In September 2015, hardware veterans Nigel Toon and Simon Knowles were doing the rounds of venture capital offices in Silicon Valley and London, touting their latest startup. The pair had a dazzling track record – among other achievements they’d sold their previous semiconductor company Icera to NVIDIA for $435 million (£346 million) four years earlier. And their vision for Graphcore – a new Bristol-based venture – was bold: they were building a new generation of microchips known as intelligence processing units (IPUs), designed for the rapidly approaching artificial intelligence age.
Yet early reactions to their pitch for series A financing were distinctly muted. “In many cases we were laughed out of court,” recalls Toon, Graphcore’s CEO.
Typically, Toon says, they’d find a partner in a VC firm who was excited by what they were doing. “But then they’d go to their partner meeting, where the first question would be: ‘What’s AI?’ It’s stunning to think that was a conversation that was happening [as recently as] 2015.” From there, it was an uphill struggle. “Even if they got the fact that AI might be interesting, they’d then say: ‘Your business model is to build a chip for this AI thing? Well, nobody’s made money from chip investments in the last 10 years.’”
Toon, who is 55 and has the mellifluous voice of an old-school BBC continuity announcer, says that chip development, in the eyes of most investors at the time, was considered highly capital intensive, with returns failing to justify the upfront financing required. “It’s not more capital intensive than software,” says Knowles, Graphcore’s co-founder and CTO. “But software has this joyful property that you can try it out in small scale first, whereas with a chip you’re all in. If it doesn’t work, you’ve spent all your money.”
That was 2015. Fast forward to today and, of course, AI hardware is a white-hot category for investors, with VC funding for US AI companies jumping by 72 per cent in 2018 to a record $9.3 billion (£7.4 billion), a fifth straight year of growth, according to a report by CB Insights and PwC.
What changed over those three years? Toon points to two things. First, in 2016 traditional chip giant Intel acquired an AI software and hardware startup called Nervana for $350 million (£280 million), raising eyebrows all over the Valley. Second, Google announced it was going to build its own chips – evidence, Toon says, that existing chips weren’t up to the task.
Knowles describes the impact of Google’s decision as “seismic”. The fact that Google thought AI was going to be a sufficiently big deal to justify the pain and expense of building its own chip team helped make the Graphcore founders’ case for them. He and Toon had been arguing that it was worth digging deep financially to develop new processor hardware because existing graphics processing units (GPUs) – used, for example, in mobile phones, games consoles and personal computers – weren’t designed for AI workloads such as machine learning and deep learning.
By then, their startup was already ahead of the pack in developing a new processor architecture. Soon top-tier investors – including Atomico, one of Europe’s best-known VCs – were beating a path to their door. Atomico, which went on to lead Graphcore’s $30 million (£24 million) Series B round in July 2017, was followed six months later by one of the Valley’s biggest guns, Sequoia Capital. At the time, Graphcore, having recently closed its Series B, didn’t need investment – but the west coast investor wasn't taking "No thanks" for an answer. “They came to see us here in Bristol and said, ‘No, you don’t understand, we want to invest in your business,’” laughs Toon. “So we work out terms and they invest $50m into the company. And that’s one of the very few investments they’ve made in the UK, because they’ve got so much opportunity on their doorstep.”
Sequoia partner Matt Miller, who now sits on Graphcore’s board, admits he was somewhat bemused to find himself chasing down a company based in Bristol. “We knew there was an opportunity for a new architecture that would be designed from the ground up that could massively accelerate our entry into this AI age, and we were trying to landscape all of these companies in China, the US and Europe,” he says. “But our references were all pointing to this one company in Bristol, whom we hadn’t met yet.”
A roar of laughter distorts the line from the Valley. “Lemme tell you, if you’d asked me a month prior if I’d ever [sit on] a board in Bristol I’d have said ‘No way!’ It’s not your typical destination on your tour of Europe. But to be honest, it’s been surprising for us in the Bay Area because the quality of talent in the UK, and particularly in Bristol in the semiconductor space, is very strong. The team they’ve been able to build there is on a par with the best in the world.”
Following a $200 million (£160 million) Series D round in December 2018, Graphcore was most recently valued at $1.7 billion (£1.36 billion), with investors, innovators and large corporates now seemingly convinced it will be the company to power the AI era in much the same way as Cambridge-born chip giant ARM dominated mobile devices, shipping over 130 billion chips and reaching 70 per cent of the global population. The opportunity at stake is nothing less than the future of AI, with applications ranging from medical advances to autonomous vehicles, space exploration and just about everything in between.
