AI sticker shock hits corporate America
Corporate AI experiments are running into “sticker shock” as enterprises realize soaring token and infrastructure bills aren’t matched by clear productivity gains or revenue. Commenters describe perverse incentives like “token leaderboards,” pressure to overuse tools such as Claude or Copilot, and chaotic agent deployments that can burn vast amounts of compute with little business value, while deeper issues—poor governance, bad KPIs, and misaligned executive incentives—go unaddressed. Many see this as an early phase of an AI hype correction, where companies must learn to target specific high‑value use cases, fix data and process plumbing, and stop treating AI as a magic replacement for workers.
AI Spending, Token Costs, and “Tokenmaxxing”
- Many commenters argue enterprises pushed indiscriminate AI usage (“tokenmaxxing”), sometimes with leaderboards and performance pressure tied to token consumption.
- Some claim low token usage has even contributed to performance improvement plans or layoffs; others doubt this is widespread, calling it exaggerated or anecdotal.
- High per-token enterprise pricing and a reported incident of a client burning hundreds of millions in a month via agents/automations are seen as emblematic of poor governance and lack of limits.
- Vendors themselves are now giving talks on “token optimization,” which some see as incompatible with promised productivity gains.
Productivity vs Real Outcomes
- Multiple comments note that more code, PRs, and deployments are not translating into faster milestones or better outcomes.
- Goodhart’s law is invoked: once activity metrics are optimized, they stop correlating with value.
- Several engineers say AI helps for boilerplate, translations, tests, and small utilities, but often fails on deeper refactors or complex design, making manual work faster.
- In large organizations, coding is a small fraction of project time (requirements, alignment, approvals dominate), so even big coding speedups yield modest overall gains.
Corporate Governance, Layoffs, and Incentives
- Strong skepticism that AI is being used to genuinely improve productivity rather than justify layoffs, boost executive compensation, or signal “innovation” to markets.
- Commenters highlight disconnects between executive incentives and company or worker outcomes, referencing principal–agent problems, wage theft comparisons, and “dictatorial” corporate structures.
- There's frustration that AI is being pushed hardest on workers, not on high-cost executive roles.
ROI Uncertainty and Bubble Concerns
- Some see early AI spending as classic hype-cycle behavior: massive capex, vague strategies, weak measurable ROI.
- A multi-step “AI bubble” scenario is sketched: universal adoption pressure, then realization of limited fit, earnings disappointments, token budget cuts, data-center pullbacks, and broader market fallout.
- Others say failure stories are still thin and that AI clearly speeds up capable individuals; the problem is misaligned use cases and immature organizational plumbing.
Energy, Environment, and Externalities
- High token burn is criticized as wasteful in energy and environmental terms, likened to or worse than earlier crypto excesses.
- Some argue token pricing still underestimates true societal and environmental costs.