On Google, Apple and the computer industry’s struggle for innovation
People who care about software should make their own hardware.
— Alan Kay
A couple of weeks ago, we read some rumors about Apple’s plans to use their own chips in Macs. Just some weeks prior to that, there were similar rumors about Google working on their own processors for phones and laptops.
For many years, Apple really has been taking Alan Kay’s quote to heart and with this strategy, they started dominating innovation in the computer industry.
Why other companies could not keep up with Apple
Apple’s A-series processors in iPhones are wiping the floor with the Qualcomm processors used in Android phones ever since the 64-bit transition.
But it is not just with the iPhone that they are leading the pack, they also dominate with their laptops.
The touchpads on Macbooks are so far ahead of anything the other laptop vendors had, that Microsoft themselves stepped in — they developed the “Precision Touchpad” for the laptop vendors to close the gap. In a similar way, it was Intel who pushed the development of the “Ultrabook” — the answer to Apple’s sleek Macbook Air devices.
Intel and Microsoft make more profits from Laptops and PCs than the companies who actually make these devices. Why? Because laptop hardware is commoditized, while the operating system (Microsoft) and processors (Intel) are monopolized.
But since Microsoft and Intel take such a big share of the profits, they are also responsible for driving innovation. The device manufacturers down the line simply do not have the money and the mindset for R&D. The R&D they do is often more focused on improving the assembly line than the product, since this is the way they can squeeze out some more profits.
Apple on the other hand keeps most of the profits made from their products and does not have to share it. They can reinvest it in product innovation.
And while Apple has been doing that for years now, recent rumors show that Google is about to do the same by creating their own processors. With severe consequences for the whole industry.
Apple’s and Google’s model of innovation puts the industry at risk
Google moving into their own processors is a big problem for Qualcomm, in the same way that all vertical integration is a problem for suppliers. Today, Qualcomm has 45% market share in smartphone processors, but they are still struggling to keep up with Apple’s chips.
Qualcomm still has very attractive modems, e.g. the radio modules that allow a phone to connect to the 4G and 5G networks. These have long been a selling point for their processors: they included the processor and the modem in the same part, which has benefits for power consumption and makes the design easier.
But their processors are becoming less and less attractive while more and more of their customers (e.g. Google) are unbundling their offer, buying only the modems, and not the processors. Samsung already did this a while ago with their Exynos processors and Google is the next company to make that step.
Actually, Samsung is the best example for the other problem Qualcomm faces: Qualcomm does not manufacture their own chips. In fact, Samsung and TSMC are two of the biggest semiconductor manufacturers for smartphone chips. So basically, Qualcomm does not have a lot to offer anymore.
And it gets even worse for Qualcomm
The next key to the puzzle is ARM, who develop the ARM instruction set architecture for computer chips. Apple, Google, Samsung, Qualcomm and the rest of the industry all use this architecture.
While Apple is designing their own processors based on this architecture, Qualcomm uses the standard designs that ARM sells. This means that a Qualcomm processor is an ARM standard design with a Qualcomm modem, manufactured in Samsung’s factory.
What stops Samsung from getting the ARM standard designs directly from ARM and designing their own chips? Right, nothing. Samsung cuts Qualcomm out of their supply chain, their revenue and profit go up, and they can still buy the modems from Qualcomm. Good for Samsung, bad for Qualcomm.
At this point you are probably ready to sell your Qualcomm stocks to buy some ARM. But wait — ARM is facing a similar problem.
Then new Open Source suppliers come to kick ass
ARM’s value comes from a pretty good instruction set architecture and good implementations of it. But there is a new contender by the name of RISC-V, an instruction set architecture with similar capabilities as ARM. However, there is one big difference: RISC-V is Open Source.
Isn’t that great? With RISC-V, there are no license fees and there is much greater freedom in extending the instruction set. Consequence: it allows companies to create innovative new features, which can give them the edge on the market.
ARM’s big splash was to make it easier for companies to create new processors. Qualcomm was able to use their designs, combine them with their modems and have the chips manufactured by TSMC and other fabs. Now, RISC-V has the potential to make this even easier and more approachable to smaller companies. ARM has seen this threat and already started making it easier for startups to use their Cores by offering them special licenses.
A company who starts designing a new processor today and for whatever reason cannot or will not use an existing design from ARM, has a lot of good reasons to forgo ARM completely and directly bet on RISC-V.
In-housing processor design will affect Drones, VR-Goggles and many more tech innovations
Although RISC-V and semiconductor foundries do make it easier than ever for companies to innovate with new chip designs, the in-housing of processors’ design from companies like Apple or Google does have some negative consequences for innovation from smaller players.
While Apple’s chips are awesome, they are also completely irrelevant for the rest of the industry. Why? Well, you cannot buy them. If you have a product idea that would be enabled by the fast chip from Apple, you simply have no chance to actually develop it. And this issue is bigger than you might think.
For instance, two of the most interesting innovations from the last years were made possible by smartphone components: Drones and VR-Goggles. The processors and sensors used in both of them were developed for smartphones. Then, through simple economies of scale, they became cheap enough to allow people to build new products from them.
A lot of innovation is driven by that process: components that were developed for other products enhance the development of new products. But when the big players start to keep more of their development in-house, there are less parts available on the market for smaller players to come up with new products.
This, in turn, can create new chances for companies like Qualcomm: when they lose big clients to in-housing, they can try to compensate for that by being more supportive of their smaller clients. They can meet them eye-to-eye and start innovating using my favorite model: collaboration.
What about you, are you ready to collaborate?
