Hark Raises $700M Series A at $6B Valuation

Hark, a secretive AI hardware lab, said it raised more than $700 million in a Series A round that values the company at about $6 billion post‑money, the startup announced May 22, 2026.

The round was led by Parkway Venture Capital and attracted a mix of strategic and ecosystem investors, including ARK Invest, AMD Ventures, Intel Capital, Qualcomm Ventures, NVIDIA, Brookfield, Greycroft, Prime Movers Lab, Salesforce Ventures and others, according to reporting and the company release.

Hark frames itself as an AI lab building a vertically integrated stack that pairs foundation models, software and purpose‑built hardware to create what it calls a “universal” personal intelligence interface. The company says the approach aims to deliver more personalized, on‑device experiences than cloud‑first assistants.

Founder and CEO Brett Adcock, previously known for Figure AI, is steering the effort and told reporters the funding will accelerate both model development and hardware design work that the company plans to ship later. Hark’s announcement and multiple outlets described the financing as oversubscribed.

Company materials and coverage say Hark will begin releasing its models to users this summer, with hardware devices following after model launches are proven in the field, a timeline that underscores the lab’s model‑first but hardware‑native strategy.

The investor list highlights why chipmakers and their venture arms joined the round: several participants are directly tied to semiconductor design and supply chains, signaling an appetite to influence or align with hardware‑native AI use cases. That mix contrasts with purely cloud‑centric investments.

The size and structure of the deal are notable — a $700 million Series A is unusually large and mirrors a broader trend of deep early‑stage bets in AI infrastructure and hardware this cycle, according to industry reporting. Observers say such capital both de‑risks expensive hardware roadmaps and raises expectations for near‑term product milestones.

Hark’s thesis — that tightly coupling models and silicon will unlock better personalization and lower latency — is familiar to a subset of founders and investors who argue that some AI experiences require control of the whole stack. Critics caution that manufacturing, thermal design and supply logistics make scaled hardware a hard operational challenge.

For chip and systems partners, the deal is both strategic and experimental: participating corporate backers can learn whether Hark’s software and models favor particular accelerators, memory architectures or power budgets, and they can position themselves early if the product-market fit appears. Reporting suggests that alignment with hardware vendors was an explicit part of the round.

Investors and analysts say the round also reflects a duality in investor thinking — many back cloud‑first models at scale while a smaller number are willing to fund vertically integrated hardware plays that promise differentiated user experiences. Hark’s financing is an example of the latter drawing meaningful capital.

Operationally, the cash should give Hark a multi‑year runway to build models, prototype devices and secure supply deals, but it also sets an aggressive bar: backers will likely watch user adoption of the initial models, hardware manufacturability and per‑unit economics. Hark’s public statements foreground those milestones.

What to watch next is straightforward: whether Hark’s model releases this summer meet usability and personalization claims, whether hardware timelines hold, and if the company can convert strategic investor interest into production partnerships or distribution channels that scale beyond early adopters. The funding round makes those answers urgent.