Big Tech Is Spending $650 Billion on AI in 2026 — And Wall Street Is Getting Nervous

$650 billion. In one year. From four companies. That’s not a typo. Alphabet, Amazon, Meta, and Microsoft have collectively committed to a capital expenditure figure so large it would rank as one of the biggest single-year investments by any group in modern economic history. And according to Bridgewater Associates, the world’s largest hedge fund, the AI boom has just entered its most dangerous phase yet.

This isn’t hype. It’s hard numbers from earnings calls, investor letters, and balance sheets.

The Breakdown: Who’s Spending What

Amazon leads the group with $200 billion in planned capital expenditure for 2026, mostly earmarked for AWS to handle surging AI workloads. Alphabet follows with a forecast of $175 to $185 billion, while Meta plans $115 to $135 billion, and Microsoft’s annual run rate puts it on pace for roughly $145 billion. The White House

Put it side by side:

At the high end, the group would spend around $665 billion — a 74% jump from the $381 billion spent in 2025. The vast majority is going to AI chips, servers, and data center infrastructure. The White House

That last line matters. This isn’t R&D spending on moonshot ideas. It’s concrete, physical infrastructure — land, power, silicon, steel.

Why Bridgewater Is Waving a Red Flag

In a letter to clients, Bridgewater co-CIO Greg Jensen said the AI boom has entered a “more dangerous phase,” marked by exponentially rising investments in physical infrastructure and growing reliance on outside capital. “Compute demand continues to significantly outpace supply, driving hyperscalers to invest even more rapidly to try to someday get ahead of the demand.” GovInfoSecurity

Jensen didn’t say the AI boom is a bubble. But he said it’s starting to rhyme with past cycles that ended badly. He compared the situation to the Dot-com bubble, though he emphasized that current conditions remain far smaller in scale. Even so, the combination of high expectations and massive investment commitments introduces notable risks. International Economic Development Council

The word to focus on there is reliance on outside capital. When companies stop funding growth from their own cash flow and start depending on debt markets or equity dilution, vulnerability increases. Any macro shock — rising rates, a market correction, an AI product that flops — hits much harder.

The Free Cash Flow Problem Nobody Wants to Talk About

Here’s the uncomfortable math. Analyst projections warn that Big Tech free cash flow could drop up to 90% in 2026 as capital expenditure outpaces revenue growth from AI spending. White House

Let that land. Ninety percent.

The four companies have already curbed share buybacks more aggressively to help fund the surge in capital expenditure. GovInfoSecurity For investors who bought these stocks expecting steady buybacks and dividends, that’s a meaningful shift in the value proposition. You’re no longer investing in cash-returning machines. You’re investing in a very expensive bet on AI monetization.

Amazon stock fell more than 8% on Friday following its announcement. Alphabet shares fell 3%, and Microsoft stock fell over 11% after its quarterly results. The White House Meta was the exception — its stock actually rallied, because AI is visibly boosting its ad revenue right now.

That split tells the story. The market is rewarding companies that can show today’s return on AI investment, and punishing those asking for patience.

What Are They Actually Building?

Think of this as the largest infrastructure buildout since the interstate highway system. Meta has started construction on a new 1 gigawatt data center campus in Lebanon, Indiana — calling it one of the company’s largest infrastructure investments at more than $10 billion. A gigawatt is not a rounding error. It’s a location decision. The White House

The surge in spending is primarily directed toward advanced data centers, specialized chips, and liquid-cooling systems to power the next generation of generative AI. White House

Liquid cooling alone is now a multi-billion dollar procurement category. The physical scale of what’s being built is genuinely hard to visualize.

Google Cloud revenue grew 48% year-on-year to $17.7 billion in Q4 2025, with Gemini models boosting profitability. Microsoft is targeting $25 billion in AI-related revenue by end of FY26, driven by Copilot and Azure AI uptake. White House

So the revenue side is growing. The question is whether it can grow fast enough to justify the spend.

Macro Impact: GDP Boost, Inflation Risk, and a Jobs Caveat

This level of tech investment has real-world macroeconomic consequences — both good and concerning.

Bridgewater estimates that technology investment contributed roughly 50 basis points to U.S. GDP growth in 2025 and expects that figure to rise to around 100 basis points in 2026 as AI-related spending accelerates. International Economic Development Council That’s significant. AI capex is now a meaningful driver of U.S. economic growth, not just a sector story.

But there are caveats. A $1.5 billion data center project may support roughly 100 long-term jobs, compared with more than 1,600 jobs from a similarly sized manufacturing plant. Bridgewater notes that AI-driven investment creates far fewer employees per million dollars of GDP than most other sectors. Holland & Knight

And on inflation: Increased demand for hardware and data-center capacity may push up prices for technology equipment and electricity in certain regions. International Economic Development Council In some U.S. markets, data center power consumption is already straining local grids and pushing up electricity costs for residents and businesses nearby.

OpenAI and Anthropic: The Pressure Is On

The $650 billion being poured in at the infrastructure layer creates enormous pressure upstream. Jensen highlighted that leading AI developers such as Anthropic and OpenAI will require major product breakthroughs to secure the final rounds of funding needed before potential IPOs. Without clear paths to outsized profits, these firms may struggle to justify their valuations and capital needs. International Economic Development Council

This is the part of the story that gets glossed over. Hyperscalers can absorb capex write-downs because they have cloud revenue, advertising revenue, e-commerce revenue — multiple fallback lines. Pure-play AI labs don’t have that cushion. They need the models to be transformatively valuable, and soon.

The window between “promising AI startup” and “revenue-generating platform” is narrowing as investor patience shrinks.

Is This a Bubble? The Honest Answer

Gartner VP analyst Ewan McIntyre says the $650 billion commitment is emblematic of what he calls the “Intelligence Supercycle — a period of disruptive investment realigning markets. But the real challenge isn’t adoption; it’s meaningful value creation.” White House

Shai Luft, COO of Bench Media, offers a counterpoint: “$650 billion represents roughly 15 to 20% of the combined annual revenue of these four companies. The bigger risk isn’t overconfidence — it’s under-investing and leaving core products like search, cloud, and advertising exposed to disruption.” White House

Both views are reasonable. This is what a genuine technology transition looks like from the inside — enormous capital commitment, unclear near-term returns, and deep strategic conviction that the alternative (falling behind) is worse than the risk of overspending.

What’s Next

Bridgewater projects AI capex could boost GDP by roughly 140 basis points in 2026 and 150 basis points in 2027 — levels comparable to business investment contributions seen during the late 1990s tech boom. Holland & Knight

Watch these developments over the next 12 months: Q2 and Q3 earnings calls will be the first real test of whether AI revenue is keeping pace with capex. OpenAI’s IPO timeline will signal whether the market still has appetite for AI valuations. Power infrastructure bottlenecks in Virginia, Texas, and the UK are emerging as the next constraint on data center expansion. And if any major AI product launches significantly underperform — a real possibility — expect the market narrative to shift fast.

The $650 billion bet is placed. Now we wait to see what it buys.

Sources: Bridgewater Associates client letter, Bloomberg, Reuters, company Q4 2025 earnings calls.

Check Latest News About White House Launches “American AI Stack” to Dominate Global AI Race and Push Back Against China