25,000 jobs eliminated per month. OpenAI's $85B burn rate. The "GPU debt treadmill." And the company building the automation now telling governments how to handle the fallout. The AI economy is not coming — it's already here.
This week, OpenAI — an $852 billion company — released a policy proposal suggesting governments should consider taxing AI and robots to fund public wealth funds, expand social safety nets, and subsidize a four-day workweek with no loss in pay.
OpenAI's proposal blends traditionally left-leaning mechanisms — public wealth funds, portable benefits, expanded healthcare — with a fundamentally market-driven framework. Framed as distributing AI prosperity more broadly, it's also the most expensive reputational insurance policy a $852 billion company has ever purchased. When you're building the technology eliminating jobs at scale, proposing the policy response is a power move, not philanthropy.
The jobs being eliminated aren't on the factory floor. Microsoft Research published new findings this week identifying the roles most exposed to the next wave of generative AI and autonomous agents: finance, legal services, and software engineering. The professions that require degrees, charge high hourly rates, and sit behind large desks in expensive cities.
The Goldman Sachs number — 25,000 jobs per month — is not a projection. It's a current measurement. Companies across five sectors that Morgan Stanley tracks have already reported a 4% net reduction in headcount, with early-career roles hit hardest. The entry-level analyst, the junior associate, the new hire learning the ropes — those positions are the first to disappear, and they're not coming back.
The uncomfortable finance reality: this is being rewarded. Companies that announce AI-driven headcount reductions are seeing stock price increases. The market is pricing human capital as a cost to be minimised, and AI infrastructure as an asset to be built. That incentive structure doesn't reverse.
Meanwhile 66% of banks surveyed by American Banker say AI is now a high strategic priority, with most increasing AI spending by at least 10% over the last 12 months. The institutions managing your money are automating the people managing your money.
Model Context Protocol just crossed 97 million installs — making it the default way AI agents connect to your tools, databases, and APIs. If you're a finance professional building AI workflows, this is the standard that will define your stack. Connect Claude to your CRM, your financial data, your email — and let it act on your behalf across all of them simultaneously. This is what "agentic AI" actually looks like in practice.
Learn MCP →OpenAI proposing robot taxes while simultaneously eliminating jobs at scale is the defining paradox of the AI economy. It's not cynical — I actually think some of those proposals are worth taking seriously. But the power dynamic is worth naming: when the company building the disruption gets to define the policy response, the policy will protect the company's ability to keep building.
The Goldman number — 25,000 jobs a month — is the one that should focus minds. Not as a reason to panic, but as a forcing function. The finance professionals who will be fine are the ones who understand these tools well enough to direct them, not just use them. There is a real and growing difference between those two things.
The AI economy is not a future event. It's the current one. This briefing is the evidence.
Join professionals across North America who get The Ledger Wire every Wednesday — AI and finance intelligence in 5 minutes flat.
🎉 You're in! Briefing #5 lands next Wednesday. See you then.
🔒 Free forever. No spam. Unsubscribe anytime.