A remarkably fast-moving dark horse is emerging in the personal AI agent sector.
On September 28, San Francisco-based startup Instinct announced the completion of a $1 billion Series C funding round, co-led by Sequoia Capital, Benchmark Capital, and Coatue, reaching a valuation of $10 billion. The company, founded only about a year ago with its product still in invite-only beta, already has annual transaction volume exceeding $1 billion on its platform, with daily user growth holding steady at roughly 10%—and zero marketing spend.
Instinct founder and CEO Noah Shinn publicly detailed the company's operational specifics for the first time in an exclusive interview on the podcast "Invest Like The Best" with host Patrick O'Shaughnessy. He said the funding will be used to expand user coverage and continue advancing product development, adding, "We're just getting started." The podcast episode also aired on September 28.
Behind this funding round lies investors' strong bet on the emerging personal AI agent sector. In the interview, Noah Shinn stated bluntly: "This might be the most exciting software race ever, and the end result is a multi-trillion-dollar market."
No App—Just a Phone Number and Email
Instinct's product logic is fundamentally different from most AI tools on the market—it has no standalone app; users interact with it via text message, phone call, or email. Noah Shinn explained: "This is not a new app, not a new tool, but a new experience. It has its own phone and computer. You can text it, you can call it, and it can also proactively call you."
He gave an example of this "proactiveness": "If there's something truly urgent, it will call you and say: 'I don't want to bother you too much, but you need to sign this document before 3 PM. It's 2:55 now. Can you handle it? It's in your inbox, and I can even send another email to push it to the very top.'"
There is a core principle behind this design. Noah Shinn said: "I don't want to confuse a simple interface with limited capability. It has a computer, so it can do almost anything you want to do on the internet."
Annual Transaction Volume Over $1 Billion, Travel Accounts for Half
Despite remaining invite-only, Instinct's business scale is already considerable.
Noah Shinn revealed in the interview: "Our annual transaction volume is currently approaching and exceeding $1 billion, while our user base is still very small. Fifty percent of that comes from travel."
He described a typical scenario: a user simply sends a voice message saying "I need to be in New York tonight," and Instinct automatically identifies the user's current location, books a flight based on their preferred airline, seat type, and credit card, then connects to a hotel, an Uber for airport transfer, and syncs everything to the calendar—"what the user actually does is just record a voice message. Everything else is handled."
He believes this experience will fundamentally change the travel booking industry landscape: "The value travel agents provide lies in integration and presentation, but now there's a better experience."
Growth: Zero Marketing Spend, 10% Daily Growth
Instinct's growth trajectory is equally striking.
Noah Shinn described the starting point of growth: "We initially gave the product to about 200 people, all close friends and family. The next day about 5 people came, then 210... then growth began to accelerate, from 1%, 2%, to 3%, 4%, then 6%, 7%, 8%, 9%, and now we're at roughly 10% to 11% daily compounded growth."
Each user only has 5 invitation slots. Noah Shinn said: "About 10% of users each day decide to give one of their precious invitations to someone else, and this happens every day."
This sense of scarcity has spawned a secondary market. "I saw someone selling invitation codes on eBay a few days ago, about $300 each," he said.
Throughout the entire process, Instinct's marketing spend has been zero.
Trust Is the Core Variable: 40% of Users Share Credit Cards After 3 Weeks
The core challenge facing personal AI agents is trust. How much sensitive information users are willing to entrust to it directly determines how much it can do.
Noah Shinn shared a set of data: "We found that building trust takes about a few weeks. Three weeks after joining, 40% of users have already shared their personal credit cards with Instinct."
He views "time to first credit card share" as a proxy metric for trust, and emphasized the significance of this figure: "That 40% also includes the proportion of users who had already churned before that point."
More critically, the retention data: "We found that once users share at least one piece of sensitive information with Instinct and truly trust it, retention rates reach as high as 80%. For a consumer product, that number is quite remarkable."
Business Model: Benchmarking Apple Pay, Rejecting Advertising
Noah Shinn explicitly rejected the advertising monetization path.
"I don't want to build that kind of reality," he said. "Imagine a world where Instinct is smarter than the user—if it uses its intelligence to influence users to buy things they don't want to buy, that would be very dangerous."
He characterized the advertising models of platforms like Google, TikTok, and Instagram as "users don't pay, users are the product," and stated that Instinct's design principle is to "act entirely on the user's side."
On the business model front, his approach is transaction commission. "What I see is a transaction commission model covering the entire platform, just like Apple Pay—users use it for free, and merchants pay to gain distribution channels." Using the travel industry as an example, he pointed out that some boutique hotels are willing to pay commissions as high as 30% per transaction, while Shopify's commission rate is about 2% to 3%, Amazon's is about 10%, and Apple's in-app purchases are 30%. "We don't yet know where we'll land on that curve, but our goal is to be at the high end."
Compute Power Is the Biggest Anxiety
Noah Shinn admitted that he spends about 40% of his time thinking about compute power.
"This isn't like a traditional consumer product where you double users and buying 2x resources is enough," he said. "The problem is that compute procurement has a delivery cycle of several months. If you're running at 10% daily growth now, in 3 to 4 months that's 100 million users. Are you going to buy compute for 100 million users now? If you buy wrong, the cost is 3x to 4x."
He also pointed out that Instinct's compute needs are fundamentally different from code generation products: "Instinct can 'wake up' and 'go to sleep' at any time of day. It might wake up at 6 AM because it knows you get up at 7, and it will scan everything to make sure everything is ready for you. This kind of background workload is something code generation products simply don't have. The compute required will be orders of magnitude more than we anticipated."
He also revealed that Instinct has found a cost optimization solution: through customized inference deployment, separating background tasks that don't need real-time responses from immediate interaction tasks, achieving 3x to 8x efficiency improvement with the same compute power, thereby delivering services at extremely low cost with performance levels approaching frontier models like Opus 5.
"Instinct Interconnect" Network and Industry Restructuring
Instinct recently launched a "Trusted Person Network" feature, allowing two users' Instinct agents to communicate directly, automatically coordinating schedules and sharing files.
Noah Shinn described a typical scenario: "You just state your intent—I want to meet this person before the weekend. It will coordinate with the other person's Instinct, find a time when both are available, and then put the meeting on the calendar."
He believes this feature will have a profound impact on existing digital services, but his attitude is cooperation rather than confrontation. "I think most digital services won't be disrupted, but will be changed and transformed," he said. "Our approach is to find these businesses and tell them: this is what we think your business looks like now, this is how Instinct users currently interact with your business, and what can we do to make both sides better off?"
His core judgment is: businesses whose revenue depends on users spending time on their apps (rather than on actual transaction volume) face the greatest risk; those that can benefit from higher transaction volumes will gain from it.
Comments