Instinct, the startup behind a viral personal AI agent, just raised $1 billion at a $10 billion valuation. Sequoia Capital, Benchmark Capital, and Coatue led the round.
Here is the part that should make you sit up. Last month, founder Noah Shinn told the Wall Street Journal that Instinct had raised $250 million at a $2.5 billion valuation. In four weeks, the valuation quadrupled.
That tells you how fast the money is repricing this category. The company had been explicitly shopping a $2.5 billion target. Investors saw the product's traction and simply paid $10 billion instead.
So what does a $10 billion personal agent actually do? Trip planning. Reservations. Grocery orders. Subscription cancellations. Instinct is building an AI that quietly handles life's admin.
The mechanics are wilder than the pitch. Instinct can operate its own phone and computer to carry out tasks, including calling businesses to get things done. It has a white-glove Instinct Concierge for hard-to-book reservations, and a Trusted Person Network that lets multiple agents talk to each other, coordinate plans, and pass files around on your behalf.
If the architecture sounds familiar, it should. Shinn is the researcher behind Reflexion, one of the seminal papers on LLM agents that learn by reflecting on their own mistakes. The man who defined agent self-improvement is now shipping the consumer version.
The competition is already forming. Meta's Muse agent connects to email, calendar, and messaging apps to shoulder routine chores, and every hyperscaler has a personal assistant somewhere in its roadmap.
But there is a structural difference: Meta wants the assistant inside its products. Instinct wants to be the independent layer that acts across every product you already use, with a hallucination-detection system baked in to catch its own errors before executing.
Zoom out and the bet is obvious. Language models commoditized information. Agents commoditize action. The company that owns the execution layer owns the relationship with the user, and the relationship is where consumer spend flows.
Note the market logic: the AI wave so far monetized through enterprise seats and token usage. Personal agents move spending to consumers, and if booking trips and restocking the pantry without lifting a finger becomes a habit, that is durable revenue.
Instinct is still in early access, and $1 billion buys a long runway to convert novelty into habit. The past two years of AI have been about models that think. The next decade is increasingly about agents that do.