OpenAI has acquired Glass Imaging, a startup that uses neural networks to improve smartphone camera image quality, in a deal worth more than $300 million, according to a report from the Wall Street Journal. The price marks a sharp jump from the company’s valuation of about $100 million in a funding round last year. TechCrunch confirmed the report, and OpenAI has not issued its own statement on the terms.
Who built Glass Imaging
The company was founded in 2019 by Ziv Attar and Tom Bishop, two former Apple engineers who led the team that developed Portrait Mode for the iPhone. Their premise at Glass Imaging was that smartphone camera improvements were hitting a physical ceiling. Sensors can only get so large inside a phone body, and adding more lenses brings diminishing returns.
Their answer was GlassAI, a neural image signal processing system that processes bursts of raw image data and corrects lens aberrations, sensor imperfections and noise at the moment the shutter fires. Unlike generative AI photo editing, the software is designed to recover detail from real sensor data rather than invent content. The company claimed in May 2025 that its technology could boost camera performance by a factor of ten, and it licensed the software to manufacturers. Honor smartphones shipped with the technology, and the company demonstrated real-time video enhancement on other platforms.
Attar described the company’s mission plainly in a 2025 announcement: unlock the full potential of all cameras to deliver detailed, sharp imagery without hallucinating detail. That distinction, restoration versus generation, matters commercially, because phone makers cannot afford cameras that invent faces or textures that were not in the scene.
What OpenAI gets
OpenAI has been building toward consumer hardware for two years. In 2025 it bought io, the device startup founded by Sam Altman and Jony Ive, for $6.5 billion, and rumors have since pointed to smartphones, earbuds and companion devices in development. Glass Imaging’s team and patents add camera expertise that OpenAI does not have in-house, and computational imaging sits directly at the intersection of the company’s models and the physical devices it wants to build.
A camera system that understands what a sensor sees, combined with models that understand what users want to do with what they see, is an obvious pairing for an AI device. Voice assistants live in earbuds; the visual equivalent needs camera software that works in real time on small, power-constrained hardware. GlassAI was designed for exactly that constraint set, running on edge processors rather than in the cloud.
There is also a defensive angle. Google’s Pixel line set the standard for computational photography, and Apple’s image pipeline is a decade ahead of everyone else’s. Any new entrant shipping a phone without comparable camera software starts from a visible deficit. Buying the team that built Portrait Mode is the fastest way to close that gap.
The deal economics
Glass Imaging raised roughly $31.5 million in total, including a $20 million Series A led by Insight Partners in May 2025 and a $9.3 million round led by GV in 2024. A $300 million-plus exit represents close to a tenfold return on total fundraising for investors. Dealroom notes the company had 38 employees and modest revenue, meaning the price is almost entirely for the team and the technology.
Insight Partners and GV, which led the two most recent rounds, stand to gain the most from relatively recent bets. The Los Altos company’s technology had also been pitched to drone makers and wearable manufacturers, giving the acquisition a wider application surface than phones alone.
The acquisition also fits the broader pattern around OpenAI’s expected initial public offering, reportedly planned for 2027. Recent purchases include io for about $6.5 billion and Statsig, an experimentation platform, for $1.1 billion. Acquiring capabilities ahead of a public listing builds out a story for investors that goes beyond chatbot subscriptions.
What it means for the market
For smartphone makers, the deal removes one of the more interesting independent suppliers of camera AI. Honor was a public customer, and other manufacturers had evaluated the technology. Whether OpenAI will continue licensing GlassAI or pull it exclusively into its own devices is unknown, and the answer matters for the Android ecosystem, where camera quality is a primary differentiator.
For Apple, the deal is a reminder of how much talent and IP has left the building. Portrait Mode’s creators now work for the company building the most credible alternative to the iPhone’s AI stack. Google has its own computational photography team and is unlikely to be affected directly, but the mid-tier Android market just lost an option.
Independent camera software vendors face a changed landscape too. With OpenAI internalizing one of the few proven neural ISP providers, manufacturers seeking similar capability now have fewer places to buy it, which strengthens the remaining players and in-house teams.
Timeline and unknowns
Neither OpenAI nor Glass Imaging has announced a closing date, and acquisitions of this size typically take weeks to complete regulatory review. Attar and Bishop are expected to join OpenAI along with the team, though neither has made a public statement since the report. What OpenAI plans to do with the technology in its first-party devices has not been detailed, and the company has not confirmed a launch window for any hardware product.
The deal closed the loop on a startup that spent seven years building niche imaging software for other people’s phones. Its next act will be inside whatever OpenAI ships.