The AI race is no longer just a software story. It’s a capital-expenditure arms race—and the price tag is getting huge.
Bridgewater Associates says Alphabet, Amazon, Meta, and Microsoft are on track to collectively invest about $650 billion in AI-related infrastructure in 2026, up sharply from roughly $410 billion in 2025. That jump alone tells you where the industry is now: not in the “experiment” phase, but in the industrial buildout phase.
Why this matters
When spending rises this fast, it changes more than tech company budgets. It affects:
- data center construction
- chip demand
- power infrastructure
- financing markets
- expectations for growth and inflation
Bridgewater’s Greg Jensen reportedly told clients the AI boom has entered a “more dangerous phase”, with investment in physical infrastructure rising exponentially and firms becoming more reliant on outside capital.
That’s the key shift: the story is no longer just “AI is exciting.” It’s “AI now requires a scale of spending that can create system-wide risk if returns disappoint.”
The hidden signal: buybacks are already getting squeezed
One of the more revealing details is that these tech giants have reportedly curbed share buybacks more aggressively to help fund the capex surge.
That matters because buybacks are often how companies return cash to shareholders. If buybacks get dialed down to fund AI infrastructure, management teams are effectively making a big bet:
future AI profits will justify today’s cash burn.
If they’re right, this looks visionary.
If they’re wrong, investors may start asking why so much capital was poured into assets that take years to pay back.
The pressure is spreading beyond Big Tech
Bridgewater’s warning isn’t only about the hyperscalers. Jensen also points to spillover risks for software companies and data providers, especially as AI leaders push hard enough to create “existential” pressure on adjacent sectors. Reuters notes this comes amid a recent selloff in software stocks.
That’s a big market takeaway: AI winners don’t just rise—they can also reprice entire industries around them.
In other words, the AI boom may keep lifting parts of the market while simultaneously damaging business models elsewhere.
Growth tailwind now, inflation headache later?
Bridgewater estimates tech investment added about 50 basis points to U.S. GDP growth in 2025 and could add roughly 100 basis points in 2026.
That’s a major macro claim, and it helps explain why markets have treated AI capex as more than a sector story. But the same spending wave may also push up prices in technology/communications equipment and even electricity in some regions.
So the boom can do two things at once:
- support growth
- create fresh inflation pressure
That’s exactly the kind of combination central banks and investors struggle to price cleanly.
The uncomfortable question: what if the returns don’t arrive fast enough?
Bridgewater’s framing echoes a fear spreading through markets: what happens if AI spending keeps accelerating, but monetization lags?
Reuters reports Jensen warned that a severe stock correction could hit growth and reduce companies’ ability to raise capital, drawing a comparison to the dot-com era (while noting recent moves are much smaller so far).
That doesn’t mean “AI = bubble collapse.” It means the financing loop matters:
- Investors fund the buildout
- Companies spend on infrastructure
- Markets expect breakthroughs and profits
- If those profits lag, funding gets harder
- The whole cycle tightens
And when the cycle involves hundreds of billions of dollars, even a slowdown can have ripple effects far beyond Silicon Valley.
Bottom line
Big Tech’s projected $650 billion AI infrastructure spend in 2026 is a sign of enormous conviction—but also growing fragility. The upside is obvious: faster AI progress, bigger platforms, and a stronger growth impulse. The risk is just as clear: if the economics don’t catch up, this capex boom could become a pressure point for markets, financing, and the broader economy.
