The enterprise AI landscape is undergoing a significant transformation, as the exuberant 'tokenmaxxing' era gives way to a disciplined focus on return on investment. In a recent conversation on Bitcoin World’s Equity podcast and in separate industry discussions, NEA partner Tiffany Luck detailed how companies are recalibrating their AI spending after a period of unchecked experimentation.
Earlier in 2026, tokenmaxxing — encouraging employees to maximise AI tool usage without strict budget limits — was the unofficial motto of Silicon Valley. Uber reportedly exhausted its annual AI budget in just a few months, multiple firms scaled back Claude licenses, and Meta quietly decommissioned its internal AI usage leaderboard. The mounting costs have forced CFOs to ask a simple question: what did we actually get for our money?
Luck, who focuses on enterprise technology investments at one of the world’s largest venture capital firms, explained that the industry is now in a phase of 'disciplined deployment and measurement.' Many organisations lack the internal frameworks to capture AI’s true impact, because traditional metrics like cost-per-transaction fail to account for changes in workflow, decision-making, and customer experience. Industry reports indicate enterprise AI spending surged over 200% year-over-year, but only a fraction of that has translated into measurable productivity gains or revenue growth.
A new ecosystem of startups is emerging to help companies track AI ROI. According to Luck, forward deployed engineers are acting as 'Trojan horses' for adoption, embedding within client organisations to demonstrate practical value. Rather than committing to a single model provider, enterprises are increasingly mixing and matching models depending on the task. Value, Luck stressed, is being created at every layer of the AI stack — not just at the model layer, but also in infrastructure, deployment tools, and vertical applications.
Looking ahead, Luck expressed particular optimism about consumer-facing personal agents, describing the potential for 'magic moments' where AI assistants seamlessly anticipate user needs. She also offered a measured view on the wave of AI IPOs in 2026, noting that investors are demanding clearer paths to profitability than in previous tech cycles. The days of funding AI companies on promise alone are fading.
For enterprises, the message is clear: innovation without measurement is just expensive experimentation. The tokenmaxxing hangover is forcing a healthier, more sustainable approach to AI adoption, where every dollar must be tied to a demonstrable outcome.
