The $650 Billion Question: Why Tech Giants' Massive AI Spending Is Rewarding Completely Different Stocks

While Amazon and Microsoft hemorrhage market value over infrastructure costs, an unexpected beneficiary emerges: the humble utility company

Prepared by: Liam Boggan, Quantmatix CEO and Founder, Head of Research

The S&P 500 limped through another week of losses on Friday, closing down 0.10% despite a late rally, as investors grappled with a staggering revelation: America’s biggest technology companies plan to spend nearly $650 billion this year on artificial intelligence infrastructure.

Amazon, Microsoft, Alphabet, and Meta collectively unveiled capital expenditure plans that would have seemed fantastical just 18 months ago. Amazon alone committed $200 billion—a figure that triggered an 11% stock plunge and contributed to a $900 billion wipeout across the trio of Amazon, Google, and Microsoft.

Yet beneath this headline chaos, something fascinating is happening in the market’s plumbing. The companies expected to benefit most from AI spending—semiconductor manufacturers and cloud computing giants—are stumbling. Meanwhile, an entirely different set of winners is emerging, and they’re not the names that dominate Silicon Valley cocktail conversations.

The Infrastructure Paradox

Consider this: While Nvidia surged 8% on Friday (ahead of its February 25th earnings), semiconductor equipment makers are flashing warning signs that systematic momentum analysts track religiously. Micron Technology, LAM Research, and KLA Corp—the companies that manufacture the tools to make AI chips—are all showing what market technicians call “exhaustion patterns.” High institutional ownership meeting decelerating momentum.

AMD illustrated the divergence perfectly. The company posted 34% revenue growth and 39% gains in its data center business—numbers that would have sparked champagne corks in any other quarter. Instead, shares cratered 17% because the forecast didn’t match expectations that had become, quite literally, priced to perfection.

Microsoft, which received a rare downgrade to “Hold” from Stifel, exemplifies the paradox. The company beat earnings expectations with $81.27 billion in revenue, yet sold off over 11% as analyst Brad Reback warned that Azure cloud supply constraints would persist “through at least June 2026”—contradicting the “perfect operating environment” thesis that had propelled the stock higher.

Enter the Utilities

While Big Tech wrestles with whether $650 billion in spending will generate adequate returns, a quieter story is unfolding in America’s electric utilities sector. US data center power demand is projected to reach 106 gigawatts by 2035—a 36% increase from estimates published just seven months ago. Nearly a quarter of new data center projects now exceed 500 megawatts, double last year’s share.

This represents the largest acceleration in American electricity demand in over two decades, and it’s transforming utilities from defensive backwaters into growth stocks.

Consider Xcel Energy, worth noting as an example of this shift. The company isn’t developing chatbots or training large language models. It’s doing something far more mundane and, currently, far more valuable to institutional investors: providing the gigawatts of firm capacity that hyperscalers need to keep their AI ambitions from literally overheating.

“The AI boom is creating a once-in-a-generation shift in power demand,” explains the research note from Quantmatix, a systematic momentum analytics platform. “What we’re tracking is institutions rotating capital into utilities experiencing genuine inflection from historically low positioning, while simultaneously reducing exposure to semiconductors showing exhaustion despite strong fundamentals.”

The divergence is stark. While utilities demonstrate what analysts call “bullish inflection surge”—rapid momentum building from a low base—semiconductors face “bullish exhaustion”—elevated positioning meeting deceleration. It’s not about the quality of the companies. It’s about where institutional flows are moving.

The Winners Nobody Expected

Some examples worth noting illustrate the rotation. Pure Storage, which makes enterprise data storage systems, is gaining traction as hyperscale data centers prioritize power efficiency in their AI inference operations. The company’s DirectFlash hardware offers superior cooling efficiency—critical when Azure capacity constraints are forcing Microsoft to ration access in key US regions.

Roblox, oddly enough, surged almost 10% on Friday, decoupled entirely from Big Tech’s infrastructure burden. Netflix continues demonstrating margin expansion without the capital expenditure anchor weighing down Meta (which announced up to $135 billion in 2026 spending) or Alphabet ($175-$185 billion projected).

