Instinct personal agent hits $1B in annualized transaction volume, growing 10% daily with $0 marketing spend

Instinct hits $1B in annualized transaction volume, growing 10% daily with zero marketing spend

Instinct, a personal agent startup founded by 23-year-old Noah Shin, has reached $1 billion in annualized transaction volume while still operating invite-only and spending nothing on marketing. The company is growing 10% per day.

The scale is striking relative to comparable launches. Stripe took 18 months to hit $1 billion in annualized transaction volume and spent two years in private beta before that. Instinct reached the milestone in roughly six months, though it's worth noting the transactions flow through third-party payment cards — Instinct is not yet extracting fees from the volume itself.

The monetization path

The clearest near-term revenue opportunity is issuing its own card. Even at fractional basis points — say 0.1% of transaction volume — a billion dollars in annual throughput produces tens of millions in revenue for a 14-person team. That math likely explains why Benchmark and Sequoia backed the company at a $10 billion valuation and why Code2 also participated in the round.

Longer-term, Instinct can follow other fintech patterns. Travel accounts for 50% of transaction volume, suggesting an Expedia-style take-rate play on bookings. Restaurant reservations and other commerce could absorb take-rates comparable to Shopify (2.5–3%) down to Apple (30%), though Noah Shin has stated his personal goal is to keep Instinct free for life.

Retention and user behavior

The core product has already generated strong behavioral lock-in. Forty percent of users shared a personal credit card with Instinct within three weeks. Users who connect at least one piece of sensitive information retain at roughly 80%. Some users send more than 90% of their messages to Instinct by voice.

The flywheel runs on social proof. Instinct launched to 200 friends and family, and invites have resold on eBay for around $300. Screenshots of agent wins — and agent failures — drive organic interest. The company spent nothing on paid marketing because users post about the product constantly.

Compute as the constraint

Compute demand is doubling roughly every week. Noah Shin spends 40% of his time buying compute, a signal that provisioning has become the binding constraint on growth. Buying compute at the last minute costs 3–4x more, which Shin flags as a specific efficiency problem. The company runs background work on setups that are 3–5–8x more efficient on the same underlying capacity than default configurations.

One detail from Shin: Instinct's custom model matches Anthropic's Opus 5 performance at a fraction of the cost, though the specifics are not disclosed in the segment.

The restaurant reservation problem and agent-to-agent negotiation

Shin raised a conceptual point about how agents will reshape booking systems. Restaurant reservations have historically operated on first-come-first-served logic — whoever books first gets the best slot, regardless of the occasion. His argument: restaurants should optimize for high-intent diners, such as someone celebrating an anniversary or birthday, rather than whoever happened to book first.

This opens a deeper structural question: what happens when every diner has an agent making their case? One host suggested that restaurant and reservation systems will develop their own agent infrastructure to handle the inbound, creating agent-to-agent negotiation at scale. In this model, your agent calls the restaurant, the restaurant's agent answers, and they negotiate availability based on your preferences, special dates, and constraints — all without human intermediaries.

The other host was more skeptical, arguing that premium restaurants intentionally reward long-term planners and regulars with preferential access, and that agent-driven optimization could erode the friction that makes high-touch hospitality valuable.

Both positions hold. Early dynamics will likely surface agent spam (restaurants flooded with bot calls from services checking availability), leading to filtering infrastructure and eventual standardization — similar to how spam filtering became table stakes for phone systems.

Risks: The coordination problem

One warning flagged in the segment: mass adoption of personal financial agents could trigger systemic instability. The Apollo chief economist and former SEC chair Gary Gensler both raised the same concern — agents optimizing simultaneously, without coordination between them, could cause flash crashes in markets or rapid shifts in capital allocation that destabilize financial infrastructure.

This is not a far-fetched scenario. If millions of users delegate investment decisions to agents, and those agents make similar choices at scale, the outcome resembles the 2010 flash crash or the liquidity crises that follow unanimous positioning. The conversation acknowledged this happens slowly: Instinct's user base is still in single-digit millions, trust-building takes years, and agents still ask permission for major moves. But the trajectory is clear and worth monitoring.

Key metrics

• $1 billion in annualized transaction volume
• Growing 10% daily
• $0 marketing spend
• 14-person team
• 40% of users connected a credit card within 3 weeks
• 80% retention for users with at least one sensitive data connection
• 50% of transaction volume is travel
• Raised $1 billion at $10 billion valuation (from Benchmark, Sequoia, Code2)
• Noah Shin is 23 years old, attended Northeastern