Revenium’s take on yesterday's New York Times feature, and what we're seeing behind the headlines.
The New York Times ran a piece yesterday on the rise of "tokenomics," the new field trying to answer what companies are actually getting for all that AI spend.
The story features Revenium’s co-founder, Jason Cumberland, who shares insights into the spending explosions customers keep running into.
We're glad the conversation is finally moving here. But if the last 18 months taught us anything, it's that measuring AI spend is only step one. The harder problem is control.
What we see across customers
Every week we talk to teams living through the exact whiplash The New York Times describes. A few patterns show up almost every time:
- Spend outruns understanding. Finance sees the invoice before engineering sees the impact. By the time anyone reconciles the two, another month has been billed.
- Guardrails get bolted on late. Most orgs discover they need model-level limits, per-team budgets, and per-agent policies only after a runaway workflow burns through six figures in a weekend.
- Outcome data is missing. Token counts are easy. Tying those tokens to features shipped, tickets resolved, or revenue influenced is where almost everyone gets stuck.
What actually works
The teams pulling ahead are treating AI spend the way finance treats every other volatile input. They set budgets by team, model, and use case. They enforce those budgets automatically, not through slide decks. They map token consumption to a specific business outcome, so every dollar has a job.
That is the operating model Revenium was built around. Live visibility into every call, guardrails that hold the line when a workflow tries to run away, and a straight line from AI cost to the outcomes that actually matter to the business.
Where this goes next
Tokenomics as a field will keep maturing. The Tokenomics Foundation is a good sign. Shared definitions and disclosure standards will help. But standards alone won’t stop spend from getting out of hand. Enterprises need a system of record for AI economics, the same way they have accounting systems for financials, CRM for customer info, and observability tools for what their software is doing.
That is the category being built right now, and it is the one we are betting on.
The takeaway
Being in The New York Times is a nice moment for our team as it validates what we’re building. The more important moment is the one every CFO and CIO is walking into next quarter, when the AI bill lands and someone has to explain it. The companies that get in front of that conversation are the ones building the muscle for AI economic control now, and we’re excited to help them get there.
If that sounds like a problem you are chasing, we should talk.
Read the NYT feature → nytimes.com/2026/08/03/business/economy/ai-spending-tokenomics.html
Learn how Revenium helps → revenium.ai
