Tech companies seeking better control of spending on AI
Tech companies are moving away from unchecked AI spending and adopting more disciplined approaches like "tokenmaxxing" and "tokenomics" to better control costs and measure returns on their AI investments.
Background
- "Tokenmaxxing" and "tokenomics" are playful industry terms referring to the race to spend as much as possible on AI compute (cloud computing resources like GPUs) and the questionable business models that have arisen around it.
- Many AI startups and even large tech firms have been judged not by revenue or profit, but by how aggressively they spend on Nvidia GPUs and cloud compute — a metric critics call "tokenmaxxing."
- "Tokenomics" blends "token" (AI's basic unit of text/data processing) and "economics" — it describes the hype-driven belief that spending massively on compute will eventually pay off, even without a clear path to profitability.
- This article reports a shift: investors and tech leaders are becoming skeptical of unlimited AI spending and are demanding measurable returns, a major departure from the "spend at all costs" mindset of the past two years.