Funding · 21 May 2026 · 2 min read

Hark lands $700M for a universal AI interface

Hark's $700M Series A is a massive bet on new interaction paradigms, pushing builders to think beyond text-based AI assistants.

Pen-and-ink illustration: a single, monumental keystone. For the story "Hark lands $700M for a universal AI interface".
— Pen-and-ink illustration: a single, monumental keystone. For the story "Hark lands $700M for a universal AI interface". —

What happened

AI lab Hark has raised a $700 million Series A, valuing the company at $6 billion post-money. The round was led by Parkway Venture Capital and included names like Nvidia, AMD Ventures, and Salesforce Ventures, as reported by TechCrunch.

Founded by Brett Adcock, Hark is building what it calls a 'universal interface' for the digital world. The company plans to release its first multimodal models this summer, with custom hardware to follow. The funding will go towards hiring, compute, and components.

How the room's reading it

The reaction on X is a mix of sticker shock and scepticism. A $700 million Series A for a company with no public product is turning heads. Many developers are pointing out the immense execution risk — building new models, a software platform, and custom hardware all at once is a monumental task.

The consensus among VCs and founders, however, is that this represents a necessary, high-risk bet on the next interface after the smartphone. Hark's framing of serving the 'normal person' — a contrast to labs chasing developer and enterprise tools — is seen as a direct shot at building the first truly mainstream AI consumer product. The privacy question around wearable hardware remains the biggest unanswered challenge.

Sailfish's take

We've seen this play before, but never with this much capital. The $700 million isn't the story — the ambition to own the entire stack is. Hark is betting that the winning AI product won't be a better model API, but a completely new interaction loop built on bespoke models and hardware. It's a direct challenge to the idea that chat is the final interface for AI.

We think they're right about the problem, even if the execution is a long shot. The real work for builders isn't just fine-tuning prompts. It's designing for context, ambient awareness, and low-friction action. We'd start prototyping with multimodal inputs today. The race for the post-smartphone interface just got a very well-funded new entrant.

Our take — your read?

Be the first to weigh in.

Sources
— END OF DISPATCH — Funding