Show Me the Money

Meta surges, Microsoft craters, and AI’s era of patient capital comes to an end

Last week delivered a verdict that has been building for two years. After hundreds of billions poured into artificial intelligence infrastructure, Wall Street finally demanded receipts.

The results were stark. On Wednesday evening, Meta and Microsoft both reported earnings, both beat expectations, and both announced plans to spend staggering sums on AI. The market’s response could not have been more different: Meta’s stock jumped 10%, while Microsoft suffered its worst single-day drop since March 2020, shedding roughly $357 billion in market value.

The divergence reveals something important about where the AI boom stands as it enters 2026: the era of patient capital is ending. Investors are no longer satisfied with promises that returns will materialize eventually. They want to see the money.

The Tale of Two Earnings

The numbers tell one story; the market’s reaction tells another.

Microsoft reported revenue of $81.3 billion, up 17% year over year. Earnings per share came in at $5.16, comfortably above estimates. By any conventional measure, an excellent quarter. The problem lay in the details. Azure revenue growth slowed to 39%, down from 40% the prior quarter, and the company guided for further deceleration to 37-38% in the current period. Meanwhile, capital expenditures surged 66% to a record $37.5 billion.

The math troubled investors. Microsoft is spending faster than its cloud business is growing. CFO Amy Hood explained that a meaningful chunk of new capacity is going to internal teams supporting products like Copilot rather than external Azure customers, but this did little to soothe concerns. If anything, it raised questions about whether Microsoft’s AI investments are cannibalizing its own cloud growth story.

Meta, by contrast, posted $59.89 billion in revenue, up 24%, with earnings per share of $8.88. Like Microsoft, Meta announced plans for massive spending, guiding capital expenditures of $115 billion to $135 billion for 2026, nearly double last year’s $72 billion. The market cheered anyway.

The difference is straightforward. Meta’s AI investments are already generating returns in its core advertising business. Ad impressions rose 18% in the quarter while average price per ad increased 6%, a combination that suggests Meta’s AI-driven recommendation systems are delivering more valuable inventory. When Mark Zuckerberg talks about pursuing “personal superintelligence,” investors can look at the income statement and see evidence that AI improvements are translating into revenue today.

Microsoft’s position is more complicated. Its AI story depends heavily on enterprise adoption of tools like Copilot and the continued growth of Azure, both of which face longer sales cycles and capacity constraints the company admits will persist through at least June. The $250 billion OpenAI deal that accounted for 45% of Microsoft’s customer backlog is impressive, but it also concentrates risk in a single relationship with a company that remains unprofitable and perpetually raising money.

The Circular Logic of AI Capitalism

Which brings us to the week’s other major development: the unraveling of Nvidia’s $100 billion OpenAI deal.

In September, Nvidia and OpenAI announced with considerable fanfare a “landmark strategic partnership” in which Nvidia would invest up to $100 billion as OpenAI deployed new data centers built with Nvidia components. The arrangement was unusual, to put it mildly. Nvidia would invest in OpenAI, which would use the funds to buy Nvidia’s chips. Critics immediately noted the circular logic, but both companies brushed off concerns.

Last Friday, The Wall Street Journal reported that the deal had collapsed. Sources said some inside Nvidia had expressed doubts about the transaction, and CEO Jensen Huang had privately criticized what he described as a “lack of discipline” in OpenAI’s business approach. Nvidia confirmed it would still participate in OpenAI’s current funding round, with Huang calling it potentially “the largest investment we’ve ever made”, but clarified the amount would be “nothing like” $100 billion.

The timing is notable. Days before the Journal report, multiple outlets reported that Nvidia, Microsoft, and Amazon were in discussions to invest up to $60 billion in OpenAI as part of a round that could reach $100 billion. The latest figures suggest Nvidia might contribute up to $30 billion, Amazon somewhere between $10 billion and $20 billion, and Microsoft less than $10 billion. SoftBank is reportedly considering another $30 billion.

Even at the reduced scale, the numbers are staggering. OpenAI is seeking to raise $100 billion at a valuation of approximately $830 billion, for a company that reportedly lost money last year and whose primary revenue source, ChatGPT subscriptions and API access, faces increasing competition from open-source alternatives and well-funded rivals.

The circular dynamics that troubled critics of the original Nvidia deal apply equally to the new funding structure. Nvidia invests in OpenAI, which buys Nvidia chips. Amazon invests in OpenAI, which signs deals to run workloads on AWS. Microsoft invests in OpenAI, which accounts for 45% of Microsoft’s cloud backlog. The money flows in circles, flattering demand metrics at each stop without necessarily demonstrating that end users are paying enough to sustain the whole edifice.

The Capex Arms Race

What makes this moment particularly fraught is the scale of spending the major players have committed. Microsoft, Meta, Alphabet, and Amazon are collectively expected to spend roughly $475 billion on capital expenditures in 2026, more than double the $230 billion they spent in 2024. Global AI infrastructure spending is projected to exceed $1.3 trillion this year.

The companies insist this spending is necessary. Microsoft’s Satya Nadella said the company added nearly a gigawatt of computing capacity in the second quarter alone and still faces constraints that will last at least until June. Meta’s Zuckerberg argued that being the most efficient company in building AI infrastructure will become a strategic advantage.

But the Microsoft earnings reaction suggests investors are growing skeptical that returns will materialize at a pace commensurate with the spending. As Wedbush analyst Dan Ives noted, “They wanted to see less capex spending and faster cloud/AI monetization, and coming out of the gates, it’s the opposite.”