Apple has acquired another AI startup, which may have confirmed the tech giant's goal to process everything on-device. It's a continued push towards generative AI as the unveiling of iOS 18 approaches.
As reported by French publication Challenges, Apple has officially acquired Paris-based startup Datakalab. According to official filings, the acquisition was completed back in December 2023. Pre-acquisition, Datakalab wasn’t just playing in the minor leagues. The startup was working with big names like the French government and Disney.
Datakalab, founded in 2016, has been cooking up some impressively efficient AI algorithms. Rather than your average energy-guzzling, server-farm-hogging AI model, this is a slicker, low-power, high-efficiency model. The goal? To snugly fit into your handheld gadgetry and process everything on-device.
That means there's no need to phone home to the cloud every five minutes, which leads to long wait times for responses. Just ask Humane about that. It also keeps all your data local and secure, without any risk of being infiltrated. On-device AI processing is perfect for Apple, whose mantra is very firmly "privacy first."
On-device AI plans confirmed?
The news of Apple's Datakalab seemingly confirms Apple's plans to use on-device processing when it rolls out AI features. It's just the latest indication of the tech giant's goal.
Over the weekend, Bloomberg’s Mark Gurman revealed that Apple’s next iteration of AI might skip the cloud commute altogether, keeping things strictly on-device. He previously highlighted Apple's recent acquisition of DarwinAI, who also specialises in on-device AI processing. Gurman also goes on to say that a chip will also help with generating the results – which could allude to what the iPhone 16 Pro’s powerful chip upgrade could feature.
Apple is said to be adding generative AI features to the upcoming iOS 18 update for the best iPhones which is likely to be revealed at WWDC 2024 in June. Apple is expected to focus on how these features can help you in your daily tasks at WWDC, instead of simply explaining how powerful its LLM (Large Language Model) is.
