Persona AI Assistant Raises $10M Ahead of $179 Wristband Launch

A second act, built faster than the first

Zach Yadegari is 19. In March, MyFitnessPal bought the calorie-tracking app he co-founded, Cal AI. He has now raised $10 million for Persona, a personal AI agent startup. Vine Ventures led the round, and Z Fellows founder Cory Levy and Collective Global also took part.11 Inc. adds that wearable companies WHOOP and Eight Sleep joined the round.12 Persona has not disclosed its valuation or the terms of the deal.414

The product comes in two parts. Today it is a free beta assistant that people use by texting it over iMessage. In December it is supposed to gain a $179 wristband that gives the same assistant a physical, screenless way in.11 Persona says the assistant can order food or rides, take notes, book travel, and answer questions and emails. Yadegari describes these as the standard jobs of a consumer AI assistant.11

The timeline is quick. Yadegari stayed at MyFitnessPal after the acquisition but left in June.11 Inc. reports that Persona launched publicly on X in late September, with the iMessage assistant and wristband preorders going live together. The funding announcement came a few weeks later.12 He is building the company with a co-founder who has not been named yet.27

The traction story, and why the coverage sounds cautious

Early numbers are small, and every figure comes from the company. Yadegari says the iMessage beta has "a few thousand" users and that band preorders have brought in five figures in revenue.11 Several outlets say these figures can't be checked independently. RuntimeWire calls them early indicators rather than proof of lasting use or real hardware demand.4 Value Add VC calls five-figure preorders a thin signal and says the key test is whether the band actually ships in December.5

Inc.'s reporter tried the product, and the result was mixed. The assistant asked for a name, saved itself as a contact, and offered a setup call. Replies took several minutes, though. When asked for three things to do in Dallas, it stalled for about five minutes and eventually suggested one museum.12 That is one test, not a benchmark. Still, it is the only independent hands-on account in the coverage, and it shows how far the beta is from Yadegari's description of an agent that might "start emailing and then hop on a call and then continue on the website."12

Yadegari puts the small user base down to limited promotion, not weak demand. He says the company is starting to buy ads aimed at a mass audience beyond X.12 That is the most important go-to-market detail in this story. Cal AI grew through consumer virality, and Yadegari is due to speak at TechCrunch Disrupt on how to create viral growth and capitalize on it.11 With Persona, he is betting that the same playbook carries over to a much harder product.

The playbook: cut friction first, add hardware second

The coverage generally describes Persona as the Cal AI approach applied more widely. Cal AI took one tedious habit, logging calories, and made it easier. Persona wants to do that for routine tasks across many services.4 Yadegari's own explanation is personal. After moving from his childhood home on Long Island to Miami, he found himself buried in errands such as booking doctor's appointments, keeping up with email, and arranging help to move furniture.12

In this reading, starting on iMessage is a smart way to keep friction low. There is no app to download and no hardware to wait for. Users text a phone number.12 The approach lets Persona test what users actually ask for and how well its agent handles it, while the band and a separate app are still being built.11 It also means Persona can collect usage data before committing to large hardware orders, which looks deliberate.

The hardware is the riskier half. Yadegari's mission statement is to make computing "seamless with human life," and he argues that physical hardware is what makes that possible.12 The obvious counterpoint is the category's track record. Inc. notes that Humane shut down its AI pin business in 2025 after the product flopped. Meta's Ray-Ban glasses have drawn privacy and surveillance concerns, and subway ads for Friend's AI necklace have been defaced in New York.12 Value Add VC adds that AI wearables have a history of missed ship dates.5

Privacy as the main selling point

On hardware, Persona leads with what the band won't do. It is not an ambient device that is always listening and recording. Users activate it with a button or, if they choose, a flick of the wrist.11 TechCrunch contrasts this with Bee, the wearable Amazon bought in 2025.11 Every outlet in the coverage treats deliberate activation as the key difference.168

The rest of the privacy pitch is a set of company claims about the back end. The AI model runs in Persona's cloud rather than on the band. Data is encrypted in transit and at rest, and users can delete it.1 Persona says it won't sell conversations to advertisers or data brokers.7 Inc. reports a further promise that user data won't be used to train its models.12 Value Add VC notes a trade-off: running in the cloud lets Persona improve the AI without new hardware, but it ties response speed to network latency.5

For agents that can act, not just answer, security may matter more than privacy. Persona says it is building defenses against phishing and prompt injection, attacks that labs including OpenAI have warned about.2 The agent can't see users' credit-card details. Purchases need the user's approval and go through outside providers such as Stripe Link, Stripe's wallet for agent-assisted payments.7 Keeping the agent separate from payment credentials is a sensible design. How well it holds up against real attacks is still unknown.6

The business model: free software, ads in shopping

How Persona plans to make money is the most interesting part of the story, and the reporting differs slightly on it. Three pieces appear across the coverage.

The first is the band, a one-time $179 hardware sale that is already bringing in preorder revenue.11

The second is advertising. Yadegari says Persona will show sponsored products when users research purchases. He argues the recommendations stay unbiased because the agent doesn't know which items are sponsored. His example is office chairs: the agent knows the user's preferences and shows matching chairs, some of them paid placements.11 Konsulteer correctly calls this a design principle the company states, not a verified guarantee, and says users will need clear disclosure of sponsored placements.6

The third is what Persona says it won't charge for. Inc. reports that the top priority for the new money is paying ongoing AI computing costs, so the software can stay free and the wristband won't need a subscription. The rest goes to manufacturing and hiring, especially in marketing.12 This is where the coverage differs. Konsulteer lists subscription economics among the factors that could shape adoption6, while Yadegari's comments to Inc. suggest he wants to avoid subscriptions altogether.12

This analysis takes the Inc. account at face value. If it holds, the model is the riskiest part of the plan. A free agent pays inference costs on every request, and if the ads only appear during shopping, those ads have to cover the cost of every other task too. The "agent doesn't know which items are sponsored" design is meant to protect user trust. It also means Persona can't easily steer users toward paid placements, which could limit ad revenue. Both points are inferences from the reporting, not anything Persona has said.

A crowded field, and a big claim

Yadegari doesn't play down the competition. "It is the most competitive industry," he told Inc., adding that Persona is going head to head with the frontier labs.12 TechCrunch names Instinct and Meta's Muse as software rivals and Bee as the closest hardware comparison.11 Value Add VC points to Instinct's valuation of roughly $10 billion to show how small Persona's early numbers are by comparison.5 Inc. also notes that OpenAI is reportedly building its own screenless consumer device.12

His argument for why a small startup can survive is that this is not a winner-takes-all market, and people will pick whichever assistant works best for them.12 His bigger claim is that shopping, Uber, DoorDash, and email are "all collapsing and will collapse into a single super app powered by an AI assistant."11 Later, he wants Persona to be proactive and solve problems before users notice them.11

The verdict

The coverage agrees on the facts and differs mainly in tone. One factual discrepancy stands out. Most outlets cite MyFitnessPal's figure of more than $30 million in annual revenue for Cal AI in under two years10, while Inc. reports $40 million in revenue over a 12-month period.12 Either number explains why investors backed Persona this early.

Our reading is that the $10 million is a bet on the founder more than on the product. Persona's real advantages are a founder who has proven he can win consumer distribution and a privacy-first design that stands apart in a category known for surveillance worries. Its weak points are a slow beta, untested hardware, and an ad-funded free model that has to cover inference costs. The next two months will show which matters more: whether the band ships by Christmas, as Yadegari is aiming for12, and whether paid marketing can grow the iMessage user base from "only in the thousands"12 to something big enough to support it.