In fact, Bristol has a strong history as a hub for hardware engineering, which can be traced back to 1978 and £50m of seed investment (another £150m would later follow) made by the UK government in Inmos, a microprocessor startup with fabrication facilities in Newport, South Wales. “We often forget the importance of government investment,” says Hermann Hauser, the Austrian-born entrepreneur and investor best known for spinning out ARM from Acorn Computers – and Graphcore’s first backer. “It was the £200 million that the Callaghan and, later, Thatcher governments originally spent on Inmos that created the infrastructure and ecosystem around Bristol that really understood semiconductors. It created brilliant people like [leading computer scientist] David May, Simon and Nigel, who would not have been there had it not been for the government initiative at the time.”
Knowles first came to Bristol in 1989 to work for Inmos. “Historically, Bristol has been the centre of chip design [in the UK], and in many ways ARM and CSR [formerly Cambridge Silicon Radio] were anomalies,” he says. “I mean, they’re very successful, large anomalies, and now everyone associates Cambridge with chips. But in terms of numbers of chip startups, and how many years back it goes, Bristol is the dominant place in the UK.”
Graphcore emerged from a tangled family tree of semiconductor companies. Toon and Knowles were introduced to each other by Stan Boland, former CEO of Acorn Group and now CEO of autonomous vehicle startup FiveAI, who had worked with Knowles at chip company Element 14. When this was acquired by Broadcom for $640 million (£512 million) in 2000, the pair went on to found Icera in 2002 with Toon, who was previously with electrical equipment manufacturer Altera Europe. When Icera was sold to NVIDIA, it meant that Knowles had already exited two chip design startups at a total value of over $1bn. But he and Toon were far from finished. What motivated them to start all over again with Graphcore?
Sitting across the table from one another in a fifth floor meeting room at their Bristol HQ, the founders exchange a fleeting glance. After a while in their company, it's clear that this long-established business double act has acquired some of the hallmarks of a marriage: they have an easy rapport, finish each other’s sentences, and occasionally talk over and correct each other.
“Simon maybe has a different view,” says Toon, “but my sense of it is that this is what we get up in the morning for. The fact that the opportunity in front of us is so enormous, I feel like I’ve been waiting my whole life for this.” He adds that it comes down to purpose: “You might get some satisfaction from connecting people together in a social network, for example, or delivering food to them through an internet app. What we’re doing is potentially changing the future of compute – we’re potentially allowing lots of people to create major breakthroughs; maybe someone will come up with a cure for cancer using the tech we’re creating.”
“We’re building the motors of AI, really,” says Knowles. “And what people will build out of those motors is far greater than our motors. We want to be the Rolls-Royce jet engines of AI machinery.”
In essence, the problem Graphcore is solving is that previous generations of microprocessors – central processing and graphics processing units – weren’t designed for machine intelligence, which requires a new way of processing data.
Knowles holds up a Graphcore chip. The size of a small cracker with a dark grey, metallic centre, it contains 23.6 billion transistor devices all connected by several miles of wiring. As transistors were progressively shrunk over the decades so that more of them could fit on to each chip, the chips themselves grew correspondingly hotter as energy demands increased. “We’re almost at the end of that gravy train now,” says Knowles. “The objective of chip design always used to be to go as fast as possible; now it’s to make the most use of the energy available.”
“To make them as efficient as possible,” clarifies Toon.
“Exactly,” says Knowles. “And actually you design things in a completely different way if you’re most interested in energy and less interested in speed per se. So why do we want more computing performance? We’ve just started to work out how to mechanize intelligence. And what do we mean by intelligence? A machine that can learn by its experience, or by being given examples, or by itself, discovering things. In no sense, historically, has a computer solved a problem – it was always the person who wrote the program. AI flips that on its head.”
Suddenly, there’s a surge in demand for more processing power due to the AI workload, at precisely the moment when traditional silicon shrinking won’t offer it. “Explaining to a computer how to learn is quite different to explaining to it how to do traditional supercomputer maths for example,” says Knowles. “So we’ve set about trying to solve those two problems – intelligence is a different workload, and focusing on efficiency and not speed – with our IPU.”
Whereas other AI hardware companies have focused on neural networks – a type of knowledge model for capturing the sort of intelligence in the human cortex, which is essentially designed to recognise numerical patterns – Graphcore has built an architecture that is more flexible. It can run current machine-learning approaches, as well as new and emerging approaches that simply don’t work efficiently on today’s hardware. “What most of the [rival] startups are doing is building a machine to do fast neural networks, and that’s what you do if your ambition for your company is to sell it for a couple of hundred million in a year or two,” says Knowles. “What we’ve tried to do – because our ambition for the company is to be permanent, and broad enough to encompass engines for AI as opposed to just chips for perception – is build a much more general purpose machine. Nigel and I were very clear about our ambition for this company: we’ve grown and sold companies before, but this one is our magnum opus.”
