AI patents just became an M&A risk you can’t ignore

Investors will increasingly ask whether your AI can scale without infringing, and that changes deal diligence fast.

Article link: https://open.substack.com/pub/johanosteyn/p/ai-patents-just-became-an-m-and-a?r=73gqa&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Audio summary: https://youtu.be/LUNwJd89mdU

Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/

If you are a business leader buying an AI capability, the conversation has usually sounded familiar: Do they have the right talent? Is the model good enough? Is the data defensible? Can we integrate it quickly? The UK Supreme Court has just added a sharper, more uncomfortable question: can you legally operate at scale without stepping on someone else’s intellectual property? 

In a recent ruling, the Court opened the door to certain artificial neural network inventions being patentable in the UK, sending the issue back to the UK Intellectual Property Office to reconsider a specific application. That may sound technical, but the commercial impact is simple. When patentability expands, the cost of getting M&A diligence wrong goes up.

CONTEXT AND BACKGROUND

The ruling, reported by Reuters, is being interpreted as a meaningful shift in how UK law treats software and AI inventions, with lawyers suggesting it could make the UK a more attractive place to protect AI innovations. Specialist legal reporting has gone further, arguing that the judgment throws out a long-standing UK approach and aligns the UK more closely with European practice, which could change how companies draft and prosecute software-related patents.

Zoom out, and you see why boards should care. We are in a period of heavy AI deal-making, with buyers chasing capability, speed, and defensibility. Crunchbase has highlighted expectations of continued big AI deals and consolidation dynamics as part of wider 2026 tech and venture trends.

INSIGHT AND ANALYSIS

Most M&A teams are comfortable assessing what a target owns. They are less comfortable assessing what a target might be infringing. That is the difference between ownership and freedom to operate. You can own code, have brilliant engineers, and still be blocked by someone else’s patents, licensing terms, or litigation threats.

If AI-related patents become easier to obtain, you should expect two things. First, more filings. Second, more “patent thickets”, overlapping claims that create uncertainty about who can ship what. That uncertainty does not stay in the legal department. It shows up as delayed product launches, higher insurance costs, more aggressive indemnity negotiations, and bigger holdbacks or price chips in deals.

This is not theoretical. We have already seen how fast AI-related legal risk can swing the narrative and the cost structure of an AI business. The Financial Times’ coverage of the Getty Images case in the UK underscored how unresolved questions about rights and precedent can hang over companies building or acquiring AI capability.

The next phase is that patents may join copyright and data rights as a mainstream AI diligence category. In plain language, buyers will start asking: What patents do you rely on? What patents might you be stepping on? What licences, open-source obligations, and supplier restrictions sit in the stack? And if we roll this out globally, do those answers change?

IMPLICATIONS

For CEOs and boards, this is not a reason to stop investing in AI. It is a reason to professionalise AI diligence. Treat freedom to operate like cybersecurity: you do not need perfection, but you do need visibility, mitigation plans, and informed sign-off.

For CFOs and deal teams, expect diligence checklists to expand. The target’s patent portfolio matters, but so does their exposure. Have they had infringement threats? Have they built around known patents? Do they have a clean chain-of-title for the core IP? Are there contractual warranties from suppliers, and are those warranties meaningful?

For general counsel and procurement leaders, vendor risk will creep into M&A risk. Many AI products are built on third-party model APIs, embedded libraries, and platform tooling. If a vendor gets pulled into a patent dispute, your roadmap can get pulled with it. The legal story around AI is widening, and recent reporting shows how quickly AI-related disputes and settlements can become a cost line item.

CLOSING TAKEAWAY

The UK Supreme Court’s decision is a reminder that AI advantage is not only technical. It is legal, contractual, and operational. As patentability expands, investors and acquirers will increasingly treat freedom to operate as a core value driver, not a footnote. The winners will be the firms that can show both innovation and defensibility: clear IP ownership, credible licensing positions, and a sober map of where the risks sit. In a market racing to buy AI capability, the smartest leaders will remember that scale is only an asset if it is legally sustainable.

Author Bio: Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net