The labs are writing the rules now

Anthropic tripled its revenue to $30 billion in four months. On the same day, OpenAI published a 13-page economic blueprint proposing robot taxes, public wealth funds, and a four-day workweek. Everyone covered them as separate stories. They are the same story.

Bloomberg reported that Anthropic's run rate surged from $9 billion at the end of 2025 to over $30 billion, with more than 1,000 enterprise customers each spending over $1 million annually, a number that doubled in under two months. The company also locked in 3.5 gigawatts of next-generation Google TPU capacity through Broadcom, on top of the 1 gigawatt already being delivered. That is not the compute profile of a company building a product. It is the compute profile of a company building infrastructure for an economy.

Then there's the policy paper. TechCrunch reported that OpenAI released "Industrial Policy for the Intelligence Age," calling for a public wealth fund holding equity stakes in AI companies, a shift in taxation from labour to capital, and employer trials of a 32-hour workweek with no pay cut. A private company valued north of $300 billion is telling the government how to restructure the tax code and the working week to accommodate the disruption that company is causing.

The pattern should be obvious but somehow isn't. When a company reaches the scale where its growth reshapes labour markets and fiscal policy, it doesn't wait for regulators to catch up. It writes the framework itself. OpenAI isn't lobbying. It's drafting.

And this isn't civic-mindedness. It's rational positioning. If someone is going to design the rules for how AI's economic gains get distributed, OpenAI would rather it be OpenAI. A public wealth fund holding equity in AI companies sounds progressive until you realise the companies proposing it would set the terms. A robot tax sounds like accountability until you recognise it would primarily burden competitors without the scale to absorb it.

The preview is already running

Japan shows what happens when the demographic pressure arrives before the policy debate resolves. TechCrunch reported that after 14 consecutive years of population decline, with the working-age population set to shrink by 15 million over the next two decades, Japan has pushed physical AI from pilot programmes into deployment across factories, warehouses, and services. The government committed $6.3 billion under PM Takaichi, aiming to capture 30% of the global physical AI market by 2040. Japan already holds 70% of global industrial robotics.

Japan didn't wait for a white paper. It didn't debate robot taxes. It deployed robots because there weren't enough humans to do the work. The economic restructuring that OpenAI proposes as future policy is already present reality in countries where the demographic maths doesn't leave room for philosophical debates about the workweek.

I think the question builders should sit with is this: when the companies generating the disruption are also designing the response to it, who is the policy actually for? Anthropic tripling to $30 billion tells you the velocity. OpenAI publishing an economic blueprint tells you these companies know the velocity, and they want to shape what comes next. The labs aren't waiting for government. They're offering to do the job themselves. Whether that's a feature or a bug depends on whether you believe the authors of the disruption should also be writing the rules for living with it.

---