Apple Embarks On New Strategy Of Using Inhouse-Designed Chips

It’s been well documented that the new line of Macs unveiled by Apple this week would be powered by new custom chips it has designed inhouse. After 15 years of using Intel chips, the Californian tech giant is going it alone. What was a surprise was the claims made by the company – that a new MacBook Pro running on Apple’s M1 chip is able to run for up to twice as long on a single charge, 20 hours, compared to an equivalent Intel-powered machine.

Manufacturer claims over battery life, with tests often conducted under very restricted conditions unlikely to be replicated in practical scenarios, are notoriously divergent from what actual users experience. And Apple’s claims have yet to be poked and prodded at through independent testing.

But if Apple is right, or not far off being right, doubling the running time of hardware thanks to a new, proprietary chip architecture, would be a major development. And a major competitive advantage for Apple. Proprietary semiconductor IP will immediately become a new battleground between the biggest technology companies. Or at least those that sell hardware.

Over the past 2-3 decades, since the beginning of the mainstream PC era, computing has been split along the lines of hardware makers, software makers and specialist chip companies. It worked pretty well with specialist chipmakers succeeding in multiplying the processing power of semiconductors by many times over the years.

However, Apple’s new approach may herald a new era of greater vertical integration. It hasn’t happened by coincidence, with Apple’s move into chip design catalysed by evolving computing needs. Big factors to those changing needs have been the evolution of smartphones and cloud computing.

These technologies have paved the way for the age of AI, big data and internet of things (IoT). With power-hungry AI now part of almost every device, improving performance while simultaneously reducing power consumption is high on the list of priorities for chip designers, and their customers.

Customisation is the quickest way to gain a competitive edge in the contemporary tech landscape. An example of this is Tesla’s approach to driverless technology, which is a bet it can built a tech stack to achieve autonomy based only on AI analysing images from the car’s cameras. Most other companies pursuing the technology think Tesla’s approach is optimistic.

But Tesla believes the computer carried by each car, running on chips it has designed in-house for the exact role they will perform, means it has a good chance of proving the doubters wrong. With the key the fact the chips are customised to the exact needs of software and hardware they will power.

Amazon and Google also now design their own AI chips for their cloud computing centres and Facebook is also working in that direction. China’s tech giants, such as Alibaba also now have their own AI chips.

Specialist manufacturers that produce chips for clients, rather than design chips themselves, is also a factor which is supporting the new trend. Apple’s M1 chips are being made by Taiwan Semiconductor Manufacturing Company (TSMC) and their manufacturing process is more advanced than what Intel can offer.

The big tech companies still rely on chipmaking specialists to deal with many of the elements outside of design and manufacturing. Google work with Broadcom and Microsoft has ‘co-designed’ chips with Qualcomm.

It’s not clear if big tech will ever start to compete directly with the specialist chipmakers, by selling their own more general-purpose chips to third-party buyers. Rather than customising proprietary models to the specific needs of their own hardware. But proprietary chips do look set to become a differentiator between big tech competitors and a new area of competition.