AI Spending Catches Up With Oil & Gas — and Wall Street Launches Compute Futures

According to Goldman Sachs projections circulated in late August 2026, Big Tech's AI capital expenditure will reach roughly $680 billion in 2026 — a 78% year-over-year increase — with total AI capex across all firms projected at $765 billion. That $680 billion figure matches, for the first time, the annual capex of the entire global oil & gas industry, a century-old, capital-intensive sector. Investor Chamath Palihapitiya highlighted this symbolic threshold in a post on August 24, 2026.

This capital catch-up comes with new financial infrastructure: CME Group and Silicon Data plan to launch a regulated US futures contract on GPU computing power on October 5, 2026, pending CFTC approval — the first standardized derivative on computing power. ICE, through its Ornn unit, has announced a competing regulated offering. Architect Financial has already been running unregulated GPU perpetual contracts on a Bermuda exchange since January 2026.

The capital intensity now reached by hyperscalers raises questions: Meta's capex stands at 54% of revenue, Microsoft's at 47%, Alphabet's at 46% — levels that in any other capital-intensive industry would already trigger overinvestment warnings. The source article explicitly notes that 'returns on that buildout remain unproven.'

For Cardan-AI clients in energy, O&G, and aerospace & defense, this shift is worth watching: the same institutional investors are now comparing AI compute capex and traditional industrial capex on directly comparable scales — a potential reshuffling of capital allocation to monitor over the next 12-18 months.