The New Frontier - Your weekly science and innovation update

There were clear signals of confidence in biotech this week, alongside a flurry of UK government announcements further positioning AI at the centre of industrial strategy. Below we round up the key market moves, policy updates, and the latest numbers we’re tracking at Knight Frank.

Biotech: leasing, sentiment, and more dry powder

There were some signals of confidence in the biotech sector this week. Our US partner Cresa’s Q4 2025 San Francisco Lab Market Overview reports a stronger finish to the year. Deal counts up 17.5% year-on-year and total leasing volume up 10.5%. Versus 2023, deal count rose 62.9% and volume 89.4%. A notable shift worth pointing out is that advanced manufacturing occupiers are now competing with life sciences for lab space, drawn by the same high‑power and technical infrastructure. This is widening the demand base for labs.

Endpoint’s latest Biotech Sentiment Survey shows the index rising to 90 in Q1 (from 78 in the prior quarter), still below the neutral 100. Executives report improving outlooks on company performance, fundraising, and partnering. Ongoing concerns include regulatory uncertainty, with little optimism for improvements in 2026. The survey covers 165 professionals spanning founders, investors, and corporate executives. If the confidence builds and sustains, we could see expansion-led demand.

Epidarex Capital closed $145mn at the first close of Fund IV, targeting company creation and early‑stage deals across the UK and US. Initial investors include the British Business Bank. As Christine Hockley, Managing Director and co‑head of Funds, notes: “The UK has a very strong science and research base, so what is needed is investors who specialise in nurturing and scaling those ideas, creating new, IP‑rich companies.” This and other recent fundraising activity add to the pile of dry powder targeting the sector.

The next phase in the UK’s AI growth strategy

Coinciding with the annual update to the AI Opportunities Action Plan, the government announced a series of measures:

• Lanarkshire named an AI Growth Zone, unlocking special support (including expedited planning) for relevant infrastructure.
• Free online AI courses via a new government‑industry partnership, with an ambition to upskill 10mn workers. Some have questioned the accessibility of the portal underscoring that execution will be critical to delivering on this ambition.
• AI in public services: a further partnership with Meta and Anthropic to drive AI adoption across transport, national security, and more.
• £36mn to upgrade the DAWN supercomputer at the University of Cambridge.
• More firepower for the Global Talent Taskforce, including specialist private‑sector head‑hunters, new relocation and set‑up functions, and an enhanced concierge offer focused initially on international AI talent.

The action plan launched in January 2025, set out three priorities: “lay the foundations to enable AI”, “change lives by embracing AI”, and “secure our future with homegrown AI”.

The plans most visible bet is physical compute and data centres. Ministers have designated five AI Growth Zones. On compute capacity, the government cites a tenfold increase in UK AI compute capacity from 2024 to 2025, including switching on Isambard-AI in Bristol.

On data, The National Data Library has been backed with over £100mn and is identifying high-value datasets that can be made “AI-ready”.  The headline move is health in the form of the new Health Data Research Service that will create a secure access point to large-scale health datasets.

And on skills, the government has a target of upskilling 10mn workers by 2030. It reports over 1mn AI courses have already been delivered, supported by industry partners, and wrapped in scholarship schemes designed to feed the high-end talent pipeline.

A central pillar of the plan is to make the state smarter through the application of AI. Through an AI Diagnostic Fund, 2.4mn scans are now AI-assisted. New tools include Extract (turns unstructured planning documents into usable data) and a separate tool targeting a 50% reduction in processing times for straightforward applications. The state is drawing on US partnerships. Anthropic with DSIT, while Meta‑based AI teams build open‑source tools for public authorities.

The plans most strategic ambition is to make the UK an “AI maker, not an AI taker”. To that end the government has created a Sovereign AI Unit to support UK AI firms and allocate compute and data assets.

