What are the key themes for institutional investors when it comes to AI investments in the UK?

Gareth: The surging demand for AI applications is driving the need for the infrastructure on which AI runs. This translates to an even stronger interest from private equity, pension and sovereign wealth funds in semiconductor firms and their entire supply chain: wafer manufacturers, chip designers, equipment suppliers, testing companies - the whole ecosystem. We are also seeing investments in cloud computing providers and their supply chains, including data centres and networking equipment.

Finally, there is, of course, interest from VCs and strategic investors in companies developing LLMs and related applications.

How about companies outside the core AI value chain? For those integrating AI into their operations to achieve productivity gains or cost reductions, has a specific investment trend transpired?

Gareth: The deployment side is seeing mixed results in capital flow and profitability. I reckon we are in early stages there, echoing Amara’s Law: we overestimate the short-term impact and underestimate the long-term. Some successes exist, like in the digital advertising space, but caution remains. In highly regulated sectors like banking or insurance, key industries in the UK, it is going to be a gradual rollout, starting with lower-hanging fruits like customer support chatbots, automated credit card applications and KYC processes.

Are the benefits of AI outside the core AI industry tangible enough for institutional investors to formulate investment themes around them?

Mike: I believe we have reached an inflection point where there is consensus among investors about the transformative potential of AI technology. My impression from speaking to clients is that AI is considered vital. Investors are actively exploring themes around cost reduction, revenue generation and how to stay competitive in the era of AI across virtually all industries.

Have institutional investors started reorganising themselves to better reap the benefits of AI?

Gareth: We recently had dinner with a major private equity client and I got the opportunity to speak to their Digital Operating Partner. It was fascinating to hear how he's working with portfolio companies on implementing AI across different corporate functions to reduce costs and grow revenues.

Mike: Digital Operating Partners can help centralise AI expertise at the portfolio level. I think because LLMs have made AI now more accessible, not only CTOs and tech specialists but also the wider C-Suite and the investment team, can now brainstorm together on what’s working, how to implement digital transformation, and how to set the agenda.

AI talent is a scarce resource. How can investors and companies who get ahead of the AI curve ensure that talent can be retained?

Gareth: Talent retention in the AI space is indeed a major challenge. We’ve seen “acqui-hires” where entire teams move quickly, taking their expertise and sometimes licensing core IP to the acquirer. It's a difficult situation for investors to defend against.

Mike: During due diligence, investors are paying closer attention to employment agreements and any “tie arrangements” in place to retain talent. In the U.S., non-competes have become more problematic, and their enforceability is often questionable. Structuring the appropriate incentive schemes is key here.

What can be done during the due diligence phase to ensure that the target company has the necessary AI capabilities?

Gareth: It’s sometimes difficult to pinpoint what is truly unique and protectable about a company’s technology stack. This is especially true with generic tools or approaches, where the line between genuine innovation and smart application can be blurry. So, the key questions we like to ask are: Is the technology truly proprietary? Is the intellectual property vested in the target and not easily replaceable? If key personnel were to leave, how much of the technology’s value would remain with the target company?

How much are companies willing to disclose during due diligence? Is there an established market practice in the context of UK deals to help mitigate against IP leakage during due diligence?

Mike: It’s a difficult tension to navigate. On one hand, there’s immense value in fully understanding the company’s technology stack around AI and assessing whether there might be latent liability accumulating as a result of the particularities of its data sourcing and AI model build. On the other hand, companies are understandably cautious about revealing their data sources, model build, and other elements of their "secret sauce."

One approach we often see is a phased due diligence process. Early stages involve limited information on financials, followed by more detailed disclosure as the bidder pool narrows. The real “secret sauce” is often only revealed in the very late stages, in a heavily restricted data room with limited access. This approach protects more sensitive information until the final stages of negotiation. The downside is that it requires investors to commit significant time and resources without full visibility into the core technology. It’s a balancing act.

If you were promoting the UK AI industry to global investors, what policies or specificities make it attractive? How can it compete with the larger markets in the US and China?

Gareth: Overall, the combination of a potentially pro-business government, a more flexible regulatory landscape, and a strong focus on promoting the AI economy should make the UK a very attractive destination for AI investment in the coming years. This is especially true for US investors who may be looking for a more favorable regulatory environment than the one currently being proposed in the EU.

The recent elections provide potential, too. The new Labour government has already signalled the intention to introduce targeted AI regulations in the UK. This is coupled with a strong desire to kickstart the economy, and promoting the AI economy could be a visible way to do that. The King’s Speech announced plans for regulating developers of powerful AI models, but no firm bill has been introduced yet.

Tech Secretary Peter Kyle stated that Labour will develop a statutory code for AI safety, requiring developers to share safety test data with the government. This contrasts with the voluntary approach under the previous government. Kyle also mentioned plans to create a regulatory innovation office.

However, there is currently no indication that the Labour government is considering across-sectoral, comprehensive regulation on AI akin to the EU AI Act.

Are there specific sectors in the UK, like financial services, where the regulatory landscape is more conducive to AI innovation and experimentation than elsewhere?

But given the EU AI Act’s extraterritorial reach, how do UK companies who have customers in the EU avoid the tempering effect of EU regulation?

Gareth: It is true, as a UK company, you can't touch the EU market without being pulled into its regulatory framework orbit. For those companies with a strong EU customer base, compliance starts with the EU GDPR and the EU AI Act as a baseline, then any sector-specific requirements in the UK need to be considered. Possibly there will be less regulatory divide in the long term. We're likely to see elements of the AI Act mirrored in UK legislation, especially regarding safety requirements for systemically important AI technologies.

Returning to your earlier point, do UK companies face any regulatory advantages?

Gareth: I think it is about the how, as much as it is about the what is being regulated. For example, the UK’s AI Safety Institute is already working with model developers to assess risks and safety measures. But the UK hasn’t formally legislated like the EU. Instead, they’re calling in major players and requesting voluntary submission of details. It is a smart “below the radar” approach to regulation and allows the industry to participate in the rule-setting.

Anything else that makes the UK an attractive destination for AI investments? France appears to be making significant strides in attracting investment in AI technology as well.

Gareth: If you're coming from the US and looking for a base in Europe, the UK is definitely compelling.

But the UK has that Anglophone advantage, a familiar legal system with lots of case law, and a history of common law flexibility. That means you can structure deals with a good degree of certainty about the outcome.

The UK legal system is also known for being relatively efficient. You can get things done quicker than in other jurisdictions, which is appealing to investors. And the UK’s IP protection is strong but also foreseeable for the most part, despite some current areas of uncertainty in the most novel AI frontiers, which is important for AI companies.

Mike: Ultimately, it comes down to a reasonable outcome and speed. The UK offers a balance of legal certainty, flexibility, and efficiency that’s hard to find elsewhere, and that’s very attractive to investors in the AI space.

Turning our attention to the geopolitical environment, how open is the UK ecosystem to cross-border partnerships and foreign investment?

And with the recent change in government, do you anticipate any shifts in policy?

Mike: I don’t foresee a drastic change in approach. These companies require capital to thrive. Should domestic sources prove insufficient, access to foreign capital becomes essential. For smaller markets like the UK, there’s always the risk of becoming overprotective and encouraging companies to leave. The key will be finding a balance between safeguarding national security and fostering an environment that welcomes vital investment.

‍

Gareth Kristensen and Michael J. Preston are partners at Cleary Gottlieb.