Toon chips in: “This is a once-in-a-generation opportunity. If we get this right, the IPU will define the future of machine intelligence, powering world-changing innovations for decades to come.”
VCs are rarely sparing in their use of hyperbole. But when a big-hitting Valley investor like Sequoia’s Miller says “We think [Graphcore] can be a company with a market cap in the tens of billions of dollars”, and flies halfway around the world to make an investment in a startup that wasn’t raising money in the first place – with the likes of BMW, Microsoft, Bosch, Dell and Samsung also queuing up to invest – there tends to be a pretty good reason.
The answer lies in the almost limitless fields Graphcore’s IPU can be applied to – anywhere, in fact, that machine intelligence can enhance human activity. “There are still some things humans are going to be better at, typically creative things,” says Atomico partner Siraj Khaliq, a computer scientist and former entrepreneur. “But when it comes to looking at patterns and making predictions – for example looking at a radiology scan and deciding if there’s cancer there or not; looking at someone’s viewing habits and deciding what they should watch next; even looking at the attributes of a person, what they do and what they like, and recommending who they should marry via dating apps – all of these things machines will now do because they’re just better at it. So I don’t think I’d be doing it justice by saying ‘Here are one or two things that Graphcore’s IPU will be used for’, because it is really pretty much everything.”
Back in Bristol, Knowles cites medicine and law as two areas on the brink of AI-driven transformation. “What is the definition of a good doctor or a good lawyer?” he says. “It’s someone with a lot of wisdom acquired by experience, someone who’s seen a lot of cases, read and digested a lot of research material and comes up with good answers. They can’t always be correct, but given the knowledge that exists they come up with the best reasonable answer based on their experience.”
The most exciting opportunity for machine intelligence is being able to do that with all of human knowledge, he says. “Take a medical oracle which can read all of the medical research that’s ever been published and can resolve and identify discrepancies. It can read all of the patient records that have ever been recorded. And it can come up with the best answer based on all of human knowledge. It’s not perfect, because all of human knowledge isn’t all knowledge, but it’s the best we can possibly do and the opportunity there for totally solving a whole load of human conditions must be enormous.”
Graphcore’s founders say that more than 100 developers or end users are currently working with their IPUs, although they decline to identify any of them. “I’m not sure we’re allowed to say [who they are],” says Toon. Is it a fair assumption that big brand investors and individuals such as Demis Hassabis, a co-founder of DeepMind who invested personally in Graphcore, are testing the technology? He bats away the question. “They are strategic investors in our company. They’ve made a decision that our tech could be strategically important to their businesses, so you might surmise that something’s going on, but we couldn’t possibly comment.”
However, days after WIRED’s conversation with Toon and Knowles, an approach to BMW i Ventures (the car giant’s venture arm focused on automotive tech) suggests a possible application. While BMW wouldn't confirm whether it was working with Graphcore (this information is commercially sensitive), it’s understood from a separate source that BMW is indeed exploring the possibilities of the startup’s IPUs. Tobias Jahn, a principal at BMW i Ventures, says his firm became interested in Graphcore as an investment because of its technology's potential for automotive applications. “For highly and fully automated driving, commonly called levels 4 and 5, efficient AI acceleration is going to be indispensable,” he says.
Graphcore is currently a niche player in a vast global semiconductor market which grew by 13.4 per cent in 2018 to $477bn, according to Gartner. Over the past two decades, the chip industry has undergone a fundamental shift that has seen manufacturing gradually move from the US and Europe to Asia. “That partly reflects the lower cost base for production in Asia and partly where incremental demand is being driven from these days – and clearly China has played a significant part in that,” says Jim Fontanelli, senior analyst at Arete Research.
With that in mind, could Graphcore’s competition ultimately come from China? It’s complicated. In 2018, there were no Chinese companies among the world’s leading 15 semiconductor corporations (which were headed by Samsung and Hynix in South Korea, Intel in the US, and TSMC in Taiwan, which manufactures leading-edge chips including Graphcore’s). Fontanelli doesn’t see China catching up with South Korea, the US and western Europe any time soon. “The ability to design chips is largely independent of the ability to manufacture, and China still has a significant gap to where the leading guys like TSMC, Samsung or Intel are from a manufacturing perspective. Certainly I don’t think they have the ability to realistically catch up in the next five years and possibly not in the next decade. The requirements around leading-edge manufacturing are far more than just having capital available.”
However, when it comes to chip design – particularly for AI – Hermann Hauser reckons the chip giants will not be able to rest on their laurels for long. “[Chip design is] still something that the west seems to be doing better than China. But having said that, China produces more STEM graduates than Europe and America put together. Chinese universities are now overtaking American universities in terms of publication of scientific articles. And China leads the way in the number of patents that it files.”
