For roughly two years, Wall Street directed substantial capital toward any venture featuring “AI” in its business proposal. Currently a more sobering question is taking over trading desks: when does Big Tech AI spending actually start paying for itself? According to a Fortune report, major asset managers think they already know the answer, and it involves a number even bigger than the eye-watering budgets making headlines this earnings season.
Summary
Key takeaways
- Big Tech’s AI-related capital spending is projected between $635 billion and $665 billion by 2026, according to Fortune.
- Global AI investment across every sector is expected to top $1 trillion by 2026.
- Asset managers expect Amazon, Microsoft, Alphabet and Meta to add roughly $340 billion to annual operating cash flow by 2027.
- Janus Henderson’s Richard Clode expects hyperscaler profits and cash flow to outpace incremental capex by 2028.
- Neocloud providers CoreWeave and Nebius have rallied on AI capacity shortages, while AI monetization still needs to grow between 5x and 13x to justify current spending plans.
Big Tech AI spending faces a new scrutiny phase
The scale of the buildout is hard to overstate. Big Tech’s AI-related capital expenditure is projected to land somewhere between $635 billion and $665 billion by 2026, Fortune reported. Zoom out further and the picture gets even bigger: global AI investment across all industries is expected to blow past $1 trillion that same year.
Four companies are effectively writing the checks that define this entire buildout. Amazon, Microsoft, Alphabet and Meta, collectively known as the hyperscalers, are pouring money into data centers, chips and power infrastructure at a pace few industries have ever matched. Recent earnings from Microsoft and Amazon reinforced a familiar theme: demand for cloud services remains intense, and capacity constraints are still limiting how fast these companies can actually grow, even with all that spending.
That combination, massive AI capital expenditure through 2026 paired with demand that still outstrips supply, is exactly why investors are no longer just asking how much these companies are spending. They want to know what comes back.
Why asset managers still bet on a payoff by 2027 and 2028
The bet, according to major asset managers cited by Fortune, indicates that the leading technology firms are positioned to boost their combined yearly operating cash flow by approximately $340 billion through 2027. Meanwhile is the kind of figure that makes even the most aggressive capex plans look defensible, provided the timeline holds.
Richard Clode of Janus Henderson has put a number on the inflection point investors are underwriting: he predicts hyperscalers will grow profits and cash flow faster than their incremental capex by 2028. In other words, the spending curve bends in favor of profitability, just not immediately.
The market’s reaction has been more cautious than that thesis might suggest. Hyperscaler stocks have lagged a 75% rise in the Philadelphia Semiconductor Index during the most recent earnings season, meaning chipmakers, the companies selling the shovels in this particular gold rush, have been rewarded far more generously than the companies actually digging. Yet forward valuations for the principal hyperscalers span from 17.6x at Meta up to 24.6x at Microsoft, each discounts to their historical highs. That gap between spending, expected returns and current valuation is precisely why hyperscaler profits tied to AI have become one of the most closely watched storylines in tech investing right now.
The neocloud squeeze: CoreWeave, Nebius and the capacity shortage
Capacity shortages have created an entirely separate trade. CoreWeave has climbed roughly 50% over recent periods, while Nebius has jumped more than 200%, both benefiting from an AI momentum surge compute shortages and the pricing power that comes with it. These neocloud providers, essentially specialized GPU cloud platforms, have thrived simply because demand for AI computing currently outstrips what’s available.
That advantage carries an expiration date. As hyperscalers bring new capacity online to meet their own cloud services demand for AI, the scarcity that fueled CoreWeave’s and Nebius’s rallies could start to fade. Pricing power built on shortage tends not to survive the shortage ending, and Amazon, Microsoft, Alphabet and Meta are precisely the companies with the balance sheets to close that gap.
Can AI monetization catch up with the spending?
Here’s the number that cuts through the optimism: AI revenue generation must rise between 5x and 13x to support the infrastructure investment strategies these organizations have committed to already announced, according to wealth managers who’ve run the math on what needs to happen for current spending to pencil out. That’s not a rounding error. It’s a wide range that reflects genuine uncertainty about how fast revenue can actually scale.
The bull case rests on timing. Enterprise AI adoption is still early, with most companies running pilots rather than full production deployments. If that shifts, and enterprise customers move from experimentation to actual deployment, the resulting revenue growth could make even a $635 billion capex bill look reasonable in hindsight. If that transition drags on, the gap between spending and monetization stays exposed for longer than investors might prefer.
Either way, the debate over Big Tech AI spending has clearly moved past the “how much” question. The industry, and the market pricing it, is now fixated on a harder one: how long before the spending turns into something investors can actually bank.
FAQ
How much is Big Tech expected to spend on AI infrastructure by 2026?
Big Tech’s AI-related capital spending is projected between $635 billion and $665 billion by 2026, according to Fortune.
What financial benefits are expected from these AI investments?
Major asset managers expect these companies to add about $340 billion to annual operating cash flow by 2027, with profits growing faster than capital expenditure by 2028, according to Janus Henderson’s Richard Clode.
Why have neocloud providers like CoreWeave and Nebius gained so much recently?
These providers have surged due to AI compute capacity shortages and the premium pricing power that comes with strong demand outstripping available supply.
Is enterprise AI adoption mature enough to justify these investments?
Enterprise AI adoption is still early, with most companies running pilots rather than production deployments, which leaves room for future revenue growth but also uncertainty about the timeline.
Article produced with the assistance of artificial intelligence and reviewed by the editorial team.
