OpenAI's $60M bet on webcam maker Opal

AI giants are searching for physical bodies to house their intelligence, and OpenAI just set its sights on a boutique webcam maker.

According to a report from The Information, OpenAI is expected to lead a $60 million Series B funding round for Opal Camera. Existing investors, including Founders Fund and Kindred Ventures, are also slated to participate.

While Opal plans to keep producing its signature webcams, the fresh capital will fund a new division: building dedicated AI hardware designed as "creative tools" for everyday consumers.

High-end optics meet on-device compute

Opal made a name for itself in the remote-work era with ultra-compact, design-forward hardware.

Its current flagship is the Tadpole, a $149 tiny clip-on webcam weighing roughly the same as an AA battery. Despite measuring just 1.25 inches across, it shoots 4K video and connects to a software companion app called Composer for fine-grained lighting adjustments.

Before Tadpole, Opal launched the C1, a webcam twice the price that ran machine learning right on the device. It packed an Intel Movidius Myriad X vision chip with a dedicated Neural Compute Engine and 16 programmable cores to handle edge vision tasks like background blur and image tuning.

In short, Opal already knows how to squeeze neural compute chips and precision lenses into tiny form factors.

OpenAI's growing hardware ecosystem

OpenAI has backed more than a dozen startups through its venture fund over the last three years, largely to get its models embedded in the wild. Just this week, portfolio company Figure AI demoed a humanoid robot executing voice commands powered by a custom OpenAI model.

Opal's upcoming gadgets are expected to hook directly into OpenAI models. Given the explicit focus on creative workflows, deep integration with OpenAI image and video generation tech seems like the obvious endgame.

This push comes alongside OpenAI's broader hardware ambitions. Reports surfaced last month that the company has been in talks with Broadcom to build custom silicon, even snapping up former Google engineers who worked on the TPU.

Why it matters

Chat interfaces only get you so far. If OpenAI wants multimodal models to genuinely assist creative workers, those models need eyes and ears that sit natively in the physical workspace.

We think backing Opal is a remarkably practical move. Building consumer hardware from scratch inside a research lab is notoriously brutal. Partnering with a team that has already proven it can design, manufacture, and ship polished hardware radically cuts execution risk.

For creators and remote workers, an ambient desktop sensor powered by cutting-edge vision and generative models could transform desk work into an interactive studio. Still, consumer AI hardware has a terrible track record recently. The pressure is on Opal to prove that an AI creative gadget can be a daily utility rather than an expensive gimmick.