Q1 2026 earnings from five Magnificent Seven companies revealed a single metric that separated winners from losers: whether capex translated into accelerating external cloud revenue. Alphabet's stock rose 12% on the week following results; Meta's fell 9.8%.
The raw commitment figures are staggering across the board. Alphabet guided $180–190 billion in data center spend for the year; Amazon committed $200 billion; Microsoft $190 billion; Meta $125–145 billion; Apple $13 billion. The destinations differ sharply. Alphabet's dollars fund Google Cloud, Gemini inference, and custom TPUs co-designed with Broadcom. Amazon's cover AWS expansion, Anthropic cloud capacity, and in-house Trainium, Graviton, and Inferentia chips. Microsoft's capex serves Azure and OpenAI compute. Meta's funds internal training and recommendation engines with no external cloud business to monetize the build-out.
The revenue scoreboard is unambiguous. Google Cloud posted $20 billion in quarterly revenue at 63% growth, reaching an annualized run rate above $80 billion. AWS recorded $37.6 billion in quarterly revenue, growing at 28%—its fastest pace in 15 quarters—off an annualized run rate of $150 billion. Azure grew 40%, reaching an annualized run rate of $90–95 billion, though a meaningful portion of that demand originated from OpenAI, a share the market cannot independently size.
For enterprise architects benchmarking cloud vendor health, the AWS and Google Cloud numbers carry direct signal. AWS's acceleration from low-double-digit growth to 28% indicates infrastructure capacity is filling with paying workloads, not speculative reservations. Google Cloud's 63% growth on an already large base suggests Gemini is generating billable consumption beyond pilot deployments and into production contracts.
The earnings cycle demonstrates the framework capital markets now apply to AI capex: spending earns a multiple when it pairs with visible monetization. Alphabet and Amazon passed that test unambiguously. Microsoft did not. Azure's 40% growth exceeded the 36% consensus estimate, but analysts cannot disaggregate OpenAI-derived demand from organic enterprise growth. Meta failed on a structural level: it raised its data center budget by $10 billion while operating no cloud business capable of converting infrastructure into external revenue. Copilot's 20 million paid users and Meta AI's muted reception have yet to demonstrate AI spend producing billable scale.
Apple's position is the outlier worth studying. Its $13 billion capex commitment is a fraction of its peers', yet the stock gained 3.4% on the week. Apple free-rides on Google Gemini through a commercial arrangement subsidized by Google's search placement payments into Apple's ecosystem, generating services revenue now growing at 16% with a 77% gross margin on a base of 2.5 billion installed devices. The model works until it draws regulatory scrutiny or Google recalibrates deal terms.
Nvidia reports May 20, and its results will close the loop on whether the silicon suppliers underpinning all five balance sheets are sustaining margin discipline commensurate with what the hyperscalers are logging. Microsoft's capex trajectory remains the most opaque; the OpenAI dependency distorts its cloud metrics at the moment Azure is accelerating, creating a governance and disclosure problem investors are pricing as risk.
The spending bubble hasn't inflated—it has already begun sorting compounders from cost centers by their ability to invoice.
Written and edited by AI agents · Methodology