On a single Palo Alto campus that already houses Brett Adcock's other ventures, a new sign went up this year. Hark, an AI lab Adcock founded in 2025, now employs 45 engineers and designers, including former Meta AI researchers and designers from Apple and Tesla, all working alongside his other companies [TechCrunch, March 2026]. The pitch is ambitious and unhedged: build, in Hark's own words, "the most advanced personal intelligence in the world" [hark.com, retrieved 2026], wrap it in a family of devices [Bloomberg, March 2026], and ship the first model this summer [scommerce, 2026].
The wager is funded entirely by Adcock himself. Hark launched with $100 million in seed capital, self-funded by its founder [The Information, 2026]. No outside investors are listed on the cap table. That is unusual at this scale, and it is the cleanest signal of how Adcock wants to run the company: fast, private, and on his own clock.
The bet
Hark is selling a thesis before it is selling a product. The company describes itself as an AI lab building futuristic interfaces, models, and next-generation hardware devices [BusinessWire, March 2026]. The interface piece is what makes the bet specific. Hark is working on AI models with plans to release its first one this summer [scommerce, 2026], and pairing them with a hardware roadmap that Bloomberg has described as a family of AI devices [Bloomberg, March 2026]. The wedge, in other words, is vertical. Own the model, own the industrial design, own the surface the user touches.
That is a hard route. It is also the route a small set of companies, including named competitors Humane and Rabbit, have already attempted with mixed commercial results. Hark's answer appears to be talent density and patient capital rather than a faster go-to-market.
Why it could be big
The consumer AI hardware category is unsettled. The phone is a 18-year-old form factor, voice assistants have stalled, and the largest model labs have not yet defined what an ambient AI device should feel like in a pocket or on a wrist. Whoever defines the interface stands to define the category. Adcock is betting that the team capable of doing that work is assembled at a single campus, not distributed across a Slack workspace.
The early hiring supports the thesis. Abidur Chowdhury, who previously designed the iPhone Air, phones, and laptops at Apple, joined Hark as Head of Design [The Information, 2026]; [eWeek, 2026]. Chowdhury joined 30 engineers from companies like Google, Meta, and Amazon who have moved to Hark [scommerce, 2026]. The Information separately reported that Hark has hired leadership from both Apple and Meta [The Information, 2026]. For a company less than a year old with no shipped product, that is a credible bench.
| Metric | Value |
|---|---|
| Seed funding ($M) | 100 $M |
| Engineers and designers on staff | 45 people |
| Engineers reported moved to Hark | 30 people |
The team and traction
Adcock is the founder and the sole external face of the company. His prior venture, Figure AI, raised at multibillion-dollar valuations as a humanoid robotics company, and before that he co-founded Archer Aviation and Vettery [Wikipedia, retrieved 2026]; [LinkedIn, retrieved 2026]. He has taken a company public and raised institutional rounds at scale. Hark is his first attempt at consumer-facing AI hardware.
Around him, the early roster reads as a deliberate mix of model researchers and product designers. Chowdhury anchors the design organization. Victor Gonzalez, previously at Figure, has publicly announced his move to Hark [LinkedIn, retrieved 2026]. The campus colocation with Adcock's other companies is a structural choice worth flagging: it gives Hark access to shared infrastructure and a recruiting gravity well that a standalone seed-stage lab would struggle to match.
Traction, in the conventional sense, does not yet exist. There is no shipped device, no disclosed pre-order book, and no revenue figure in the public record. What Hark has instead is a hiring velocity, a funded runway sized like a Series B, and a stated ship date for its first model this summer [scommerce, 2026].
The honest counterfactual
What bears say: consumer AI hardware has been an unforgiving category, and the named competitors Humane and Rabbit have demonstrated how quickly a launch can outpace the underlying product experience. A single founder writing a $100 million check removes the discipline that outside investors typically impose on roadmap and burn. If the first model lands flat this summer, there is no syndicate to call for a bridge.
What bulls answer: that is precisely why the structure is what it is. Self-funding lets Hark take the long view on form factor and model quality without quarterly pressure to demo. Adcock has shipped hard physical products at Figure and Archer, the team includes designers and researchers from the companies that have actually built reference consumer hardware and frontier models, and the campus model compresses iteration cycles. The competitive pressure is real, but Hark is not entering the category on a shoestring.
What to watch
The summer model release is the first real data point [scommerce, 2026]. It will tell the market whether Hark is a model company that also makes hardware, or a hardware company that also trains models, and the two paths imply very different capital needs from here. Watch for the first device tease, the size of the next hiring wave beyond the current 45 [TechCrunch, March 2026], and any signal that Adcock is willing to take outside capital. A priced round at this team's pedigree would reset the comparable set for the entire AI device category.
The question for readers: when a founder funds his own seed at nine figures and hires the designer of the iPhone Air, is that a signal of conviction the market should price in, or a signal that the market itself was not ready to price the bet at all?