Funding · 21 May 2026 · 2 min read

Hark raises $700M for a universal AI interface

Hark's massive $700M Series A signals a huge bet on a new hardware-native platform for AI, challenging the phone as the primary interface.

Pen-and-ink illustration: a single massive keystone locks a complex. For the story "Hark raises $700M for a universal AI interface".
— Pen-and-ink illustration: a single massive keystone locks a complex. For the story "Hark raises $700M for a universal AI interface". —

What happened

Hark, an AI lab building a personal AI assistant, has raised a $700 million Series A round. TechCrunch reported the funding values the company at $6 billion post-money. The round was led by Parkway Venture Capital and included investors like Align Ventures, ARK Invest, and Intel Capital.

Founder Brett Adcock launched Hark in late 2025 to develop an agentic AI system. The company, which currently has 70 employees, plans to release its first multi-modal models this summer, followed by custom hardware devices.

How the room's reading it

The size of the round for a pre-product company is turning heads. Investors are clearly betting that the next major platform shift is a universal AI interface — a layer that sits on top of existing services. Hark is framing its mission as building for consumers, a contrast to what its design director calls the developer-tool focus of labs like OpenAI and Anthropic. This narrative of helping "the normal person" is central to their pitch.

Still, practitioners are asking hard questions about execution. The problem of building a context-aware wearable that doesn't violate privacy or social norms is unsolved, as Meta's experience with smart glasses shows. Without a public demo, the consensus is one of intrigued scepticism — a massive bet on a very difficult product category.

Sailfish's take

We don't see this as a bet on another software agent. This is a bet that the phone is the wrong form factor for agentic AI. The $700 million isn't just for model training; it's for a hardware play to build a native device for a universal AI interface. We've shipped enough products to know that software-only solutions often lack the deep system-level integration and contextual awareness needed for a true assistant.

A dedicated device could solve that, but it introduces a colossal adoption hurdle. We think the interface problem is real, but the answer isn't necessarily new hardware. The more interesting challenge is building agents that work seamlessly on the devices people already own. We're watching Hark's progress, but we remain sceptical that consumers will carry another gadget.

Our take — your read?

Be the first to weigh in.

Sources
— END OF DISPATCH — Funding