Imagination Targeting Heterogeneous Computing

Article By : Nitin Dahad

EE Times visited Imagination Technologies U.K. headquarters for an exclusive interview with CEO Simon Beresford-Wylie.

Last October, Imagination Technologies hired a new CEO, Simon Beresford-Wylie. He is Imagination’s sixth CEO in 6 years, and just like his predecessors he is committed to building on the company’s heritage in GPUs. The biggest question is if Beresford-Wylie can come up with a different way to do the same thing, only this time leading to sustained success.

Now almost nine months into the job, and with the lockdown in the U.K. easing, Beresford-Wylie was available for an exclusive interview at the firm’s revamped headquarters in Kings Langley, a rural setting about 20 miles north of London.

Simon Beresford-Wylie comes from a telecoms and broadcast industry background. He is fully aware his office seems to be equipped with a revolving door. However, he quipped as we got into the introductions, “I’m a resilient person though; I will be here.”

Every new CEO always seems to have turnaround plans for a company, and Beresford-Wylie is no exception. When I talked to a previous CEO, Ron Black, in January 2019, he had the same sentiment (see: Imagination’s New CEO Plots Turnaround Strategy).

So, is it different this time? Beresford-Wylie is confident as the company has costs under control, turned a profit in 2020, has a clear plan on how to utilize the company’s GPU and mobile heritage, has investor backing for a long-term plan which includes a 10-year outlook for its RISC-V strategy; and he has an executive team aligned with his revamped strategy.

The latter is apparent as I walked around the building and as I met each executive in turn along the way, all were either amazingly on-message or really believed in the new strategy and were behind it wholeheartedly. I got the sense it was the latter, since they could see the initial positive results. For example, Steve Evans, the chief revenue officer, was bullish about sales growth and especially areas like electric vehicles (EVs) in China; and Tim Whitfield, the chief of engineering and recent recruit from Arm, also seemed geared up to apply his experience from over 20 years at Arm.

Simon Beresford-Wylie Imagination Technologies
Simon Beresford-Wylie in the lounge area of Imagination’s revamped HQ building. (Source: Nitin Dahad)

Beresford-Wylie was also keen to demonstrate how he is attempting to bring about changes to the office culture at Imagination. As an example, offices are now all open plan, with no offices or allocated desk for the CEO; and he’s done away with special designated car parking spaces for senior management. And if anyone wants a quiet workspace, there are enclosed cubicles available. And to facilitate hybrid working across geographies, they’ve invested in video conferencing facilities.

It all felt very much like a WeWork building, but fairly empty for now as people only have to come into the office optionally so not many people are in at any one time.

So, to the interview.

The 100-day plan

First, we asked him to assess his first nine months.

Simon Beresford-Wylie: We’ve turned a corner in terms of profitability which is great. We exited 2019 in profit, we’ve completed a strategy review and the board signed off on that in March (2021). We’re executing against that and we’re off to a great start. All in all, I’m feeling very happy with acquiring new customers, getting lots of design wins. The pipeline is incredibly strong and growing across multiple geographies in multiple segments.

EE TimesYou mentioned that you had a 100-day plan, what did that entail?

Simon Beresford-Wylie: The 100-day plan was effectively what I came into the business with on the 1st of October 2020, and I think you need to remember that was a quite turbulent period, the first half of last year, so there are a number of elements to it.

The first part of that was just to fill out the executive management team, as we had a number of vacant positions. We’ve done that with Tim Whitfield joining us for example to run engineering, Tim Mamtora joining us to run our labs environment, Mark Logan as the CFO, and Nick Merry into HR. So, we are really just getting the team bedded in there.

The second thing we needed to do as a team was to stabilize the near-term financials of Imagination. It’s probably no secret we’ve not made profit or generated cash for a number of years. And in all businesses, you need to be able to feed yourself, and pay salaries. Certainly, the beginnings of this process had started before I joined. We have since been laser-focused on getting to profit and cash, and happily as we exited 2020, we eked out a small profit and we generated cash.

The third aspect of the 100-day plan was to effectively go through a proper “bottoms up” strategy review. Who are we? Where do we want to point the business? What segments should we be pointed on? What geography should we be serving? What IP should we have in the portfolio? That was a really deep exhaustive exercise.

We kicked that off in November and the board signed it off in March.

Then the fourth thing we’ve done around this was working with the executive management team to revisit our mission — like who are we, and where are we going — and revisit our vision. Then bind that all together with the strategy and bring that to the employees and at the same time refresh values.

So really almost a reboot of the company, and we are through it now.

I think that the 100-day plan worked well, it aligned us all, and we’re on the other side. We’re in execution mode.

