🕶️ Nvidia bet $1.1B on a 2-month-old

Good morning. It's Wednesday, which means you're exactly halfway through caring about this week's goals.

Two things happened that shouldn't be possible yet. A two-month-old company raised eleven figures from people who usually ask for traction.

And Mercury just issued a credit card to something that files zero expense reports. If you're wondering who's actually in charge of the budget now, you're asking the right question. Somewhere between ownership and automation, the rules quietly rewrote themselves.

burnd is built to work individually or as a complete system.


Choose the product your skin needs - or build the full routine. Right now, Buy 3 Get 1 FREE.

In case you missed it under the pile of enterprise LLM press releases, River AI just did the unthinkable: it made your AI yours. This week, the company closed a $1.1 billion round led by General Catalyst and Amppublic, with strategic checks from chip royalty Nvidia and AMD, plus Y Combinator and Temasek riding along.

The pitch is a middle finger to corporate chatbots: a personal AI stack where you own the hardware, the data, and the intelligence. No walled garden. No subscription-gated brain rental. CEO Igor Babuschkin told the New York Times the point is letting people retrain and modify AI without Big Tech holding the leash.

Here's the part that stings. Their training API is already live, which means this isn't a slide-deck fantasy. Two hardware giants just bet real money that intelligence belongs at the edge, not locked inside hyperscaler clouds.

And the ripple hits closer than you think. Every AI startup whose entire business is reselling access to someone else's model? That moat just sprung a leak. When users can own and retrain their own stack, the middleman tax evaporates.

Bake user-ownership into the product: data export, local deployment, retrainable instances. Trade your data-moat obsession for interoperability. The incumbents are reinforcing their gardens. Days from now, watch River's first developer adoption numbers. That's your signal.

🔋 Battery advice is a mess.

Never charge past 80%. Always keep it above 20%. Never leave it plugged in overnight. Fast charging destroys your battery. Rechargeable AAs explode.

So... which of these are actually true?

Mercury just turned corporate cards into API endpoints. Instead of plastic for employees, its new “AI agent cards” mean startups are issuing cards to agents that sit inside ops, finance, or infra workflows and spend autonomously.

This shifts spend management from “who swiped?” to “what process executed?”, treating purchases as outputs of code, not one-off human decisions. Policy becomes prompts, limits, and routing logic wired into agents; finance teams move from approving transactions to designing systems.

If this model holds, the real competition won’t be card perks but who enables the safest, most programmable autonomous spend workflows. That raises new questions: how audits trace agent behavior, how risk teams interrogate models, and how much financial authority founders are willing to hand to software that never sleeps, and never files expense reports.

Scale your growth: Join names like HubSpot, Ahrefs, and Deel. Put your product in front of 85,000+ startup founders. Let's talk.