Toon says the Chinese government went through the AI equivalent of a "Sputnik moment" when DeepMind’s AlphaGo became the first computer program to defeat a professional Go player, in Seoul in 2016. “They’ve been investing a lot of money and the thing China is doing differently to other countries is that they’re making data available to companies they’re trying to support,” he says.
“They see, as we do, that this is a fundamental shift in computing and this is their opportunity to try to become independent using their own technology, rather than being dependent on other people’s. So I would say they are very actively trying to support and build their own technology at the semiconductor level, at the algorithm and application level – I wouldn’t say they are a long way behind, they’re running very quickly.”
Knowles adds that while China “can certainly build chips”, designing “state of the art microprocessors” like Graphcore's is a different matter. Historically, China has not had chip design capabilities – it has not had chip manufacturing capabilities until fairly recently – so it hasn’t got that indigenous expertise. But the Chinese diaspora has been studying and working in the west. “And now that China is becoming a more attractive place to live, I’m sure they are going back to China, bringing their skills with them, and China will learn to do this.”
While there may not be viable Chinese competitors to Graphcore – at least not in the near term – in Europe it's a different story, and AI hardware-focused startups are emerging fast. “It’s a highly competitive space, and there are quite a few startups trying to do this now,” says Siraj Khaliq. “But they have different approaches, and I haven’t seen one with a better approach [than Graphcore].”
However, he concedes that people will eventually copy Graphcore’s approach, which means that Knowles and Toon will only succeed long term by moving faster, continuously innovating and having an array of products in the pipeline.
Hauser, too, accepts that there are “lots of startups trying to do this”, but says Graphcore has two big advantages. First, it was fastest out of the blocks. Second, it “got extremely lucky” in that the members of one of the best chip design units in the world – the Knowles team that went to NVIDIA in the Icera exit – were made redundant at the precise moment Graphcore needed them. “Normally with a startup you’re not given one of the world’s best design teams on a plate,” he says. “They [went on to] produce the world’s largest and most complex chip in one-and-a-half years – and they were right first time.”
It’s fair to say Toon and Knowles have ridden that initial luck. By their own calculations they have gone on to raise a total of $329 million (£263 million) over four rounds as they scale at a ferocious pace. At around 270 employees today, they expect to swell their ranks to up to 500 by the end of 2019. “Last week we added 10 people,” says Toon. “We’re in the process of building up a team in Cambridge, and we’re hiring here in Bristol at a massive rate. We’re also ramping up our team in Oslo who are building a technology of how we connect these IPU processors together, so you can have thousands of processors that all work together.” The startup also has a customer support and business development team in Palo Alto, California, and is building up an equivalent operation in Beijing.
Graphcore plainly has a decent shot at becoming one of European tech’s outsized success stories, perhaps even eclipsing the likes of Spotify (approximate market cap: $26 billion), Yandex ($12 billion), Zalando ($9.5 billion), Delivery Hero ($7.5 billion) and ARM itself, which was acquired by SoftBank for $32 billion in 2016. And yet the pick of the UK-born AI startups – such as DeepMind (acquired by Google), Magic Pony (Twitter), Evi Technologies (Amazon), Vocal IQ (Apple), and SwiftKey (Microsoft) – have generally been snapped up by one of the US goliaths before making it to global scale. Might Graphcore follow suit?
“We’ve certainly done that before,” says Toon, referring to the pair’s earlier exits. “But we think our market is massive – it’s not like this is going to be a small thing inside someone else’s chip; this is a standalone, independent piece of tech that will be sold on a very large scale. So that would suggest this is a standalone company – and all of the investors we’ve had so far are there for the long term.”
“They’re carefully chosen for that,” says Knowles.
Toon smiles. “When Matt Miller, from Sequoia, came for his very first board meeting – and Matt’s a big guy – he looks around the room at all the other investors and says ‘Look, the first one of you to talk about selling this company, I’m going to punch you on the nose.’ He said it as a piece of fun, but that’s what Sequoia does – it builds big companies that go public, and he just wanted to be sure that all the other investors were on the same page.” So is Graphcore’s goal ultimately to IPO? “That’s the path we’re shooting for, absolutely.”
The appetite in Europe now, particularly among the leading VCs, is not to build unicorns, but decacorns, says Toon. “It’s not about having one [tech giant], it’s about having lots. For us, it’s this idea of ‘Will people in future buy more CPUs or IPUs?’. They’ll buy more IPUs. CPUs will still be there. They’ll be doing the inputs and the outputs. They’ll be presenting and collating the data. But the compute will be done on IPUs.
“It’s like going back to the 1970s and the birth of personal computers, microprocessors, and companies like Apple and Intel that got created at that time. There are going to be Apples and Intels that will be created in the AI world. And our goal is to be one of them.”
This article was originally published by WIRED UK