SoftBank, one of the world’s biggest tech companies and startup investors, believes we’ll have a trillion connected devices by 2025. And these devices, connected through the internet of things (IoT, where everyday objects have processors or sensors) will generate $11 trillion in value by that time, according to SoftBank Group chief operating officer Marcelo Claure, who spoke this week at an event held by SoftBank’s Arm chip design company, which has grown from 4,000 to 6,000 people under SoftBank’s ownership.

Arm held its annual Arm TechCon event this week in San Jose, California. The company talked about its Neoverse, which will provide a “cloud to edge” infrastructure for Arm-based devices. But there was an elephant in the room, as a cloud is hanging over Saudi Arabia, one of the biggest investors in the SoftBank Vision Fund.

Washington Post journalist Jamal Khashoggi disappeared after a visit to the Saudi consulate in Turkey at the beginning of this month, and it was later confirmed he was killed in a “fight.” Evidence is pointing toward Saudi crown prince Mohammed bin Salman, who has pledged to invest $45 billion in SoftBank’s second Vision Fund.

While the circumstances of Khashoggi’s death wasn’t the topic at hand at TechCon, it came up during a conversation between a group of journalists, Claure, and Arm CEO Simon Segers after the keynote talks. This began when a journalist asked about the incident, and Claure answered the question. Don Clark of the New York Times asked a follow-up, and then an Arm public relations person tried to shut the conversation down. Richard Waters of the Financial Times objected, saying, “We are not in Saudi Arabia. We are in the United States. This is a serious point. Open journalism is a serious part of this democracy. Do you really think we’re going to shut this down after one question? It’s just nonsense.”

The PR person relented, and in the session that took place on Wednesday, Claure acknowledged that SoftBank was “watching the developments” in the case closely.

We also talked about how SoftBank believes Arm will play a key role in bringing about the Singularity — the day a few decades in the future when artificial intelligence exceeds the collective intelligence of humanity. This is the stuff of science fiction, but Claure insisted it is grounded in today’s technological reality.

Here’s an edited transcript of our interview.

Above: SoftBank COO Marcelo Claure (left) and ARM CEO Simon Segers take questions at Arm TechCon.

Image Credit: Dean Takahashi

Question: Two years into the acquisition by SoftBank, I’d love to hear any specific examples of increases in staff or budgets for R&D, or particular initiatives.

Did Neoverse get investment from SoftBank, for instance?

Simon Segers: We’ve been investing broadly across Arm. Masa’s instruction to me on day one was, “Go, go, go!” A term he used frequently. We’ve invested a lot in our engineering headcount to deliver the products we’ve been talking about and will talk about this week. We’ve hired about 2,000 people into Arm over the last two years. That’s had a negative impact on our profitability for now, but what Masa is always looking to do is to invest now to create a huge return in the future. He’s not too worried about the short-term results, as long as there’s a plan to deliver great upside in the future.

Those 2,000 people in Arm, obviously some of that is in the infrastructure of the company, because we’re bigger now, but it’s largely into engineering. We’re about 80 percent technical across the company. People are working on our IOT platform, the variants of our processes for things like Neoverse, the other IP that our partners license from us, and all the technologies we think will be needed to deliver on that next generation of computing, machine learning being one of them.

So, very broad investment across the technology space. The benefit of being part of SoftBank and not being public is that we can invest in many of those technologies in parallel. Two years ago and a bit, I would have had to pick one or two things and place some bets. We can now do that in a much broader way than we would have been able to do previously.

Question: Can you point to anything you’ve done that you wouldn’t have done as a public company?

Segers: If we think of our IOT platform, we acquired a couple of companies earlier this summer. We’ve invested significantly in our headcount to deliver on the organic technologies we were developing. I’m sure that wouldn’t have gone as rapidly as it has done, had we been public.

Question: The 2,000 people, is that a net increase, or was there also a loss?

Segers: That’s the gross number. There’s a bit of attrition along the way, obviously. But we’re over 6,000 people now in Arm. I can’t remember exactly what the number was on the day of acquisition, but we’re just over 6,000 now.

The growth has been global. SoftBank committed that we would double our headcount in the U.K. over the five years after the acquisition. We’re on a path to that. But we’ve grown pretty uniformly in our main sites. We have about 1,200 people here in the U.S. We have 1,000 people in India. We’ve grown in Israel. We’re quite diversified in where the people are. We’re looking to have a significant presence where there are significant technology presences, so that we can tap into the local ecosystem and local talent.

Arm plans to be the glue for all tech.

Above: Arm plans to be the glue for all tech.

Image Credit: Arm

Question: What are the challenges of collaborating with SoftBank?

Segers: Every company has their own set of objectives. Everyone’s trying to go really fast. But the opportunity is to look for ways to leverage each other’s skills and technologies, and look for ways to solve problems across the group in a more efficient way.

We recently announced, with Sprint, that they’ll be using the IOT platform we’ve been building as part of their IOT offering. That’s Sprint engineering a solution very similar to what we had. It means everyone can go faster.

Question: Can you detail how the IOT partnership with Intel works?

