AI Capex May Hit $820 Billion in 2026. The Funding Spigot Is the Real Trade.

The 2026 AI buildout is already measured in hundreds of billions

This is a liquidity story first and a valuation debate second. UBS now sees $820 billion in 2026 AI capital expenditure, while independent models put the top hyperscalers at about $602 billion in 2026 capex to $660 billion to $690 billion. Independent estimates also suggest roughly 75% of aggregate hyperscaler capex is AI-linked, which puts AI-specific spending around $450 billion. That is a massive concentration of spend on a small group of balance sheets.

The new catalyst is funding, not just spending plans. Alphabet has signaled that equity is back in the toolkit, with an $80 billion equity offering that includes a $10 billion Berkshire Hathaway investment. That shifts the market read. This is no longer only about who can borrow more; it is about who can bring in fresh equity and keep deploying through the cycle.

The bull-bear split is sharper too. Bulls still expect AI cloud infrastructure to remain capacity-constrained. Bears argue demand is too narrow, pointing to the view that four customers contributed 61% of Nvidia's revenue reflects concentrated dependency rather than broad enterprise payback. If funding keeps widening before adoption broadens, investors may keep rewarding spenders first and demand economic proof later.

Big Tech's funding mix is changing again

Debt has done its job

Reuters has covered Big Tech's hefty capital spending alongside strong results, and Invesco said hyperscalers want diverse sources of capital because the buildout is so large. That has leaned on bonds for a while, with expectations for robust issuance in 2026 across investment-grade credit, project finance, and related structures. But as financing needs grow, the next dollar becomes less about standard leverage metrics and more about the full capital structure.

That is why Alphabet's move matters. The company is targeting $80 billion in equity offerings, including a $10 billion Berkshire Hathaway private placement. Equity does not create a new cash-flow stream on its own, but it can create more room to keep spending before monetization fully broadens. If other hyperscalers follow, the market will care less about who can borrow another notch and more about who can raise, deploy, and earn through the cycle.

The new filter is cash generation and payback

At current intensity, hyperscaler capex is running at 45-57% of revenue. That makes operating cash especially valuable. Investors are no longer judging AI exposure by itself; they are judging whether AI spending is attached to cash engines that can absorb dilution, support financing activity, and eventually show operating leverage.

That is where the differences between companies start to matter. Microsoft, Alphabet, Amazon, and Meta are all pointing to AI-driven demand in cloud, search, or advertising. But Meta lacks the same built-in cushion as firms with a mature cloud segment, and Reuters noted no cloud arm to cushion infrastructure spending, which can make investors uneasy if payoffs take longer than expected.

What investors need to see next: broader demand and clearer returns

The funding story is only half of the trade. Investors also need evidence that spending is leaking into durable revenue streams.

Nvidia's customer concentration still matters

Four customers contributed 61% of Nvidia's revenue. That does not prove the AI thesis is wrong, but it does suggest demand is still concentrated at the top. Bears will argue that is not yet broad enough to confirm a full enterprise AI economy.

If spending remains trapped among a few giants, the market can keep financing the buildout for a while and then reprice the story quickly. The key question is whether AI demand expands beyond infrastructure buyers into wider enterprise adoption.

Reuters says investors are now asking whether the spend will pay off

Reuters said investors want to know whether cloud computing and advertising are growing fast enough to justify the outlays. That is the right lens.

Microsoft is a useful near-term scoreboard. Management is already tying returns to AI powering demand across cloud, advertising and search. If cloud growth continues to strengthen and those AI-linked products keep expanding, monetization is moving from narrative toward cash.

What to watch

• Trigger: cloud growth starts supporting margins, not just headline revenue.
• Trigger: software adoption shows up through sustained usage, not just pilot announcements.
• Watchpoint: if more companies turn to equity as hyperscaler funding shifts, volatility can rise before payback broadens.
• Watchpoint: if core ad and cloud engines weaken while capex stays elevated, the current setup becomes harder to defend.

That is why this matters now. The next few quarters should show whether AI is becoming a profit center or remains primarily a capital appetite.