What to watch next (four tests)

1. Can AI Growth Zones translate into real grids and buildings given planning delays and energy constraints?
2. Can AI pilots become national services that earn public trust amid data‑security concerns and a mixed record on public‑private delivery?
3. Can the upskilling ambition be executed effectively, training people in the right tools for the jobs of the future?
4. Can the UK build true AI sovereignty while leaning on American muscle?

A quietly important feature

The plan is rapidly generating new organisations. Some entirely new, other units, centres, and delivery teams that may require real estate or sit within existing government infrastructure. Examples include a network of robotic adoption hubs. Taken together, the UK could plausibly be heading towards 10–15 AI‑related entities under the plan’s umbrella.

Sector metrics that matter

At Knight Frank, we’re constantly crunching the data to understand what’s really driving the sector. Who is hiring, how big the future pipeline is, and where is demand building. Here are some standout stats.

1. The UK has a strong and increasing pipeline of AI start-ups. First-time equity investment into UK AI startups rose sharply in 2025, with just over 800 companies securing their first deal, up 20.88% from 2024. Half of these are headquartered in London.
2. Company incorporations also grew with a record 3,426 new companies formed in 2025.
3. There are 2,742 AI startups ($100k-15m total VC funding), 503 breakouts ($15-100m total VC funding) and 110 scale-ups ($100m+ total VC financing)
4. At the other end of the scale the UK is home to 19 AI unicorns, 17 of which are headquartered in London.  These unicorns have increased their year-on-year global headcount by just under 5,000 people. A number of these unicorns are among the top AI hirers in London over the past year. According to LinkedIn data for example Wayve has increased headcount by 52% and Synthesia by 43%.
5. 33% of all UK VC was raised by AI startups in 2025, the highest share to date.
6. While caution should be applied in making direct comparisons between global markets the latest Cresa San Francisco Office Market Update shows that leasing activity is being driven by AI firms. These occupiers are prioritising secure, move-in-ready space, with large open floor plates that support collaboration and talent attraction and plenty of amenities.

Other news that caught our eye

• Synthesia (London) raised $200mn in a GV‑led Series E, valuing the company at $4bn.
• Anthropic reportedly lifted its funding target to $20bn, aiming for a $350bn valuation.
• OpenAI is in talks to raise ~$40bn from NVIDIA, Microsoft, and Amazon as part of a $100bn round.
• LinkedIn reports AI is creating more roles than it is replacing in the near term. In the US, AI‑literacy job postings grew 70% year-on-year, over the last two years, employers created 1.3mn AI‑related roles (e.g., data annotators, AI engineers). Data centres created 600,000 new jobs globally over the past year and foundational AI model firms are adding headcount aggressively.
• The UK government launched a £20mn fund to accelerate defence tech procurement via fast‑tracked government contracts.
• Sanofi reported annual results: R&D spend up 6.6% to €2.3bn; guidance to spend €14–15bn on M&A this year.
• AstraZeneca outlined plans to invest €15bn in China through 2030.

And finally my colleague Shabab Qadar provides a synopsis of the British Land/Life Science REIT deal.

The acquisition of Life Science REIT represents a strategically compelling and financially disciplined move for British Land. The transaction is immediately earnings‑accretive, driven by cost synergies and operational efficiencies, while being neutral to EPRA NTA per share and secured at a 26% discount to Life Science REIT’s EPRA NTA, offering meaningful value‑uplift potential. Additional upside is expected from the lease‑up of newly delivered and vacant space, supported by the integration of the assets into British Land’s scalable and efficient management platform. Strategically, the acquisition significantly enhances British Land’s science and technology exposure by adding five well‑located Golden Triangle assets across London, Oxford and Cambridge, enabling an expanded occupier mix and strengthening the company’s position within the UK’s leading innovation clusters. The portfolio brings strong reversionary potential, with an 8% net reversionary yield, and provides a platform for British Land to apply its proven innovation‑campus capabilities—demonstrated on schemes such as One Triton Square—to drive long‑term performance and value creation.