Instinct AI Assistant Valuation Hits $10B After $1B Series C

A four-month climb from obscurity to $10 billion

Instinct is an invite-only AI assistant that people use by sending text messages, and Silicon Valley insiders started passing it around this summer. It has since raised money faster than almost any consumer AI company before it. In late August, The Wall Street Journal reported exclusively that the startup had raised a round valuing it at $2.5 billion and that its total funding had reached $350 million. Early users had already started raising privacy concerns.14 About a month later, Instinct confirmed a $1 billion Series C at a $10 billion valuation. Sequoia Capital, Benchmark and Coatue invested in the round.6

The speed is the story. OfficeChai traced the company's financing back to an early round at roughly $50 million, backed by Conviction and Greenoaks. Kleiner Perkins then led a $75 million Series A at about $500 million in early August. Index Ventures and Benchmark co-led a $250 million Series B at $2.5 billion on August 26. The new Series C brings total funding to more than $1.3 billion.14 On those numbers, the valuation grew about 200-fold in roughly five months. The company has 14 employees, which works out to more than $700 million of valuation per person.14

TechCrunch pointed out that Instinct only opened its invite-only service in August 2026. That makes it unusual for the company to be raising a Series C at all.6 Instinct has not published user counts or growth metrics.12 OfficeChai reports that it has passed 100,000 users and that compute limits have sometimes forced it to warn people about slower responses.14

What investors think they are buying

Instinct works differently from a chatbot. Users text it a request and the agent carries the task out. Reported examples include booking a cabin, making a dinner reservation, faxing a bank to get a late fee refunded, checking whether a dentist is in-network, arguing with an insurance representative, and canceling forgotten subscriptions.9 To do this work, Instinct uses its own phone number and its own computer. Recent additions include a "concierge" feature that places phone calls for appointments at businesses with no online booking. A "trusted person network" lets one user's agent coordinate plans with the agents of that person's friends.6

Coverage of the round agrees on the investment thesis. TechCrunch described the funding as a sign of excitement about a new class of consumer agents that complete tasks for people instead of just talking with them.6 OfficeChai described the bet more broadly: investors think the consumer agent could become the next computing layer, and that the first company people trust with it will be very hard to replace.14 Carbon Track argued that the names behind the round show institutions treating Instinct as a long-term bet, not an early-stage experiment.9

That reading fits a wider surge in AI deals. Around the time of the Series B, EliseAI was reportedly in talks at a $3.7 billion valuation. DeepSeek was said to be seeking a valuation of about $74 billion.5 ElevenLabs, a more mature comparison, is valued at $22 billion after a September secondary sale. That figure is roughly 36.7 times its reported $600 million in revenue.15

The contrast matters. ElevenLabs has a revenue line that investors can apply a multiple to. Instinct does not. The service is free, and its founder has said he does not want to charge users. Advertising has been mentioned as a possible business model, which leaves open how the company will make money.14 Sequoia, Benchmark and Coatue are therefore paying for distribution and position, not for earnings. This is an analytical point, but an important one: the $10 billion figure reflects how much investors fear missing the category, not anything Instinct has proven about its economics.

The founder and the competitive streak

The company is Spear Street Technology, Inc., a San Francisco business registered in California in April. Reports name Noah Shinn, 23, as founder and CEO, and no co-founders have been publicly identified.14 Shinn left Northeastern University in 2023. As an undergraduate he wrote Reflexion, an influential paper on how AI agents can learn from their mistakes. He then worked briefly at the AI startup Sierra.119

Business Insider's profile credits much of the early buzz to one decision: opening the product to venture capitalists first. VCs then posted on social media about an agent that did their work for them.7 The same profile describes a combative side to Shinn. Upstarts Media reported that two of his major backers, Index Ventures and Conviction, were also funding Town, a rival agent startup backed by Andreessen Horowitz. After Shinn complained, both firms abruptly pulled their money from Town.7 An Instinct spokesperson would not engage with outside characterizations of the episode.7 Instinct also declined to make Shinn available for interviews about the Series C and issued a written statement from him instead.6

The Town episode suggests something about the market beyond Shinn's personality. When one hot startup can push established firms to drop a competing bet, investors are being forced to choose sides in the agent category early.

Meta looms over the valuation

The largest risk to Instinct is Meta, not other startups. TechCrunch reports that Meta's assistant, Muse, offers many of the same features. Muse also plugs deeply into Meta's social apps, so it can summarize Instagram DMs, follow Facebook Groups and watch Marketplace listings. It has reached the top of U.S. app store charts with millions of downloads.6 Instinct still has no mobile app and runs entirely through SMS.12

Dealroom reported that Meta's AI chief has publicly mocked how slow Instinct is.11 Business Insider noted the asymmetry behind that remark: Meta has huge amounts of computing power and cash, while Instinct is free, sometimes lags under heavy demand, and relies on investor money to keep running.7 The Series C works as a war chest for a compute-heavy fight against a company that can absorb losses for a long time.

The reporting is inconsistent on Meta's product. Carbon Track's August-era account said Meta was internally testing a similar assistant called Hatch. The same article then confusingly referred to the rival's scale as "Sierra's."9 Later coverage consistently calls Meta's shipped product Muse.67 The most plausible reading is that the internal test grew into a public launch within weeks. That pace would itself show how narrow Instinct's head start may be.

Trust is the unpriced liability

All the major accounts treat privacy as the main weakness. The original Journal report noted concerns from early users about giving an assistant access to email.1 To work at full capability, Instinct asks for access to a user's email, calendar and other connected accounts. Its terms allow the company to train models on user data.9 TechCrunch described the first version of Instinct's privacy policy as especially worrying because of how much it claimed. The policy has since been revised.6

The problems go beyond policy wording. Business Insider reported that Shinn acknowledged the agent can hallucinate output that looks like another person's data.13 OfficeChai collected several incidents. Investor Katie Jacobs Stanton reportedly had the agent send an email on her behalf. A user said Instinct described a financial document that was not theirs. TechCrunch had already flagged security and privacy concerns in August.14

Whether these are hallucinations or actual cross-user leaks matters a great deal. Both are serious for a product built to handle bank letters, insurance disputes and medical scheduling. Carbon Track noted that an agent acting for a user, instead of advising them, changes both who is liable and how much data access it needs.9

Where the reporting disagrees

The broad story is consistent across outlets, but some details differ. Some summaries said Instinct raised $350 million in its "newest round."5 More careful accounts separate the $250 million Series B from about $100 million raised earlier, which together make the $350 million total.79 Timelines also conflict. Dealroom says Instinct launched in summer 2025.11 TechCrunch puts the invite-only launch in August 2026, and OfficeChai says the operating company was registered in April.614 The 2026 timeline fits better with the reported spring valuation of $50 million.

The verdict

Instinct is the clearest example so far of venture capital's appetite for consumer agents. A 14-person company with no revenue model has reached a $10 billion valuation in less than half a year. The capital is real, and so are the viral adoption and the founder's research background. The risks are just as concrete. Meta has a stronger distribution channel, compute costs grow with every new user, and data incidents keep piling up for a product whose value depends on being trusted with users' most sensitive accounts. Whether this Series C turns out to be prescient or a peak will depend on Instinct's trust record and economics, not on more viral demand.