Hark Raises $700M at $6B Valuation: Brett Adcock Wants to Build the Universal Interface Between Humans and AI
Brett Adcock's Hark (Figure.AI, Archer Aviation) raised $700M Series A at a $6 billion post-money valuation on May 21, 2026. Round led by Parkway Venture Capital with NVIDIA, AMD Ventures, Salesforce Ventures. Mission: AI-native hardware steered by speech, vision and persistent memory.
On May 21, 2026, Hark — Brett Adcock's new AI startup, the founder behind Figure.AI and Archer Aviation — announced a $700 million Series A at a $6 billion post-money valuation. The round, one of the largest Series A in history, is led by Parkway Venture Capital with a consortium of 12 investors: NVIDIA, AMD Ventures, ARK Invest, Brookfield, Greycroft, Intel Capital, Prime Movers Lab, Qualcomm Ventures, Salesforce Ventures, Align Ventures, Tamarack Global. The mission: build the universal interface between humans and AI. And redefine what a device even is. Full breakdown.
Brett Adcock: The Track Record That Forces the Check
Hark isn't an ordinary startup — because its founder isn't an ordinary founder. Brett Adcock's resume:
- Vettery (jobs marketplace) sold to Adecco in 2018 for $100M.
- Archer Aviation (electric eVTOLs), IPO'd in 2021 above $3 billion.
- Figure.AI (humanoid robots), valued at $40 billion in early 2026 after its OpenAI partnership and subsequent break, now in partnership with Microsoft.
When Adcock puts $100 million of his own money into Hark in late 2025 and closes a round at $6 billion 18 months after launch, VCs don't ask the same question they'd ask any other founder. It's no longer "is this doable?" but "how much of the allocation can we grab?". Exact same pattern as Jeff Bezos's Project Prometheus that raised $10 billion at a $38 billion valuation in physical AI.
The Pitch: Not a Device, an Interface
Hark refuses to be framed as a hardware maker. Official pitch: "intelligence paired with next-gen hardware, designed as the universal interface between humans and machines".
Concretely, Hark is building in parallel:
- Proprietary foundation models (in-house, not an OpenAI or Anthropic wrap).
- Software systems for agentic orchestration (likely MCP-aware).
- AI-native hardware — a device designed from scratch for this interaction.
- New interfaces — speech, vision, persistent memory.
The most obvious comparison is the Humane AI Pin (which flopped) and the Rabbit R1 (which flopped too). But Adcock arrives with two things they didn't have: $20 billion in combined funding across his previous companies and an ex-Apple iPhone lead designer, Abidur Chowdhury, poached in early 2026.
Abidur Chowdhury: The Design Signature That Changes the Math
Hiring Abidur Chowdhury as VP Design at Hark isn't anecdotal. He's the man who led iPhone industrial design at Apple for eight years. On a market where consumer AI hardware has failed twice in two years, having an iPhone designer at the helm changes what VCs and early users anticipate.
It's also a talent-market signal. As Apple, Google and Meta struggle to retain their top designers while seniors are leaving Big Tech for AI startups, a company like Hark grabs the top design tier with little effort. Consumer hardware is becoming the most exciting topic of the decade — for the first time since the iPhone launched.
Why Now?
Hark publishes its round exactly two days after Google launched Gemini Spark. That's no coincidence. Spark proves that persistent 24/7 agents have a use case. Hark bets on what comes next: a device dedicated to those agents, that follows you, listens to you, sees you, and keeps the memory of your context.
The persistent personal-agent market is splitting into three layers:
| Layer | Players | Approach |
|---|---|---|
| Model | OpenAI, Anthropic, Google, Mistral | Frontier LLM (Opus, GPT, Gemini, Medium) |
| Platform | Manus Cloud, Claude Dispatch, Gemini Spark | Persistent VM + native integrations |
| Hardware | Hark, Figure.AI, Project Prometheus, Apple (rumored) | AI-native device (audio + vision + memory) |
Hark targets the hardware layer — least competitive, most risky, most strategic long-term. If it works, it's the next platform in the iPhone sense.
The Roadmap: Models in Summer, Hardware After
Hark announces a two-phase rollout. Summer 2026: first models open to beta testers, early access to the personal platform. Then (no public date): launch of the first AI-native device.
The approach is unusual in consumer hardware: Hark wants to validate the intelligence before selling the hardware. Makes sense — the Humane AI Pin flopped because the AI behind it wasn't ready. Adcock won't repeat that mistake.
The Investor Consortium: A Rare Industrial Signal
The cap table of this Series A is more interesting than the dollar number. NVIDIA, AMD Ventures, Intel Capital, Qualcomm Ventures in the same round — extremely rare. These four silicon suppliers are normally direct competitors on GPUs, CPUs and AI accelerators. When they co-invest in the same device, it means two things:
- No one wants to miss the next device. The iPhone made Qualcomm's fortune. No supplier wants to be off the next platform.
- Hark will likely use hybrid architectures. A device blending NVIDIA GPU for cloud inference, Qualcomm NPU for edge, and custom accelerators (Intel or AMD) for specific workloads.
Add Salesforce Ventures validating the enterprise stack. And ARK Invest lining Hark up for the next public-tech cohort.
What Hark Says About the AI Hardware Market in 2026
The Hark round validates a thesis that's been circulating for 18 months but few VCs dared activate: AI-native consumer hardware is coming, and whoever ships it first wins the decade. Three reads.
First, AI software alone isn't enough anymore. Users want a device that listens, sees, keeps memory — without unlocking a phone for every interaction. As long as we're inside the smartphone, we're in Apple's paradigm. Hark wants out of the paradigm.
Second, capital concentrated on few bets. Like Anthropic at $900 billion, Hark at $6 billion in Series A shows VCs are no longer diversifying — they're over-funding the 5-10 bets deemed strategic.
Third, serial founders become the filter. Without Adcock's track record (Vettery, Archer, Figure), Hark doesn't close $6 billion in Series A. The AI hardware market has become so capital-intensive and risky that VCs no longer take the bet on a first-time founder. That's a structural shift in 2026 venture capital.
Key Takeaways
- Hark, Brett Adcock's AI hardware startup (Figure.AI, Archer), raises $700M Series A at a $6 billion valuation on May 21, 2026.
- Round led by Parkway Venture Capital with NVIDIA, AMD Ventures, Intel Capital, Qualcomm Ventures, Salesforce Ventures, ARK Invest and 6 others.
- Mission: build the universal interface between humans and AI — proprietary models, agentic platform, and AI-native hardware developed in parallel.
- Abidur Chowdhury, ex-iPhone lead designer, owns design — a strong signal of consumer ambition.
- Rollout: models in summer 2026, hardware after. Adcock wants to validate the intelligence before selling the device.
Hark isn't a hardware startup. It's an industrial thesis: the device after the iPhone will be AI-native — and whoever builds it wins the decade. At $6 billion in Series A, the market just voted. More details in the official BusinessWire release and the TechCrunch article.
