Apple drops $2B on secretive silent-speech startup Q.ai

If you have ever tried talking to Siri on a crowded train, you know the awkward dilemma: speak up and look crazy, or whisper and be ignored.

Apple is throwing serious cash at solving that problem. According to reports from the Financial Times and Reuters, Apple has acquired Israeli AI startup Q.ai for close to $2 billion. Johny Srouji, Apple’s senior VP of hardware technologies, even confirmed the deal on the record, praising the team's work in combining imaging and machine learning.

At nearly $2 billion, this marks Apple’s second-largest acquisition in history, trailing only the $3 billion purchase of Beats in 2014.

The art of silent speech

Founded in 2022 and backed by Kleiner Perkins and GV, Q.ai has spent the last few years operating in stealth mode. Its minimalist website promised to craft a new kind of quiet in a noisy world.

Patent filings reveal what that actually means: Q.ai uses machine learning to decode "silent" voice input. By tracking micro facial movements, subtle muscle activity, and whispered acoustics, their tech lets devices capture user intent without requiring an audible speaking voice.

If the playbook sounds familiar, look at the leadership. CEO Aviad Maizels is selling to Apple for the second time. In 2013, he sold 3D-sensing startup PrimeSense to Apple for $345 million, providing the core tech that eventually replaced Touch ID with Face ID. Now Maizels and his co-founders, Yonatan Wexler and Avi Barliya, are heading back inside Apple.

Why it matters

We think this deal shows Apple tackling the most overlooked bottleneck in AI: the physical interface.

While rivals race to build ever-larger multimodal foundation models, interacting with them in public remains awkward. You either type on glass or talk aloud like a tech executive rehearsing a keynote. Silent voice input and facial micro-motion tracking bypass that social friction entirely.

For future devices like smart glasses, camera-equipped AirPods, and next-gen Vision Pro headsets, this technology could be transformative. Imagine sending a message or prompting an AI assistant simply by mouthing the words on a busy subway platform.

The real test will be miniaturization. Turning sensitive facial-tracking algorithms into ultra-low-power chips that fit inside a tiny earbud is notoriously difficult. But given that Maizels already managed to squeeze bulky 3D sensors into an iPhone notch, Apple clearly believes he can pull off the hardware magic a second time.