Processing Powers

For the past six years Mark Lippett has navigated the deep tech company XMOS towards new opportunities for its specialised chips for audio, voice AI, imaging, robotics and more. Christian Annesley caught up with him to hear about the journey – and what’s coming next

The technology revolution we are all living through has silicon chips at its core. From the smartphones in our pockets to the vast datacentres powering the internet; electric scooters to hypersonic aircraft, pacemakers to supercomputers – inside every one of them, unseen and unsung, are the miniature ‘brains’ that make it all possible: semiconductors.

So it’s a huge industry – and one that is still growing very fast year-on-year thanks, in part, to the seemingly endless proliferation of ‘smart’ devices that now run on computing intelligence. The latest figures show that global semiconductor sales climbed above $500bn in 2021, having been worth $30bn when it was a nascent sector in the mid-1980s. This chip market in turn powers a tech industry that is nowadays worth more than an astonishing $5trn annually to the world economy.

Giant numbers, then, wherever you look in the tech space. But how do companies tap into that market? The impression to consumers might be that a handful of big players dominate in semiconductors – everyone knows about Intel, don’t they?

In fact, the ever-growing number of applications for processing power means this is a market with many dimensions to it. Plus, it is one defined by constant breakthroughs and improvements – even if Moore’s law, which anticipates a doubling of the number of components on a chip every two years, no longer quite holds true these days.

If you build something people won’t just buy it. You have to take your proposition to individuals in a form where they will immediately see the benefit.
Mark Lippett, XMOS

No matter – XMOS stands as proof that the landscape for chip-making remains one of opportunity. So what’s the company’s story?

Finding the fit
XMOS was spun out of the University of Bristol in 2005, with the involvement from the beginning of the well-known computer scientist David May (he was also involved with Picochip in Bath).

The firm received early seed capital from the university’s enterprise fund, plus other seed funds like Wyvern. From the off, it was going after an opportunity for a new kind of specialist chip to solve some fresh problems coming down the line.

Fast forward to 2022 and for the past six years it has been chief executive Mark Lippett at the helm. The company is now around 85-strong and turning over £15m, give or take, with revenues being reinvested.

Lippett is quick to explain, however, that going after opportunities in processors for the Internet of Things and artificial intelligence (AI) is always a challenge that takes time and money to unlock.

“From the beginning we have been placing long-term bets, it is fair to say,” says Lippett. “That’s not to say that we haven’t been targeted – we have, of course – but to re-spin a chip costs at least $10m and takes about 18 months at a minimum. Only at that point do you get the first customer feedback on how well your product is working at solving or improving on something. It’s a big undertaking, but we understand what’s involved. And I can say this much about the process: the idea of a field-of-dreams strategy is a fallacy.

“If you build something people won’t just buy it. You have to take your proposition to individuals in a form where they will immediately see the benefit. So your starting point is to pick on some markets and to solve some problems that are inherent in those markets.”

Audio effort
With XMOS, the start of its journey was in audio, and specifically its capability to solve an audio issue that applied to Apple’s Macbooks.

“We had a general-purpose technology but with the potential built into it to solve this problem – and in a flexible way, which turned out to be our biggest USP,” says Lippett.

If that chance in Apple audio was a way in for XMOS – its first real-world application – it was an opportunity onto which it could quickly start to layer further opportunities.

“Because it was an audio application, it helped us move into the voice AI market that has roared ahead with the likes of Alexa and Google Assistant becoming a part of people’s lives.”

A platform for opportunity What XMOS offers in these market spaces, then, is fast and versatile chips that are well-suited to supporting the needs of these new enabling technologies.

Sometimes it has to position itself and think hard about where it fits with an opportunity, but other times it has found the platforms it delivers then get picked up and used in ways it hasn’t predicted (and fortunately that it doesn’t need to predict).

“That’s a boon for us, of course,” says Lippett. “Some of our marketing and investments have to be focused on solving particular problems, but often we just need to identify a space into which our technology has the potential to expand and flex. Our xcore.ai platform, for example, delivers artificial intelligence for the Internet of Things – AioT for short – and it has applications in myriad contexts, from smart homes to health monitoring devices to automotive to smart cities and smart manufacturing.”

Ready for take-off
If it sounds like XMOS is well-placed now in terms of growing some of its opportunities, that’s a fair impression. It has taken about £60m in equity funding since launch, but equity raises haven’t been on its agenda for a few years now. It can reinvest from its profits or look to take on debt when it needs to spend.

“We are investing behind our revenue,” says Lippett. “We don’t need to be highly profitable at this stage but to maintain growth. The aim for the next few years is to maintain 50 per cent annual growth if we can, although there are also some market headwinds and supply chain issues to contend with which might slow us down slightly.”

XMOS now has about 200 customers at any one time, which gives a sense of the broad base of its opportunities. It has even developed a solution for an e-bike and for some train-terminal fast-food kiosks in Shanghai, for example.

“Adaptability is the key to our success,” says Lippett. “We address problems with flexible technology, which is good for us and good for our would-be customers. Our workforce is full of innovators, and that’s our strength.”

Adaptability is the key to our success. We address problems with flexible technology, which is good for us and good for our would-be customers.
Mark Lippett, XMOS

The company is what is known as fabless, which is another way of saying it designs and sells its semiconductors but outsources production to a specialist semiconductor foundry in Taiwan. As a modus operandi, it has that in common with many other semiconductor businesses (see The origins of fabless manufacture, prior page).

To keep the business on track, and also as a reflection of its broad investor base, XMOS today has quite a large board of ten collectively in charge of steering the business.

“There are a lot of us, it’s true,” says Lippett. “But it works well. Everyone who is involved brings something different to the table – and we are, after all, a complex business operating in complex markets.”

The Bristol advantage
And how does Bristol work as a base? “Very well indeed,” says Lippett. “You generally get strong technology clusters around universities, and that’s the case with Bristol of course. The University of Bristol has a strong computer science department, and there are high-quality graduates coming out of the University of Bath and the University of the West of England, too. One of the things that Bristol has going for it in addition is that it’s an attractive, liveable city. That means many who study in Bristol are only too happy to stay on and call it home for the long term.

“I would say that the winds of change are also working in the city’s favour. What it offers is consistent with what many young people are looking for today in a place to live, with great proximity to the countryside and coast and generally great travel connections, but without a big-city vibe.

“It has become expensive, it’s true, but it is not as pricey as London and much of the South East while offering plenty of advantages. And connectivity to London is still strong, for when it is needed.”