He has been in technology long enough to be sceptical of markets that confuse scale with strength. Morris started working in IT in the 1970s, when computers were slow, memory was scarce, and efficiency was a basic requirement, not a competitive advantage. That early experience still shapes how he thinks about technology now.
While much of the AI market has focused on bigger models, more compute and larger infrastructure bets, Morris believes the more important question is simpler: who can turn that into something genuinely useful, commercially viable and efficient?
That belief sits behind Corpora, the AI research engine he leads as CEO and backs as an investor. It also informs his view that the UK may have a more credible role in AI than many assume. Britain is unlikely to outspend the US or overpower it on infrastructure, but Morris thinks that misses the point. He sees the opportunity in building smarter systems that make better use of what they have.
“My first computer was an ICL 1901, and it had 16 kilobytes of memory,” he says. “That is nothing by today’s standards, but it forced me to be efficient.”
It is a line that helps explain not only how Morris learned to build, but how he still judges the market. Efficiency, for him, is not a technical side note. It is one of the clearest signs of whether a business actually understands its product, its costs and its long-term potential.
“Systems were far less sophisticated,” he says. “You had to think much more carefully about what you were doing.”
That mindset feels increasingly relevant in AI. The sector has attracted extraordinary attention and investment, but it is still wrestling with a basic commercial problem. These systems are expensive to run, uneven in their outputs, and often built on workflows that create more cost than they remove. Morris does not dismiss the scale of the opportunity. He thinks that the current market still equates technical power with business quality.
What he thinks most AI tools still get wrong
Corpora was built around that view. Launched in late 2024, it positions itself as an AI research engine rather than a general search assistant. Morris is careful about that distinction because he thinks the category is discussed too loosely.
“If you’re looking for the cheapest place to buy a TV, that’s not research,” he says. “Research helps you tackle difficult challenges that you need to understand in a reasonable amount of detail.”
That is the area he cares about: more involved questions, larger sets of information, and outputs that go beyond quick summaries. In that category, he argues, how the system is structured matters as much as the answer it produces.
Many AI research tools still rely on a cumbersome process. They search the live web, identify relevant sources, process that information, summarise it and then assemble a response. That can work, but it is expensive in both time and compute.
Corpora’s answer is to organise public information inside a database designed to be interrogated more quickly by AI.
“Our database structure allows us to interrogate information at much faster speeds,” he says. “Other AI research tools have to search the web, go through the relevant pages and summarise their findings. Because AI is not exactly the most efficient technology in the world, the cost and time of doing that is quite extensive. But Corpora’s database means it doesn’t have to go through that process.”
It is a simple point, but an important one. Right now, many AI products are still judged by how impressive they look in a demo or how broad their claimed capabilities appear. Morris expects that to change. Over time, he thinks the stronger businesses will be the ones that can deliver reliable, useful outputs quickly and at a cost that makes commercial sense.
That is the heart of his argument. The long-term winners in AI may not be the companies with the biggest story. They may be the ones with the most disciplined systems.
Why has constraint worked in his favour before
This is not a new belief for Morris. If anything, it is one of the most consistent themes across his career.
One of his earlier successes was Udate, the online dating business he founded in 1999. It grew quickly as internet adoption accelerated, reaching more than one million users by 2000. Just as the company was preparing to float, the dot-com bubble burst.
“It was a nightmare,” Morris says. “We thought we were going to raise hundreds of millions of dollars, but we ended up raising $7.5m on a reverse takeover.”
At the time, it looked like a serious setback. With hindsight, he sees something more useful.
“The fact that we raised less money meant we had to be much more efficient,” he says.
That period appears to have sharpened one of his strongest operating instincts. Easy capital can make weak decisions easier to live with. Tighter conditions tend to expose what matters faster. A business has to become clearer about where value sits, where waste is building up and what customers actually come back for.
Udate went on to become the second most popular online dating site and the most profitable internet company in the UK before being sold to USA Interactive, the owner of Match.com, in 2002 for £93.6 million.
For Morris, the lesson was practical rather than philosophical. Efficiency is often treated as something a company worries about later, once growth is established. He sees it as part of what makes durable growth possible in the first place.
The lesson he took from King and Candy Crush
That same logic shaped his thinking at King, where he was an early investor and later chair.
When he first backed the business in 2003, King was focused mainly on skills-based online games. The issue was not whether people enjoyed them. The engagement did not last.
“We were looking at games like FIFA, which had the benefit of being able to reincarnate the same game every year with new players and features,” Morris says. “In contrast, we were making games that would effectively fizzle out.”
That realisation forced a rethink. Instead of concentrating on games that peaked and faded, King moved towards titles that were easier to pick up, more habit-forming and better suited to repeat engagement. That shift eventually produced Candy Crush Saga, one of the defining successes of the mobile gaming era.
The numbers behind Candy Crush are part of its legend, but what matters more in Morris’s account is how the business got there. It was not about novelty for its own sake. It came from identifying a weakness in the model, understanding user behaviour more clearly and building around retention rather than one-off appeal.
That feels relevant to his AI thinking now. Product strength is not just about whether people try something once. It is about whether the product solves a meaningful problem well enough, quickly enough and consistently enough to become part of how people work.
Why does he think the UK still has a real opening in AI
Morris’s broader argument about Britain is grounded in that same discipline. He is realistic about the country’s limits. The UK does not have the same domestic scale, capital depth or infrastructure base as the US. He does not think that rules it out of the AI race.
Instead, he sees a more specific opening. British technology businesses have often had to operate under tighter constraints, which has pushed them towards greater efficiency, sharper commercial judgement and more disciplined execution. In a market where the cost base is still high and the economics are not yet settled, those traits may become more valuable than they first appear.
The UK, in his view, does not need to dominate every layer of the AI stack to produce meaningful AI businesses. It needs to find the parts of the market where architecture, application and operating discipline matter more than brute force.
That is a more grounded version of AI ambition than the one often heard in policy rhetoric, but it is also more credible. Morris is not claiming Britain will win on scale. He makes the case that there is still room to improve execution quality.
A more practical way to look at the AI market
What makes Morris worth listening to on AI is that he is not approaching it like a commentator taken in by the noise around the sector. He sees it as someone who has built and backed technology businesses across very different eras, from early computing to the consumer internet, gaming, and now AI.
That changes the emphasis. The key question is not whether AI is transformative in theory. It is whether the businesses forming around it can produce useful outcomes at a cost, speed and reliability that hold up in the real world.
That is where his focus on efficiency stops sounding like a personal preference and becomes a commercial filter. How much waste is built into the system? How much friction sits between the user’s question and a genuinely useful answer? What additional cost is incurred because the architecture is clumsy? How much of the market is still judged on appearance rather than substance?
Morris has been around long enough to know those questions matter. The early excitement in any technology wave can carry businesses a long way, but eventually the market asks harder things of them. Can the product work consistently? Can it scale economically? Can it solve a real problem better than the alternatives?
For all the talk of AI supremacy, his own view is more measured than that language suggests. The winners may not be the companies with the most compute, the most capital or the loudest story. They may simply be the ones who learn how to make the technology work better.
For a country like the UK, and for businesses trying to build in a more disciplined way, that may be the most useful advantage of all.