Segers: It’s specifically looking at the challenges of onboarding devices. As you look at IOT, it’s quite complex to get devices on a network, and it’s complex to get the data to tell you something useful. Generally we’ve been looking to address those friction points. How do you get something on the network, and once you have it on the network, how do you do things with the data? How do you capture the data in the right way?

We’ve been looking to reduce that friction and improve device security. The partnership with Intel is about how you get those devices onboarded independently of the underlying architecture of the device itself.

Question: Are you still profitable, albeit less profitable?

Segers: Yes.

Drew Henry of Arm at the company's Arm TechCon event.

Above: Drew Henry of Arm at the company’s Arm TechCon event.

Image Credit: Dean Takahashi

Question: Could you explain some of the assumptions behind the big numbers you had? A trillion units, $11 trillion in value, how do you connect that back to what’s shipping today or what growth rates you’re talking about for those numbers to happen?

Marcelo Claure: We get those numbers through a combination of third parties. We get numbers from Arm. We get numbers from all of our companies, and then we make some predictions around the amount of connected devices that there will be. It’s a gathering of many different sources of data. We run the data through Arm’s management to make sure we’re looking at the same page.

When you look at what’s happening in terms of everything being connected on a global basis, and getting a trillion devices by 2025–everything we carry or that we have is going to be connected at all levels. Therefore we feel quite comfortable that those are the numbers. Then we work backwards with Arm to make sure they’re ready as far as increasing shipments. We’ve been seeing increasing shipments over all of our business, so we feel comfortable that those are solid numbers.

When you look at $11 trillion in economic value, we’ve done a lot of research over the course of the last year in order to put together all the different use cases that will exist. When you’re able to combine high-speed networks with low latency and great coverage–just combine that with what we’re seeing today in the new chips that are already being manufactured.

The compound growth rate of Arm chips has been growing at about 16 percent the last few years. Extrapolate that forward and you get to big numbers quite quickly. A lot of the growth is coming in these very small, very low power processors which are going into microcontrollers inside IOT things. IOT is obviously a spectrum of computing, but in those end points, where a lot of that connectivity is going to happen, we’re seeing explosive growth.

Question: Was the notion of a billion sensors also part of the same 2025 prediction? I think that was Drew’s number?

Segers: Yes, that will be part of the connected things.

Question: With the pullout of GlobalFoundries and chasing Moore’s Law recently, are you at all worried about bleeding-edge capacity? Will there be enough foundry capacity for the growth you’re talking about?

Segers: Not something I spend a lot of time worrying about, I have to say. GlobalFoundries pulling out of that space, I think that’s a rational decision that they’ve made for their business, to go focus on things where they can lead and show some value and differentiation into the market. The technologies they’re focusing on are going to be important for some of the end markets we care about.

TSMC is doing a great job on the leading edge right now. If the market is sufficiently large I’m confident they’ll invest to provide the capacity to the industry. If they can’t do that alone, I’m sure that others will come in to add capacity.

Simon Segers, CEO of Arm, at Arm TechCon 2018.

Above: Simon Segers, CEO of Arm, at Arm TechCon 2018.

Question: Can you share your opinion on what some have seen as a slowdown in the semiconductor industry?

Segers: The semi industry has had a couple of years of very strong growth. I’ve seen reports saying it’s going to slow down. There’s always a variance in the outlook for the semi industry. If I look back, the accuracy with which those forecasts have historically been made hasn’t been phenomenal. But to borrow some of Masa’s thinking, there may be ups and downs along the way, but you have to look through that. We have the opportunity to look through specific cycles and make sure we’re investing in technologies that will be critical for the long run, regardless of whether there’s some bumps along the way.

Again, I don’t spend a lot of time worrying about what the semi outlook is for the next 12 months. We care about that because we care about the finances of our partners, who we are working with to deliver this technology, but we’re taking a long-term view of where the market is going to go. I think those trends are undeniable.

Question: Are you worried at all about RISC V, the open-source chip design that seems to be gaining momentum, including at some of your licensees? What’s your competitive response to that?

Segers: I look at that and I can argue that there have been open-source processors in the past. There have been free processors in the past. There have been configurable processors in the past. None of them have anything like the industry traction we do.

At the same time, the Innovator’s Dilemma is one of the best business books that’s been written, especially in this space. We have to take the threat of somebody disrupting us from below very seriously. So, what are we doing? We’re focused on making sure that our technology investment means we have the best products, that our products are easy to access, easy to design in, easy for developers to build products around.

You can see the fruits of that at this event. We have 5,000 people registered to come to this event. That’s a lot of people, and many presentations being given by people using Arm technology. It’s an enormous amount of traction behind what we’re doing. We’ll make sure we keep investing to keep that community engrossed in what Arm is going. That’s helped by the investments we’re making in our overall platform.

Question: Does it change the licensing dynamic, your licensing discussions?

Segers: Sometimes? But at the end of the day, if you want to build a high-performance microprocessor, that takes real engineering effort. If you want to build a processor architecture and invest in an ecosystem, that takes effort. It takes resources. It takes cost. I’ve no issue with RISC V keeping us on our toes, either from a technology or a business model point of view. Our goal is to stay out front.