MHRA's 2030 Strategy Bets on an AI Centre and Lifecycle Regulation to Fix the Approval-to-Adoption Gap

The Medicines and Healthcare products Regulatory Agency laid out a sweeping five-year agenda on 14 October, promising structural changes to how AI-enabled medical products are approved, monitored and updated across their working lives. The MHRA Strategy 2030 arrived just eight days after ministers accepted all 44 recommendations from the NHS doctor-led National Commission into the Regulation of AI in Healthcare, making this arguably the most concentrated burst of regulatory reform activity the UK medtech sector has seen in a generation.

From One-Off Assessment to Lifecycle Monitoring

The Commission's central finding was blunt: current regulatory approaches were designed for products that are static and easier to assess at a single point in time. AI-enabled devices are neither. They iterate, drift, and can perform differently depending on the workflow, data and clinical setting around them. The government accepted that verdict in full on 6 October, committing to a framework in which AI-enabled medical devices are assessed and monitored throughout their working lives rather than through over-reliance on one-time assessments.

The practical machinery for that commitment is already in motion. The MHRA will publish draft guidance by December 2026 on predetermined change control plans, a mechanism that would allow manufacturers to establish defined boundaries within which they can update a device without triggering a full re-authorisation. The regulator will then open consultations in 2027 on how to qualify and classify AI-enabled devices. A full implementation roadmap covering all 44 recommendations, with timelines and responsibilities assigned, is due by spring 2027.

None of this is binding yet. The acceptance commits the government to developing the reforms; it does not make proposed regulatory changes immediately effective for NHS organisations, clinicians or developers. That is worth stating plainly, because the sector has a habit of mistaking a ministerial green light for a new rule.

The 2030 Strategy's AI Architecture

Beyond the Commission response, the MHRA's own strategy to 2030 introduces several distinct commitments. The headline is a new Centre for AI and Digital Health, intended to give developers of AI-driven medicines and devices much clearer regulatory expectations. An AI Medicines Development Action Plan is due in 2027. Inside the agency itself, AI will be deployed operationally: a Clinical Trial Protocol Checker is being developed to help sponsors produce higher-quality applications before submission, and AI will be used to hunt for safety signals in real-world data.

On efficiency, the numbers are striking. Scientific advice turnaround is targeted at 45 days rather than the current 90, and qualifying Phase 1 trials will get a first assessment in 14 days. Those are the kind of figures that make commercial development teams pay attention.

The strategy also commits to expanding international reliance routes with trusted regulators including the US FDA, and to continued recognition of CE-marked devices. For smaller companies, a targeted voucher scheme is planned to reduce regulatory barriers. Health innovation minister James Frith said the strategy would help the agency keep pace with the latest innovations in AI, data and medicine.

AI Airlock Phase 3: the Deadline Is Now

Running alongside both the Commission response and the 2030 strategy is the opening of AI Airlock Phase 3, the MHRA's regulatory sandbox for AI-enabled medical devices. This phase is funded until April 2029 following three years of additional government backing, and its focus is deliberate: post-market surveillance and lifecycle regulation. Phase 3 will test how AI devices can be monitored and managed safely after deployment, generating evidence intended to inform the very MHRA guidance and policy currently being drafted.

Applications are open now. Developers seeking inclusion in the current review wave must apply before 31 October 2026. A webinar for prospective applicants runs on 22 October, tomorrow from the date of this article. The first cohort will be selected in November. There is no application fee, though candidates are expected to fund their own studies within the sandbox.

What the Trademark Data Signals

Commercial intent across the AI medtech space is, unusually, subdued on one specific measure. AI Business Dispatch analysis of IPO (TMD) data as of October 2026 shows just 228 UK trademark filings in NICE Class 10 in Q4 2026, a fall of 81% against the prior period. Class 10 covers surgical, medical and veterinary instruments and apparatus, making it the closest trademark proxy for physical AI-enabled medical device activity.

The drop is sharp enough to warrant scrutiny. One plausible reading: developers are holding off on brand registrations while the regulatory architecture settles, preferring to commit trademark spend once the MHRA's classification and change-control guidance is published. If so, the December 2026 draft guidance could trigger a rebound in filings. A second reading is more cautious: uncertainty about lifecycle obligations may be slowing the pipeline of market-ready products. The MHRA's AI Airlock and its 2030 centre are, in part, a bet that clarity will unlock that pent-up activity.

The Adoption Question Remains Open

Regulatory reform is necessary. It is not sufficient. The gap between MHRA approval and NHS adoption has been the industry's standing complaint for years, and the techUK Health and Care Summit held on 19 October noted that the Commission's acceptance came just two days before the event, giving delegates little time to absorb it. The European Health Data Space was flagged as a model the UK should consider aligning with on data sharing, a reminder that post-Brexit divergence carries its own costs.

The MHRA's new adaptive framework, if delivered on schedule, should give AI developers a more predictable commercial pathway. Whether NHS trusts and integrated care boards then actually deploy what gets approved is a separate policy problem, one the 2030 strategy does not fully solve.