Imagination Technologies management team
Most of the executive management team back in the office at HQ in Kings Langley, UK. From left to right: Chris Porthouse, Simon Beresford-Wylie, Tim Whitfield, David Harold, Steve Evans, Mark Logan. (Source: Nitin Dahad)

Targeting a market shift toward heterogeneous computing

EE TimesThose are great words, but what are the headlines of that strategy review in terms of who you are, what do you want, your portfolio and your target market?

Simon Beresford-Wylie: It’s a great question. Imagination as you know is a 35-year-old company, with 25 years in GPU IP, so the heritage of this company is very much in GPU IP, particularly most recently in mobility; and then the secondary market of infotainment systems.

That heritage is great but actually there is a shift in the market now towards heterogeneous compute. Customers, for example in automotive, are saying, “Actually we don’t want to buy from you just the GPU IP for the infotainment system”. With the arrival of automated driver assistance systems (ADAS) and EVs, there is a pull for a greater solution approach. The wiring looms are coming out of cars and are being replaced by Ethernet. There’s a lot more data being pumped around those cars and being analyzed. Effectively cars are almost becoming computer platforms on wheels, so customers are saying, “We need to work with partners who can actually provide Ethernet packet processing AI solutions, CPU and GPU, or GPU doing more than just GPU.”

That’s really a solution-orientation, and with heterogeneous compute. So, part of our strategy has been looking at how we serve segments rather than just being a point IP company.

EE TimesCould you then explain the segments you are targeting?

Simon Beresford-Wylie: The first one I’ve already touched on is automotive, and we are serving this segment in North America and China, and also Japan and Taiwan.

A second segment that has really come up and is fast growing for us is the data center, cloud and desktop (DCD). Our heritage in mobility is actually serving us well here, because we’re very good in terms of power, performance and area; that means very limited power consumption, high performance and very small area. And guess what? That’s very resonant in data center applications.

The third segment is from the heritage in mobility, mobile devices. We grew up in the Apple iOS, and now we see significant pull for our offerings in the Android world, and I think that’s a reflection of the investment that we made in 2019 under Canyon Bridge in refreshing the GPU product line.

That’s really got attention in the market, so we’re taking market share and we see all sorts of opportunities in high-end devices as well as low to medium in the Android world.

More investment in RISC-V – but it’s a 10-year play

The fourth area is RISC-V, which is not so much a market segment. We have RISC-V already today within the GPU on the compute side.

But more broadly I see this as a 10-year slow motion market “discontinuity” if you like. Like a poor alternative to a very large IP company that’s based here in the UK on the CPU side of things.

So we have RISC-V already, but we are going to invest more in the coming years and develop a high- end offering both as a standalone IP offering for those that wanted to buy CPU RISC-V, and also to embed as mentioned earlier in the solutions that we sell into other segments.

EE TimesIs it the case that RISC-V hasn’t yet taken off yet because not enough larger established players are sort of taking that on board?

Simon Beresford-Wylie: I think there’s an element of that. These things do take a heck of a long time to get traction as ecosystems develop around them, and companies that are sustainable that customers and partners believe are going to be there, start to move into this.

There are a lot of startups that are active in RISC-V but don’t have a track record of delivering quality IP at a point in time.

The great thing about Imagination is this company has been around for a long time. We do have a track record of delivering quality IP. You know when we say we’re going to deliver it. So, we’ve got the engineering capability and prospective customers recognize and acknowledge that, and I think that that’s something that we will use to our advantage as we move into that high-end RISC-V segment.

Regional split

EE TimesYou said you turned a profit in 2020. Have you identified where your biggest growth is right now? For example, is it EVs in China?

Simon Beresford-Wylie: So North America is a good and important market for us, and China is definitely a high growth market, no doubt about that, both especially in automotive. What’s happening in China is extraordinary, I guess; but equally Japan is a strong market for us; Taiwan is a strong market for us; and actually Europe, we sell here as well. China is driving quite a lot of growth, particularly in the data center and desktop (DCD) market segment.

In terms of split between regions, I’d say Europe has come off a little bit, the US is very significant, and we’re weighted more towards Asia now.

EE TimesWhat about the hyperscalars?

Simon Beresford-Wylie: So, we’ve we benchmarked very well against Nvidia. I think that’s a reflection of the PPA (power, performance, area) heritage that we have and that ports nicely into the DCD market for us. And with the investment we made in refreshing the GPU line in 2019, we’re getting real benefit from now. It’s really got attention.

Geopolitics and China

EE Times: Let’s turn to the corporate side and geopolitics. There was quite a lot of perceived turmoil with your investors and management. Is that now settled?

Simon Beresford-Wylie: Yes, it is. Look, it was a very difficult time. You know the first six months of last year was quite traumatic for the company. I think it showed a resilience in the engineering workforce as these things were going on around them in terms of the executive and the commentary around the potential appointment of new board members from China Reform.

