Arm’s new path is beginning to take shape.
Weeks after Nvidia’s failed $40 bid to acquire Arm from SoftBank, the appointment of Rene Haas to replace longtime chief executive Simon Segars set the company on a new trajectory. Haas appears determined to shake up the company, with plans to lay off up to 15% of staff before plans to take the company public by the end of March next year.
It’s hard to argue that Arm isn’t thriving. Last week, the company reported adjusted profitability of $1 billion on revenue of $2.7 billion for fiscal 2021, the latter of which jumped 35%. Arm’s adjusted profit figure leaves a lot to think about, as the company didn’t reveal what it moved to arrive at the $1 billion figure (the company didn’t respond to a request for further details). ). Arm had $1.13 billion in licensing revenue, which comes from companies like Apple paying to use Arm designs for their custom chips, and $1.54 billion in royalty sales. Throughout the year, the company said 29.2 billion Arm-based chips were shipped.
Protocol met Haas in San Francisco three months after taking the top job and after nearly 9 years with the company, most recently leading its intellectual property group. He discussed Arm’s future plans, its research plans, and the new RISC-V, which makes chip designs that compete with Arm. Haas has his work cut out for him: In a now-deleted blog post from his corporate site, his predecessor described going public as an existential threat to the company. A spokesperson for Arm also declined to comment on the deleted post.
This interview has been edited and condensed for brevity.
Your predecessor Simon said unequivocally last year that if Arm went public, he would die. Now Arm is planning to go public, and I know you’re laying off some of the workforce, but that doesn’t solve the fundamental problem that Simon expressed – that the pressure to achieve revenue growth and short-term profitability would stifle the company’s ability to invest, grow, move fast and innovate. So how are you going to make it work?
That’s a good question, and I’ll answer it. Let’s talk about the results for a moment, because it just goes together – we’ve never made more than $2 billion in revenue in our history, so now we’ve made $2.6 [billion]. Our royalty-free number was over a billion. We have never exceeded one billion, ever. It’s up 60% year over year, which is kind of an indication of what product demand will look like in the future. And then our royalties were $1.5 billion, a record as well.
I took over the [intellectual property] in 2017, and we made kind of a fundamental pivot away from general-purpose CPUs and more into market-specific products. At the same time, we started to get rid of products: display IP, video IP, which were not very differentiated and commonplace. We’ve really doubled down on what I would call the compute platform – the CPU, the GPU, and the activities around that.
The restructuring really wasn’t about trying to fit into an EBITA envelope. It was a question of determining whether we had the right profile of [expenses] versus research and development. We need to invest more in R&D, so part of that was creating space to invest more, and that gives me a high degree of confidence – IPO or non-IPO – that we have a very healthy business.
This year we started the transition to v9 [architecture] royalties, but not in a significant way, which will start next year. V9’s royalty rates are better than V8’s, which gives us confidence. And at the same time, these new markets with hyperscalers like Amazon and Microsoft, automotive and others, I think we can manage that balance.
It’s a delicate balance, and I’m not going to dispute what Simon said because those are his words, not mine. But I am convinced that we have a solid company.
You plan to invest more in research and development, to devote more resources to it: how does Arm do it? How do you plan to allocate R&D dollars? Some companies have set specific performance goals, such as a 1,000x increase.
We are very focused on performance efficiency. Because I think one of the things that’s going to drive our growth, we’ve seen it before: more and more chips are using more and more cores.
In a multi-core system, you need a high degree of efficiency around performance per watt. And if you think about the data center, the electric vehicle, or the base station, they all need a lot of performance, but they really need performance per watt. So if Intel says 1,000x better performance, we’re definitely going to be competitive on performance, but we’re going to be relentless on performance per watt.
As for the data center – if a new data center is to be built in Ireland, say, for them to get the land and a power contract, only a certain square footage will be allowed, [and] there will only be a certain number of megawatts allowed. If they don’t compromise on performance, it’s really going to be around performance per square foot and performance per square watt. That’s why when [Amazon CEO] Andy Jassy came on stage and said, “Why Graviton2?” He talked about around 40% better performance in the same power envelope.
More broadly, for us, it’s all about the compute platform, in terms of GPU, CPU, and machine learning. Arm had a lot of other products that were on the periphery, which required a lot of engineering work, and I don’t think we were very differentiated or adding much value. While on the other hand, the software investment required for these ecosystems is somewhat insatiable. And it’s also going to be a high degree of focus for us.
Arm used to grow cores, then let other people build their own chips: how does that change over time? Does Arm develop the final design for customers more now than before?
More and more, what we’re seeing is that throwing that piece of intellectual property over the wall isn’t going to be enough to guarantee a world-class product. There are a lot of things around system design that are increasingly important: the interconnect, the physical design, the memory subsystem. We are much more prescriptive about how to build a [system-on-chip] using arm [designs].
Unlike a simple IP licensing model – here’s the recipe, go ahead and build – we’re now doing things around subsystems that have basically allowed people to build better SoCs. It will also give us better performance and performance per watt because we can now guarantee a certain performance threshold. We understand that you build in TSMC, you build in Samsung, you build in GlobalFoundries – we work with the library, and we do most of that ourselves.
When you think of high performance systems, they are really hard to build. Our world is disaggregated: there are Cadence tools, there are Synopsis tools, there is TSMC, there is a type of substrate. If you’re leaving it up to everyone to figure out what their secret sauce is, you’re going to have to compromise. Since the performance engine is most processor-bound, we feel we have an obligation to be a little more prescriptive in terms of the design of these products.
Is having more control over exactly how products are made more profitable?
It could be – I can’t say, it’s too advanced.
With only two dominant chip designs, who do you see as Arm’s biggest competitor at this point?
If you think only of the CPU industry, the way to think of the competition is the instruction set architecture. There are really only three that are in vogue. There are x86, only two companies in the world [Intel and AMD] can build. So it’s a competitor, but not in the classic sense, because it’s not open.
There is only one other open ISA, and that is RISC-V. And RISC-V is absolutely a competitor to what we’re doing, and they’re unique because they’re an open-source implementation. He’s a very interesting competitor because one of their biggest strengths might be their biggest weakness.
They are open source, so anyone can create whatever they want with them. It can be modified, it can be extended, anywhere in the world could have its own RISC-V flavor. But the more you differentiate yourself and the more you create different extensions that aren’t adopted by everyone in the ecosystem, the higher the potential for software fragmentation you get.
For RISC-V, the jury is still out – it’s early. There’s a lot of momentum around RISC-V, no doubt. Where we see them gaining traction is in what I would call the deeply embedded space, where external software doesn’t really matter. It’s controlled, in a black box. There is no developer, there are no software kits in this world. For hyperscalers, there is a lot of work to do around software optimizations and workloads. And then you find yourself in the situation of whether everyone would be happy with their own type of risk five implementation. There must be a standard.
I think the RISC-V thing will be a very interesting story to see unfold.
What do you think of your relationship with Intel? On the one hand, the company is a competitor, but on the other hand, it has opened its factory doors and is ready to manufacture anyone’s chips.
People would be surprised that Intel is one of our big customers: one of our biggest customers, believe it or not. I’d like to see Intel Foundry Services be a massive success, because when you take Arm, which is the most ubiquitous processor in the world, there’s going to be a marriage somewhere. They can’t succeed without having a solid offer and being very competitive with Arm. On the other hand, for IFS to succeed, we need to work closely with them. And we do.