Even in consumer staples, momentum is building. Campbell Soup isn’t exciting. But in a market where the Nasdaq has dropped 4% over the past week and investors watched $900 billion evaporate from three mega-cap names, “not exciting” starts looking rather attractive.

What Delayed Data Means

Adding to market uncertainty, critical economic data remains in limbo. Non-Farm Payrolls and CPI inflation reports, originally scheduled for February 6th, were delayed until February 11th and 13th due to the federal government shutdown. This created an information vacuum forcing investors to focus intensely on corporate earnings and, crucially, those staggering capital expenditure announcements.

The labor market isn’t helping sentiment. Weekly initial jobless claims jumped to 231,000 from 209,000 the previous week, reinforcing concerns that employment momentum is cooling precisely as tech giants embark on the largest corporate spending spree this century.

The Broadening Market

“Leadership is no longer concentrated in a few mega-caps,” notes Friday’s market analysis. “We’re seeing momentum strengthen in industrials and consumer staples—a healthier, more diversified environment as capital rotates out of technology concentration.”

This isn’t your typical defensive rotation, where investors flee to safety during downturns. This is something more nuanced: institutions repositioning for a market where returns on $650 billion in AI infrastructure spending remain uncertain, but the 106 gigawatts of power demand to support it is contractual and measurable.

Companies like WR Berkley in financials or AutoZone in retail aren’t making AI breakthroughs. They’re simply showing consistent momentum characteristics while their tech counterparts grapple with questions about whether training increasingly sophisticated AI models will generate proportional revenue increases.

The Valuation Reckoning

Meta’s stock initially rallied 10% after its earnings announcement, demonstrating that strong guidance can offset capex concerns—but those gains evaporated as broader tech weakness dragged the sector down. Alphabet’s Gemini App boasts 750 million monthly active users, yet the stock fell 3% on its spending guidance.

The market is delivering a clear message: Show us the returns, not just the expenditures.

Meanwhile, sectors demonstrating actual momentum shifts—not just narrative promise—are attracting systematic capital flows. Utilities aren’t promising to revolutionize human-computer interaction. They’re signing multi-year agreements for gigawatts of firm capacity. The difference between speculation and contract is becoming material to institutional allocators.

What This Means for Investors

The rotation creates unusual opportunities for those paying attention. Technology sector momentum has fractured internally—some pockets like industrial technology and enterprise software efficiency plays (Pure Storage, for instance) are strengthening, while broad semiconductor exposure is weakening despite compelling long-term fundamentals.

This dispersion rewards precision over passivity. Owning the Nasdaq or a broad tech ETF captures both the weakness in exhausted names and the strength in selective winners. The gap between them is widening.

European utilities face similar dynamics—capital expenditure reaching 164% of EBITDA to support grid modernization—while European banks, despite stellar 2025 performance, show similar exhaustion patterns to their US counterparts.

The Bottom Line

The S&P 500’s third weekly decline in four weeks masks a more profound shift: After years where simply owning mega-cap technology generated outperformance, the market is demanding differentiation. Amazon’s 11% Friday plunge on a $200 billion capex announcement versus Apple’s resilience (despite lower AI spending) tells the story.

Momentum is rotating from companies promising AI transformation to companies providing the infrastructure enabling it, and from those infrastructure providers to companies demonstrating stable earnings growth without the capital intensity.

As one analyst tracking the Big Tech spending spree put it: “The skepticism is probably healthier than any previous cycle I’ve seen.” Investors aren’t rejecting AI’s potential. They’re insisting on proof that $650 billion in spending will generate proportional returns.

Until that proof arrives, the market is rewarding stability, contracts, and measurable demand over narrative and promise. It’s why utility companies—yes, utility companies—are experiencing genuine momentum inflections while semiconductor equipment makers flash exhaustion warnings.

In a year where Nvidia, Alphabet, and Tesla outperformed the S&P 500 by substantial margins while Microsoft, Meta, Amazon, and Apple lagged, 2026 is shaping up as the year the market stops rewarding AI spending and starts demanding AI returns.

The $650 billion question is whether Big Tech can deliver them.