The most important AI story today is not another benchmark win. It is that buyers may finally be behaving like AI is part of the operating budget.
Business Insider reports that a new RBC Capital Markets survey of more than 100 chief information officers and technology leaders found broad momentum for enterprise AI spending in the second half of 2026. That matters because these are not casual users testing a chatbot over lunch. CIOs control the budgets that decide whether generative AI remains a fascinating demo or becomes a line item that survives procurement, security review, finance scrutiny, and renewal season.
The headline finding is blunt: every respondent in the survey is allocating budget to AI or large language model projects, and 91% are creating new AI budgets rather than simply raiding existing software spend. That is the number to watch. If AI dollars are additive, vendors get more room to grow and customers get more room to experiment without canceling half the stack to pay for tokens.
The survey also cuts against one of the loudest worries in the market: that usage bills would scare enterprises away. Nearly nine in ten respondents said token budgets are manageable, even though almost half have already exceeded original spending plans. That does not mean AI is cheap. It means the perceived value is high enough that the bill has not yet become a deal-breaker.
OpenAI appears to be the immediate beneficiary. RBC's survey found that 57% of respondents use ChatGPT most often among model-based AI services, compared with 12% for Anthropic's Claude. OpenAI also led on perceived performance, with 44% naming it the highest-performing provider versus 24% for Anthropic. For all the debate about frontier model rankings, enterprise distribution may be turning into a separate and equally powerful moat.
The other useful signal is where deployments stand. More than half of respondents said AI is already in production, while another 35% expect to get there within six months. That is the shift Daily AI Paper readers should care about: AI adoption is becoming less about who has the most impressive lab demo and more about who can survive messy production workflows, usage-based pricing, compliance demands, and department-level ROI.
There is still plenty of caution baked in. This is one survey, not the whole market. CIOs may be more enthusiastic in a bull cycle than they are when budgets tighten. And production usage can range from narrow internal copilots to mission-critical automation. But the direction is hard to ignore. If token costs stay manageable and hybrid seat-plus-usage pricing becomes normal, enterprise AI may expand the software budget instead of merely rearranging it.
That is why this story is worth publishing: it suggests the next phase of the AI race may be decided less by launch theatrics and more by budget gravity. The companies that win will not just have strong models. They will make AI easy enough to buy, govern, meter, and justify every quarter.