NEW YORK, Feb 7, 2026, 03:07 (EST)
- Amazon’s $200 billion capital expenditures outlined for 2026 reignited investor questions—how soon will AI dollars translate into actual profit?
- Amazon, Alphabet, Microsoft and Meta together have pointed to over $630 billion in 2026 spending plans linked to AI and data centers.
- Chip stocks jumped, lifted by the surge in spending, but software and data companies remained weighed down as worries over AI-related upheaval lingered.
Amazon.com shares slipped Friday, with investors uneasy after the company unveiled a roughly $200 billion push to beef up its artificial intelligence (AI) infrastructure. The hefty price tag behind Amazon’s AI ambitions raised fresh doubts for some shareholders.
Amazon, Alphabet, Microsoft and Meta have now committed over $630 billion in capex—spending that covers everything from servers and chips to sprawling data centers, and most of it is getting funneled into AI projects. Morgan Stanley analysts point out that investors aren’t cutting tech giants any slack for these hefty outlays unless there’s a visible path to strong returns on invested capital.
According to Bloomberg data, the combined total for 2026 lands around $650 billion, a figure that highlights just how massive the build-out is getting. Investors are watching closely, and they’re not hesitating to hit the brakes on spending plans that lack a direct line to earnings growth.
Amazon flagged a capex plan that points to a jump of more than 50% this year, catching some analysts off guard. The number landed well above market expectations, according to several on the Street. “The magnitude of the spend is materially greater than consensus expected,” wrote MoffettNathanson analysts. https://www.reuters.com/business/retail-co…
Big spending announcements sent tech names sharply in different directions. Nvidia soared 7.8% Friday, AMD popped 8.3%, Broadcom was up 7.1%. The Dow, meanwhile, punched through 50,000 for the first time ever. “There’s real demand for AI products,” said Ross Mayfield, investment strategy analyst at Baird. https://www.reuters.com/business/futures-s…
Timing is part of the worry here. The checks get sent out up front, but investors have become impatient with the largest platforms—revenue growth has to match up, or there’s little forgiveness.
Bernstein’s Mark Shmulik isn’t seeing investors sign off on massive spending just because companies say so. “Investors needed more than promises to underwrite this story,” he wrote after listening to Amazon’s earnings call. BNP Paribas analyst Nick Jones, for his part, described Alphabet’s capex as “rational” in light of current demand. https://www.businessinsider.com/amazon-goo…
Software stocks have taken a hit as anxiety over AI’s impact on pricing grows. The S&P 500 software and services index tumbled 4.6% Thursday, erasing roughly $1 trillion in value since Jan. 28—a rout some are calling “software-mageddon.” “A sell-everything mindset,” is how Dave Harrison Smith at Bailard put it. Near-term earnings, said Goldman Sachs strategist Ben Snider, will serve as “important signals of business resilience.” https://www.reuters.com/business/us-softwa…
Data and analytics names have faced steady selling, with investors worried that AI could eat into services they currently monetize. “The market’s viewpoint is that the AI build-out trade … got too pricey,” said Andrew Wells, chief investment officer at SanJac Alpha. He framed the move as de-risking. Nvidia CEO Jensen Huang described demand as “sky-high,” but St. James’s Place equity strategist Carlota Estragues Lopez noted that investors are approaching AI coverage “far more cautiously.” https://www.reuters.com/business/global-so…
There’s another view emerging among analysts: hyperscalers — the heavyweight cloud companies with sprawling data centers — are ditching their old asset-light playbook. Now, they’re leaning into bigger capital outlays, and the margin for error just keeps shrinking.
Still, this trade isn’t set in stone. Should the AI push end up boosting cloud sales and driving up prices, spending could be seen as investment instead of weighing things down. But if demand stalls—or if too much capacity comes online too soon—dot-com era analogies may stick around in the market chatter.
This week, a Reuters Open Interest column pointed out that the old “rising tide” in AI stocks isn’t holding up—semiconductors are off to the races, but software names are taking a hit. Oracle and ServiceNow are among those getting caught on the wrong side of the split. https://www.reuters.com/commentary/reuters…
