Bloomberg broke the news a little while ago that Apple (AAPL +1.00%) is working on a dedicated processor to handle artificial intelligence, or AI, related tasks internally referred to as the Apple Neural Engine.
This chip, Bloomberg reports, would "improve the way the company's devices handle tasks that would otherwise require human intelligence -- such as facial recognition and speech recognition."
Image source: Apple.
"The Apple AI chip is designed to make significant improvements to Apple's hardware over time, and the company plans to integrate the chip into many of its devices, including the iPhone and iPad," the report continues.
One particularly juicy tidbit from the article is that Apple "has tested prototypes of future iPhones with the chip," citing an individual "with knowledge of the matter," though Bloomberg says, "it's unclear if the component will be ready this year."
In this column, I'd like to offer some speculation on when Apple will release this chip into the wild.
It'll either be this year or next
Since Bloomberg reports that Apple has "tested prototypes of future iPhones with the chip," the chip must be pretty far along in development. The development cycle for a chip (or, as I believe is the case here, an intellectual property that'll be integrated into the main A-series applications processors rather than a separate chip) is quite long, involving many steps.
If Apple already has this chip, or a chip incorporating this chip, in-house and suitable for testing in prototype iPhones, then the company is likely very far along in the development of the technology.
Now, having a chip prototype in hand doesn't mean that it's ready for mass production. The silicon that Apple has might not work at full performance, could have bugs and so on.
Or, alternatively, Apple may simply have completed the design of the Neural Engine after the window of opportunity to include it in the upcoming iPhone closed. If the Apple Neural Engine is a separate chip, then that window would have been larger than if it were going to be integrated into the A11 Fusion chip (given what are relatively long chip development cycles).
At any rate, I don't think it'll be too long before we see the first version of the Apple Neural Engine make it into the hands of consumers. There may be a chance that it arrives in this year's iPhones, but that chance is probably small. However, if it doesn't, then I'd certainly expect it to be a part of the 2018 iPhone lineup.
Another feature to keep building
Perhaps the most exciting thing about Apple's upcoming Apple Neural Engine isn't necessarily the first version of the technology. I suspect that rather than enabling new functionality, the first iteration will mainly be used to run pre-existing tasks more efficiently.
However, over time, I expect Apple to iterate on this technology, leveraging the benefits of experience and continued chip manufacturing technology innovation, to enable AI-related functionality that's so advanced that it couldn't reasonably be run on either the CPU or graphics processor embedded in the A-series processors.
