The founder who built Figure AI's $39B robot empire is now targeting the device in your pocket and he's brought Apple's iPhone Air designer along with $800M -- TFN

The AI hardware market is projected to grow from $83.4 billion in 2025 to $361.7 billion by 2035, according to Cervicorn Consulting, at a CAGR of 15.8% and Hark enters it before shipping a single product.

When Abidur Chowdhury walked Apple's stage to introduce the iPhone Air, few expected him to leave within months. He did, after a single meeting with Brett Adcock. Now he is Director of Design at Hark, a San Jose-based AI lab that Adcock founded in late 2025 with $100 million of his own capital and a conviction that every AI product built so far has been wrong.

Hark emerged from eight months of stealth on 24 March 2026. The company has 45 staff on the same campus as Adcock's other ventures, with headcount targeted at 100 by mid-2026. It has disclosed no product, no price, and no hardware form factor, only a timeline: AI models this summer, native hardware shortly after.

Most AI companies pick a layer of the stack: models, software, or hardware and work from there. Hark builds all three simultaneously. The company is developing multimodal foundation models capable of processing speech, text, vision, and contextual memory, alongside purpose-built consumer hardware designed specifically for those models. The stated goal is an always-on personal AI that remembers your life, anticipates tasks, and operates as a background layer beneath everything you do rather than an app you open.

The problem Adcock targets is mundane but expensive in time: the cognitive overhead of daily coordination. Booking travel, filling forms, planning home renovations, switching between fragmented apps. "The AI systems I use today are far from my vision of what the future should be," Adcock said in the company's launch statement. "We want to create intelligence that lets you offload your mental workload into a system that begins to think like you and sometimes ahead of you."

Chowdhury frames the design problem in terms of stack position. "Very few people are really going after what the future is," he told TechCrunch at launch. "There's so much that we could be doing if intelligence was at the base layer of everything we touched instead of becoming an app or a website at that upper layer." He has explicitly ruled out wearables as Hark's direction, distancing the company from the form factors that have defined and in most cases killed, previous AI hardware attempts.

The AI research bench includes alumni of Meta's Superintelligence Lab. The hardware team draws from Apple, Tesla, and Meta, with engineers covering mechanical, electrical, firmware, embedded systems, and supply chain. Hark has also signed a deal to bring thousands of Nvidia B200 GPUs online in April to support multimodal pre-training.

Nvidia CEO Jensen Huang commented on the partnership: "The new era of personal AI will be defined by intelligent agents that understand context, reason across modalities, and act on our behalf. Bringing that vision to life requires enormous compute to build powerful multimodal foundation models, and we're excited to support Hark's work with NVIDIA accelerated computing."

Hark's $100 million seed is entirely personal capital from Adcock, who built his fortune through Vettery (sold to Adecco), Archer Aviation (public via SPAC in 2021), and Figure AI, the humanoid robotics company currently valued at $39 billion. No external investors participated in the seed round. Since launch, Hark has raised a $700 million Series A at a $6 billion valuation, led by Parkway Venture Capital, a New York firm founded in 2019 that previously led Figure AI's $1 billion Series C and has also backed SandboxAQ and xAI. Co-investors include Nvidia, AMD Ventures, Intel Capital, Qualcomm Ventures, Salesforce Ventures, ARK Invest, Brookfield, Greycroft, Prime Movers Lab, Align Ventures, and Tamarack Global. Total funding to date stands at $800 million.

The AI hardware category has a short, painful track record. Humane's AI Pin launched in 2024 with $240 million in funding and an $850 million valuation; by February 2025 it had shut down, with HP acquiring its assets for $116 million. Rabbit's R1 device launched to similar criticism. The clearest large-scale competitor is OpenAI, which acquired Jony Ive's hardware startup io Products in May 2025 for $6.5 billion and is developing a screenless ambient AI device expected in late 2026. Meta's Ray-Ban smart glasses represent the one AI hardware product to gain meaningful consumer traction so far, but operate in a narrower use case.

Hark differentiates itself from all three on one point: vertical integration from day one. Humane and Rabbit built devices around third-party models that were not capable enough to justify the hardware. OpenAI is integrating models and hardware retroactively through an acquisition. Hark is designing models, software, and hardware as a single system from the start -- a bet that the interface advantages of that approach outweigh the capital and execution risks of doing everything at once.

The global AI in hardware market was valued at $83.4 billion in 2025 and is projected to reach $361.7 billion by 2035, growing at a CAGR of 15.8%, according to Cervicorn Consulting. Mordor Intelligence puts the edge AI hardware segment specifically at $30.7 billion in 2026, growing to $68.7 billion by 2031 at a 17.5% CAGR -- the segment most directly relevant to the always-on, on-device AI Hark is building.

Hark's first AI models will tell investors and the market whether the foundational layer of the company is competitive. The hardware question: what it looks like, what it costs, and whether anyone buys it, comes second. In a category where every well-funded, design-credentialled team before it has either shut down or not yet shipped, the more interesting question may be whether vertical integration is the solution, or just the most expensive version of the same problem.