THE MONEY
BEHIND AI.

Venture capital and corporate investment, government spending, hyperscaler capex, market forecasts, spending by industry, the fastest-scaling AI-native companies, and the LLM market’s competitive landscape. Every figure primary-sourced and dated.

Thank you for reading this post, don't forget to subscribe!

The state of play: AI economics in 2026

The defining feature of AI in 2026 is capital intensity. Gartner projects worldwide AI spending will reach $2.52 trillion this year, up 44% year over year, with AI infrastructure alone accounting for roughly $1.37 trillion of that total. This build-out rests on a foundation laid in 2024, when US private AI investment hit $109.1 billion — nearly twelve times China’s — according to Stanford HAI’s 2025 AI Index.

The model layer, meanwhile, is consolidating. In Menlo Ventures’ 2025 enterprise survey, Anthropic led LLM API spend at 40%, overtaking OpenAI at 27%, with Google third at 21% — together close to 90% of the market. Yet unit economics are moving the opposite direction: Stanford HAI found the cost of GPT-3.5-level inference fell roughly 280-fold in 18 months, from $20 to $0.07 per million tokens.

The through-line is a market betting that collapsing token costs expand demand faster than they erode margins — spending rises even as the price of intelligence falls.

• Infrastructure is the spend. Gartner puts AI-optimized hardware and data centers at more than half of 2026’s $2.52T outlay.
• It is a three-horse race. Anthropic, OpenAI and Google together control roughly 88% of enterprise LLM API spend (Menlo Ventures, 2025).
• Cost per token is in free fall. A ~280x drop in inference cost in 18 months resets every ROI assumption (Stanford HAI, 2025).