But that’s a year ago now, and we have moved beyond that. I’ve been in post now for coming up on nine months, and this feels just like any other company to me. It’s got a normal board; we’ve appointed two new non-executive directors, Sir Peter Bonfield and Didier Lamouche, both of whom are veterans from the semiconductor sector.

My board meetings are just like the board meetings I had at Arqiva and just like the board meetings from Nokia, so it’s no different. Canyon Bridge is like any other private equity company, and as I see it at least, China Reform are just like normal limited partners investing in a private equity business.

EE TimesWhat are your thoughts on the geopolitical dimension to doing business with China?

Simon Beresford-Wylie: the reality is that the world has shifted, and I think it’s shifted forever in practice, there’s a geopolitical dimension to the China-U.S. relationship. Now they’re competitors. They look at each other through a different lens.

We, like any other company that wants to sell into both markets, need to be very sensitive to that in terms of what we develop, where we develop it, and the customers to whom we sell. We are like all of our competitors and all of the tech companies in North America, they’re still selling into China.

There are certain customers that are on entity lists and so on that are problematic and we have certain IP that we are very sensitive to in terms of what we can and what we can’t sell and there are export control mechanisms in place here in the UK which we follow religiously.

EE TimesHas that entity list affected you in any way in terms of sales?

Simon Beresford-Wylie: Well, we are growing at a rate of knots at the moment, despite the entity list.

From telecoms and broadcast to semiconductors?

EE TimesWith your background being very much the telecoms networks sector, semiconductors and intellectual property is very different. What do you think you bring to Imagination and the industry?

Simon Beresford-Wylie: Well, I challenge that a little bit. So, the business model is a bit different, but actually what I have found here is it’s R&D engineering, it’s writing code, it’s regression testing it’s verification, it’s program management. That’s exactly what we were doing at Nokia.

It’s different but the same. And it’s different but the same from what we were doing at Elster with smart metering networks, you have engineers writing code and testing the code and getting the solutions out there; and working on roadmaps and all that sort of stuff.

The business model is different [here]. We don’t manufacture things so that that that’s a bit different.

What can I bring to this company? I think I’m a seasoned executive with a strong track record. I’m a values-based leader genuinely, and I think that counts for a lot in terms of building a modern work culture and walking the talk there. But I’m super enthusiastic about getting back into a sector in growth and accompanying growth. I have to go back to the 90s to think, “Oh my goodness, this is what growth is like”.

So, it’s essentially about acting as a catalyst and creating an energy and an excitement that we can do this and that we can win. We need competitive IP, we need to have the right strategy in terms of the markets that we’re focusing on, and we need to be executing with alacrity. And I think I’ve got the capability to help that.

Why did you want to be CEO at Imagination?

EE TimesWhat was the trigger in your mind when you heard that there was a job going for a CEO at Imagination that made you think: “I want to go and try”?

Simon Beresford-Wylie: First of all, it is global, which I love, but it’s engineering. I’m not an engineer. I regret the fact I’m not an engineer. I want to be an engineer. I love working with engineers because engineers rule the world. They create things. They make things. Everything in this room – apart from you and me – has been designed and built by an engineer.

It’s incredible. What we do here at Imagination is so complex, rocket science complexity. And that excites me and I’m tech curious and it’s an iconic company.

I think it was an unpolished jewel. It’s a jewel, but an unpolished jewel and we are polishing that jewel to put a lustre on it to make the employees across the business believe that we can be winners in this move towards heterogeneous compute and the GPU, the EPP (Ethernet packet processor), the AI, the NNA (neural network accelerator) and the CPU IP that we have will put us in a great position to be a winner.

And I just wanted to be part of that.

EE Times: So the last question then. What’s your vision for the next 12 to 24 months for Imagination?

Simon Beresford-Wylie: It is to execute on the vision that we have, and in the process, become the undisputed leader in this new move towards heterogeneous compute. We’ve got a strategy. The investors have signed up to it. The board of directors met in March, and they said go for it, including RISC-V CPU.

EE Times: Simon, thank you.

This article was originally published on EE Times.

Nitin Dahad is a correspondent for EE Times, EE Times Europe and also Editor-in-Chief of embedded.com. With 35 years in the electronics industry, he’s had many different roles: from engineer to journalist, and from entrepreneur to startup mentor and government advisor. He was part of the startup team that launched 32-bit microprocessor company ARC International in the US in the late 1990s and took it public, and co-founder of The Chilli, which influenced much of the tech startup scene in the early 2000s. He’s also worked with many of the big names – including National Semiconductor, GEC Plessey Semiconductors, Dialog Semiconductor and Marconi Instruments.

Subscribe to Newsletter

Leave a comment