Instinct Went From a $2.5 Billion Valuation to $10 Billion in About Five Weeks. It Has Published No User Numbers.

On 28 September 2026 Instinct announced a $1 billion Series C at a $10 billion valuation, led by Sequoia Capital, Benchmark and Coatue. About a month earlier, founder Noah Shinn told the Wall Street Journal the company had raised $250 million at roughly $2.5 billion, as relayed by two outlets we reviewed. That is a fourfold step in about five weeks, for an invite-only assistant that works over text and phone calls, has no mobile app yet, and has not disclosed users, revenue or growth. A valuation is a price, not evidence, so the useful question is what that price is paying for.

What was disclosed

The round was first reported by The Information, and Instinct confirmed it as a Series C in a press release, per TechCrunch. The Business Wire release names the three investors and says the $1 billion is additional to the company's latest round, giving no earlier amounts. TechCrunch and the release both put the round at $1 billion; neither mentions a larger figure. The prior round was reported by Let's Data Science, citing Reuters, as $250 million at about $2.5 billion. Another outlet, Ecosistema Startup, reports the same $250 million at $2.5 billion and says total funding to that point was $350 million. The $350 million therefore appears to be cumulative funding, not the size of the earlier round.

The press release, and TechCrunch, say the product is still in early access. Instinct launched invite-only in August 2026. TechCrunch says the company has shared no user numbers, revenue or growth metrics. Shinn is quoted saying the funding will help bring Instinct to more people and continue building the future of personal AI.

What is being sold

Users text or call the assistant, which then completes tasks using its own phone number and computer, per the release. TechCrunch lists booking travel and reservations, making purchases, paying bills, cancelling subscriptions and ordering groceries. The release highlights three recent features: Concierge, a premium service for phone calls, specialised bookings and billing disputes; a Trusted Person Network that lets a user's Instinct coordinate plans with other people's Instincts, including sharing files; and location sharing through iMessage. All of these descriptions come from the company.

What a $10 billion price implies

Without usage data the price can only be read through arithmetic and comparison. If the $10 billion figure is post-money, which the release does not state, the $1 billion round represents about 10% of the company. At a 20-times revenue multiple, an assumption chosen only for illustration, $10 billion would correspond to $500 million of annual revenue. Instinct has said nothing about its revenue, so the point is the scale of what the price assumes, not a claim about the company. The earlier step, from $2.5 billion to $10 billion, was priced by the same market in roughly five weeks, which says that whatever investors learned in that window was large. It does not say what they learned.

Investors in a private round often see metrics that are never made public, and the absence of published numbers is not proof the numbers are weak. The reverse also holds: a high price from well-regarded investors is not proof the numbers are strong. Anyone outside the cap table is judging a story, not a dataset.

The security list is a liability answer

An agent that has its own phone and computer inherits the consequences of the tasks it performs: payments, cancellations, addresses, and calls placed in a user's name. The release's security section reads like an answer to that exposure. It lists isolated sandboxes, short-lived local credentials and identity-signed tool execution, which the company says were built in from the start, and an active detective system that it says catches subtle hallucinations while the agent forms its responses. These are Instinct's own claims and the release offers no third-party audit or incident data. They are also design promises, not evidence of an outcome.

The need is not hypothetical across the category. Meta's Muse, a rival agent, was the subject of a Marketplace dispute that we covered separately, in which a user says the agent gave a buyer his address; Meta says it followed instructions and neither account is verified. The point here is narrower: in this category, where an agent acts in a person's name, a company's valuation is also a bet on how it handles the failures that will occur.

The rival case: distribution against trust

TechCrunch frames Meta's Muse as the obvious comparison. Muse offers many similar features and is integrated into Meta's social products; it can monitor and summarise Instagram messages and Facebook Groups and track Marketplace listings. TechCrunch reports that Muse has reached the top of US app stores and been downloaded millions of times. Instinct has no app and no disclosed users.

That sets up the argument behind the price. If the agent market is decided by distribution, Meta starts with billions of accounts and Instinct with an invite list. If it is decided by trust, a company whose entire product is the agent, and which has built its security story around that, might be preferred over a platform whose agent is one feature among many. Which of these wins is unknown, and the $10 billion price is a bet on the second. The first remains the more conventional view.

What the evidence does not establish

• How many people use Instinct, how often, or what they pay. None of this has been disclosed.
• Whether the security measures work as described. The claims are the company's own.
• Whether the $10 billion figure is pre- or post-money; the release does not say.
• The terms of the round, including any structure that would make the headline valuation differ from the price paid.