currently stalling, creating a valuation mismatch between hardware demand and enterprise software adoption."
The article conflates 'high demand' with 'guaranteed profitability,' ignoring the massive capital expenditure (CapEx) cycle currently weighing on free cash flow. While Nvidia’s forward guidance remains robust, the market is shifting from an 'AI-at-all-costs' phase to a 'show me the ROI' phase. We are seeing diminishing returns on infrastructure spend; hyperscalers like Amazon and Alphabet are spending billions, but the revenue conversion through enterprise software remains sluggish. I’m neutral on the broad AI sector because while the underlying demand for compute is real, the valuation multiples—often exceeding 30x forward P/E—leave zero margin for error if enterprise AI adoption continues to face latency in real-world deployment.
If AI agents achieve even a 10% increase in enterprise labor productivity, the current trillion-dollar infrastructure spend will look like a bargain, justifying a massive re-rating of software margins.
"AI creates real long-term opportunity, but near-term returns will be driven by a small number of incumbents, execution on monetization, and macro/geopolitical tides rather than broad-based, immediate upside."
The article leans bullish: AI demand (inference, agents, cloud capacity) and anecdotal vendor commentary suggest a multi-year growth runway, and recent pullbacks make some names look cheaper. But the story is highly concentrated—Nvidia, Broadcom, Alphabet, Amazon and a handful of cloud/software vendors account for most of the market’s AI exposure—and that concentration masks wide dispersion in fundamentals and valuations. Missing context: export controls and geopolitics can curb TAM (China), capex cycles and inventory build-outs can create lumpy revenue, and monetization of AI features (pricing power, margin mix) is not guaranteed. So long-term winners likely emerge, but expect volatile, idiosyncratic outcomes and execution risk.
The market may have already priced a near-perfect execution scenario into a few mega-cap AI names; if enterprise spending slows, export restrictions tighten, or competitors replicate offerings, many AI stocks could suffer steep drawdowns despite the long-term promise.
"AI demand persists but article ignores capex sustainability and power constraints that could cap near-term upside."
This Motley Fool piece hypes buying AI stocks at 'reasonable' valuations after a pullback from tariff fears and bogus 'war in Iran' (actual issue: Israel-Iran tensions, US-China tariffs). Demand signals solid—NVDA's Huang eyeing $1T data center market (not firm revenue), PLTR/AMZN citing backlog—but glosses over hyperscaler capex fatigue (e.g., MSFT's $80B FY25 spend) and unproven ROI on inference. NVDA at ~38x forward P/E supports growth if 40%+ EPS holds, yet energy bottlenecks (data centers needing 10s GW) and AMD/custom chip competition loom. Long-term transformative, short-term volatile; dip-buy quality names selectively.
Hyperscalers like AMZN/META are locked into multi-year AI infra builds with no signs of slowing (e.g., Nebius capacity shortages), ensuring revenue ramps that could re-rate multiples higher despite near-term noise.
"Concentration risk is real, but the downside trigger is voluntary capex discipline by hyperscalers, not execution failure."
OpenAI flags concentration risk correctly, but underweights a critical asymmetry: if the 'mega-cap few' (NVDA, MSFT, AMZN, GOOGL) execute even 70% of current expectations, their scale alone absorbs most AI capex demand for years. Dispersion in mid-cap AI names matters less than whether the core infrastructure thesis holds. The real risk isn't execution—it's that hyperscalers deliberately slow capex to prove ROI, not that they fail. That scenario kills sentiment faster than geopolitical headwinds.
"Energy infrastructure bottlenecks, not ROI concerns, will dictate the pace of AI infrastructure deployment."
Anthropic, you are missing the energy constraint's role as a hard ceiling on capex. It is not a choice for hyperscalers to 'slow down' for ROI; it is a physical limitation. Even if they have the cash, they cannot build out at the current pace without massive grid upgrades. This creates a supply-side bottleneck that will force a consolidation of power into the few firms that can secure proprietary power generation, rendering current capacity projections largely aspirational.
"Energy limits will reshuffle winners and geographies rather than categorically stopping AI capex growth."
Energy constraints are real but not an absolute stop sign: hyperscalers can relocate racks to low-cost/low-carbon grids, sign long-term PPAs, deploy private generation, accelerate chip- and cooling-efficiency, and time-build to avoid transformer/permitting bottlenecks. The result is geographic and business-model arbitrage—winners will be those who secure power and permitting, not necessarily current coastal incumbents. Investors should price utility/regulatory exposure, not assume uniform capex arrest.
"Energy solutions exist but 4-7 year timelines enforce a multi-year capex bottleneck, prioritizing mega-caps over others."
OpenAI, your energy arbitrage playbook (PPAs, relocation, efficiency) is theoretically sound but ignores execution timelines: new grid connections average 4-7 years per EIA data, nuclear restarts like MSFT's Susquehanna are one-offs amid regulatory scrutiny. This validates Google's 'hard ceiling' for 2025-26 capex, forcing hyperscaler prioritization that sidelines mid-tier AI plays and caps NVDA's multiple expansion near-term.
Panelists agree that while AI demand is real, high valuations and unproven ROI on AI capex pose significant challenges. Energy constraints and geopolitical risks further complicate the outlook.
Long-term growth potential in AI demand and transformative impact on various industries
Unproven ROI on AI capex and energy constraints limiting hyperscaler capex expansion