The Linux Foundation hosted a virtual analyst summit on October 8. It introduced TCA as a new way to measure AI costs beyond tokens. TCA covers infrastructure, energy, capital, software, licensing, and labor. The summit argued AI has a value problem more than a cost problem. Multimodel AI makes economics an architecture decision for most firms. Model routing is now a key lever for cost, performance, risk, and value. FinOps, ITAM, ITFM, and tokenomics are converging into AI value management. FOCUS is emerging as the financial data layer for AI cost normalization. CFOs are increasingly involved in AI cost and value discussions. The next step is mapping AI costs to the general ledger.
What changed
Enterprises measured AI spend mainly by token costs and infrastructure ownership.
What it unlocks
Connecting AI production, consumption, and cost to measurable business value.
- 10–25% of total AI spend on tokens
- 472 respondents in State of Tokenomics 2026
- $4.6T in respondent revenue
- 96% use frontier models
Sources