Apple acquired WaveOne, a Mountain View-based startup working on AI-powered video compression. The company’s approach aimed to reduce the data needed to deliver video while keeping key details visible, even when connectivity became limited.
Why video compression is changing
Video is typically compressed by a content provider, such as YouTube’s servers, then decompressed on the viewer’s device. This division of work is effective, but new codecs can depend on specialized hardware to speed up compression and playback. That makes it harder for improvements to reach users quickly.
WaveOne set out to make video compression work with AI accelerators already built into many phones and a growing number of PCs. Its goal was to use existing device capabilities rather than rely on hardware designed specifically for a new codec.
Using AI to decide what to preserve
WaveOne’s “content-aware” system used scene and object detection to analyze video frames. It could prioritize faces over less important parts of a scene, for example, reducing the amount of data needed while retaining details viewers are more likely to notice.
The company also said its technology could respond to sudden drops in bandwidth. Instead of freezing when a connection weakened, it could make a “best guess” from the available data and show a less detailed picture until conditions improved.
WaveOne claimed its hardware-agnostic approach could cut video file sizes by as much as half, with larger gains in more complex scenes. These were claims by the startup; the article does not describe independent testing of those results.
What the acquisition could mean for Apple
Apple did not confirm the sale when asked for comment. However, WaveOne’s website had shut down around January, and several former employees, including a co-founder, were working in Apple’s machine learning groups.
WaveOne’s former head of sales and business development, Bob Stanko sh, announced the sale in a LinkedIn post published a month earlier. He wrote: “After almost two years at WaveOne, last week we finalized the sale of the company to Apple,” and said Apple had recognized the technology’s potential.
For Apple, more efficient compression could reduce bandwidth costs or help streaming services deliver higher resolutions and framerates, depending on the content. The article points to Apple TV+ as one possible beneficiary, but does not say how or when Apple might use WaveOne’s technology.
A wider push toward AI video codecs
WaveOne was founded in 2016 by Lubomir Bourdev and Oren Rippel. Before the startup, Bourdev was a founding member of Meta’s AI research division, and both founders worked on Meta’s computer vision team, which handled content moderation, visual search and feed ranking on Facebook.
The company attracted $9 million from investors including Khosla Ventures, Vela Partners, Incubate Fund, Omega Venture Partners and Blue Ivy before the acquisition. Its work also reflects a broader interest in using machine learning to make streaming more efficient.
The article notes that Alphabet’s DeepMind adapted a machine learning algorithm originally developed to play board games to compress YouTube videos. That effort led to a 4% reduction in the data YouTube needed to stream to users. Whether Apple’s acquisition leads to similar gains remains to be seen.