Job candidates are walking into interviews with a new line item on their compensation wishlist: daily AI token budgets. Not $10 or $20 - we're talking $100 to several hundred dollars per day, just for API access.
Charles Lamanna, Microsoft's Corporate VP for Business and Industry Copilot, shared this observation recently, and it reveals something fundamental about how AI is reshaping not just work, but the entire structure of employment.
Why Token Budgets Matter
Here's what's actually happening. Developers, designers, analysts - anyone building things - now rely on AI tools the way they rely on their laptop. Claude, GPT-4, Gemini, specialist models for code or research. These aren't nice-to-haves. They're core infrastructure.
And unlike a laptop, which you buy once, AI access is a running cost. A heavy user burning through advanced model tokens can rack up serious bills. Companies that restrict access are creating a productivity ceiling. Candidates know this. They're negotiating accordingly.
The interesting bit isn't just that people want AI budgets - it's that they're willing to trade salary for access. That tells you the value they're extracting. If someone will take a slightly lower base to guarantee unrestricted Claude access, they've done the maths. The AI makes them more valuable, faster.
What This Means for Hiring
For employers, this creates an odd dynamic. You can pay someone £80,000 and give them restricted AI access, watching them hit limits mid-sprint. Or you can pay £75,000, add a generous token budget, and get someone who ships faster because they're never throttled.
The smart move seems obvious. But many companies are still treating AI like a perk, not infrastructure. IT departments are setting blanket limits. Finance teams are questioning the cost. Meanwhile, candidates are walking away from offers because the AI policy feels restrictive.
This matters especially for smaller businesses and startups. Big tech companies can absorb AI costs easily. A 10-person startup offering competitive token budgets suddenly looks more attractive than a corporation with strict usage caps. That's a meaningful hiring advantage.
The Bigger Shift
What we're watching is AI access becoming as non-negotiable as a decent salary or health insurance. Five years ago, remote work was a perk. Then it became standard. Now, for many roles, it's a dealbreaker if it's not offered.
AI is following the same path, but faster. The difference is cost structure. Remote work doesn't cost the company more per employee. AI usage does. That creates tension. Companies want to attract talent, but they're also watching cloud bills climb.
The resolution will likely be tiered access. Junior roles get standard budgets. Senior roles, especially in engineering and product, get higher limits or unrestricted access. But any company trying to hire strong technical talent without addressing this will find themselves at a disadvantage.
What Candidates Should Ask
If you're interviewing, here's what to clarify: daily token limits, which models are approved, whether personal tool usage is allowed, and what happens if you hit the cap. Some companies have generous budgets but restrictive approval processes. Others have modest limits but let you use whatever tools you want within that budget.
The worst scenario is finding out after you start that the AI policy doesn't match how you work. Better to negotiate it upfront, alongside salary and benefits. If the company balks at the question, that tells you something about how they think about productivity tools.
AI token budgets are the new desk setup conversation - except the stakes are higher. Get it right, and you're working at full speed. Get it wrong, and you're waiting for approval every time you want to use a better model. For anyone building things in 2025, that's not a small detail. It's the difference between moving fast and moving slow.