ARM CEO interprets the company's business model, business relationship with Apple

There is an interesting strange phenomenon in the modern technology world: There is a company that does not produce anything, but it dominates the chip world. Its technology lives in your phone, in your TV, in your car, even in your laptop, and in the data centers that power it all.

It's ARM (ARM), a chip design company that has been through a lot over the past few years. ARM mainly designs the instruction set for modern chips, and customers include Qualcomm, Apple, and Samsung, whose chips are based on ARM architecture. In terms of business model, ARM licenses the instruction set to these companies and collects royalties from these companies. With ARM's architecture, these companies can produce a variety of custom chips according to their needs. This model of ARM has achieved great success.

ARM CEO Rene Haas accepted an interview with foreign media on Tuesday to discuss the company's business model, business relationships with customers such as Apple, sales, listing, and other topics.

ARM chips are everywhere

Haas explained that ARM is not a well-known company, and the outside world does not know us very well, but we consider ourselves very important. First, ARM is in the semiconductor value chain of the semiconductor world. Basically, you can find ARM's technology in almost any type of semiconductor product and/or original equipment manufacturer (OEM) product. Our technology covers smartphones, laptops, and smart TVs. Looking around my desk, there are probably ARM processors everywhere. We don't actually produce anything. We design for the product, which is intellectual property. We do the design, instead of making a chip, we license the design to those who will make the final product. Our most famous main product is the microprocessor, or CPU, which is the brain of almost all electronic devices.

We don't make chips, we license this brain to others. So there are ARM chips everywhere. Looking at data from last quarter, all semiconductor companies and OEMs around the world made and shipped a combined 7.4 billion chips with ARM CPUs, GPUs, or other technologies built in. That's a huge number. So we're in the semiconductor value chain, but we don't manufacture anything. We do design, and most of our designs are microprocessors.

Clients cover almost all large companies

In the electronics industry, it's probably easier to say who isn't our customer, Haas said. Almost every company you can think of is our client. TSMC, Samsung, and GF are entity companies that produce chips. Intel, AMD, Nvidia, Qualcomm, Amazon, Microsoft, and Google are also our customers. Then, in the rest of the world, our customers are Alibaba, Tencent, and ByteDance. Almost every company is our client.

Our business model has two components. We start by charging an upfront license fee, which is what our partners pay us to use the technology. This gives them the right to design using our technology. If those designs end up in production and become the final product, we'll charge a royalty per unit based on some math associated with the contract. So, we have two revenue streams: one is licensing revenue and the other is royalty revenue.

Consumers buy devices that contain chips from Samsung, Qualcomm, and Apple, but instead of paying ARM, the companies pay. For example, Qualcomm would report to ARM how many chips they sold and then pay ARM based on a pre-agreed royalty rate.

The importance of the principle of neutrality

Haas said ARM will try to remain neutral. As we all know, we are known as the "Switzerland of electronics" (Switzerland maintains a neutral policy), which is a good metaphor. We don't try to pick winners, dabbling in the various sub-ecosystems within the ecosystem. If you start at the bottom of the semiconductor chain, you have GF, Samsung, TSMC, Intel, and all the people who make chips, and you have to work with all of them. We have to make sure our technology can be built on every semiconductor process in the world, and that requires investment from all these partners. And then, as we move up to Android, Linux, Windows, and all the major operating systems that we support, we have to make sure we're in there too.

Many chip companies rely on TSMC, so how much does ARM rely on TSMC? "We work closely with TSMC," Haas said. All foundries are important to us and the company remains neutral.

Why change your mind to go public?

Before Haas became ARM's CEO, his predecessor was adamantly opposed to going public because once the company went public, it would be under pressure to boost revenue, putting ARM's neutral, fair model in jeopardy as the company could strike special deals to increase revenue.

Haas responded that at the end of last year, when the Nvidia acquisition basically fell apart, we made a change and announced the decision to seek an IPO. When I took office in mid-February and ARM's fiscal year ended in March, we were finally able to talk about our financial performance, which we haven't talked about for a while. We were very quiet during Nvidia's acquisition hunt. When we reported our last fiscal year revenue, our revenue was well over $2 billion. We generated $2.6 billion in revenue and we've never exceeded $2 billion in the past. Our operating margin was close to 40%, but people thought we were losing money because we didn't disclose more. This quarter, you'll find that those numbers are even higher than last year: operating margins were over 50% on revenue of $700 million and $450 million in royalties.

Haas has said it plans to go public in March 2023. As time draws near, will ARM be able to deliver on this goal? "We're in the process right now, and I can't say too much," Haas said.

Will ARM retire the x86 architecture?

When asked if ARM and RISC-V architectures will make the old x86 obsolete, Haas said it's hard to tell. x86 has been around for a long time and has a very large installed base. I certainly don't want to do anything to discredit what they do. I think the challenge with x86 is that it only comes from two companies (Intel and AMD). That in itself is a limiting factor, limiting how far it can go.

Post a Comment