Four trends reshaping UK pre-seed in 2026

According to Sifted, SFC Capital was one of the most active venture capital investor in H1 2026. This is in line with our ‘volume with discipline’ strategy that I described in a previous article.

We’re certainly seeing a lot of activity in the UK pre-seed market and I thought that it would be interesting to highlight some of the new trends driving our recent investments.

The sector mix of our investment is basically unchanged – we’ve invested across B2B software, health, hardware and robotics, green tech, fintech as well a few consumer investments. However, what is changing are the profiles of companies that we are seeing today, which are materially different from those raising money even two years ago.

Four trends stand out at the moment in my view.

1. AI is moving from simple automation into complex, high-value workflows

The first wave of AI startups largely focused on relatively ‘obvious’ opportunities: generating marketing copy, summarising meetings, answering customer questions or automating administrative tasks.

Those products demonstrated what the technology could do, but they were often straightforward to replicate. In many cases, they rested primarily on access to the same underlying foundation models available to everybody else.

The companies that we are now seeing are tackling much more difficult processes and are integrating AI into workflows that involve proprietary data, industry-specific knowledge, regulatory constraints and security requirements. The founders are often coming directly from the industry and have experienced themselves the particular problem that they decided to solve.

A great example from our recent cohorts is Kord (previously Checkboard). Kord brings identity verification, anti-money-laundering checks, client onboarding, document signing and payment processing into a single platform for regulated businesses (law firms, conveyancers, financial services firms etc.).

The problem they solve has become more pressing as AI makes fraud and forged documents increasingly sophisticated. Kord not only checks identities and documents against a broad range of data sources but also maintains the audit trails required in regulated transactions.

Another recent example is Nayara AI, which builds clinical-grade AI agents around specific patient-care pathways. Its platform supports processes such as pre-operative preparation, medication adherence and patient follow-up, while generating structured risk flags and summaries for clinical teams.

This is considerably more complex than deploying a generic chatbot. Nayara works with healthcare providers to map the complete patient journey and define rules of how the AI should respond.

These businesses point towards the next stage of AI adoption. The value is moving away from the model itself and towards everything surrounding it: handling proprietary and sensitive data, providing deep industry expertise, integrating into existing systems, ensuring security and auditability etc.

On the other side of the equation, we are also seeing a greater openness from large organisations to work with early-stage startups and actually implement these solutions beyond simply piloting them, which allows startups to generate substantial revenue sooner.

2. AI is making hardware and robotics less hard

An interesting secondary impact of AI is that it is reducing the cost of developing, launching and operating sophisticated hardware.

Historically, hardware companies were often difficult propositions for pre-seed investors. They required substantial upfront capital, long development periods and large operational teams. Prototypes could be expensive, iteration cycles slow and gross margins uncertain.

Many of those challenges remain. Hardware is not suddenly easy. But AI, computer vision, simulation, edge computing and increasingly capable off-the-shelf components are changing the economics.

Hardware can now be monitored and controlled more intelligently. A machine can adapt to its environment, detect faults, collect data and operate with less human intervention. Software-based improvements can also increase the capabilities of deployed hardware without requiring the entire physical product to be redesigned.

This can make a hardware startup more scalable and, in some respects, more software-like in its business model. The line between software and hardware is getting blurred and we are seeing more credible robotics and advanced-manufacturing opportunities at pre-seed stage that don’t necessarily require millions of pounds of capital to get to market.

ScrubMarine, for example, is developing autonomous underwater robots that clean and inspect ship hulls. Its lightweight Turtle robot is designed to remove biofouling while capturing inspection data, replacing work that can otherwise require dry-docking or hazardous diving operations.

Fairfield Vision combines specialised sensors and computer vision to detect crop disease, forecast yields and generate actionable maps for growers. The resulting intelligence can also support autonomous agricultural applications.

Hydra Manufacturing, a University of Leeds spinout, is developing a hybrid additive-manufacturing platform for advanced ceramic components. Its system uses AI-powered defect detection to identify and autonomously correct problems during printing rather than discovering them after a part has failed.

We expect to see more companies combining robotics, sensors and domain-specific AI to address labour shortages, safety risks, infrastructure maintenance, agriculture, manufacturing and environmental monitoring.

3. University spinouts are becoming more attractive

UK universities have always produced exceptional research. Historically the challenge has sometimes been to convert that research into companies with structures, teams and commercial plans that work for external investors.

We see this changing rapidly as university incubators and commercialisation programmes are becoming more professional and increasingly standardised. Founders are generally better prepared, university teams have a clearer understanding of venture financing, and investment terms are becoming more attractive.

The adoption of the University Spin-out Investment Terms guides has helped establish clearer expectations around equity, licensing, royalties and due diligence. More than 70 UK universities have now adopted the guides, according to TenU, while the Royal Academy of Engineering reports that average university equity stakes have fallen to approximately 16% - down from the 22–25% range seen before the 2023 reforms.

There is still considerable variation between institutions and individual transactions. However, the trend is clearly positive: negotiation and deal execution is becoming faster, cap tables more investable and incentives better aligned between universities, founders and investors.

The old stereotype of a university spinout limited to drug development programmes requiring enormous amounts of capital and ten or more years before commercialisation is also becoming increasingly untrue.

Our recent spinout investments are not research projects, they are startup companies commercialising valuable intellectual property with identified customers and routes to market. Some recent examples include:

• Instruct3D, originating from the University of Sheffield, which combines physics-based modelling, sensors and data to improve the reliability and certification of additive manufacturing;
• Hydra Manufacturing, from the University of Leeds, developing intelligent manufacturing equipment for advanced ceramics;

• RemePhy, an Imperial College London spinout using plants and soil bacteria to remediate land contaminated by heavy metals;

• Neubond, another Imperial spinout, developing wearable neuromuscular technology for stroke rehabilitation.

University spinouts are now a significant part of our dealflow, sourced through direct relationships with universities across the UK and through our work as an approved investor partner of Innovate UK, which provides another important source of high-quality deep tech ventures, combining our equity investments with non-dilutive R&D funding.

4. Pre-seed is becoming the new seed

The final trend has to do with the terms of pre-seed rounds and what founders are expected to have achieved before raising their first funding round.

Based on our own investment data, average pre-seed round sizes are now approximately 40% higher than they were two years ago, while entry valuations have also increased by around 25%.*

At the same time, the companies raising these rounds are more advanced. Pre-seed companies increasingly look like seed-stage companies did two years ago.

More than half of the businesses we now invest in are already generating revenue at pre-seed. Only a couple of years ago, that figure was closer to 30%.

This reflects how much founders can accomplish without external capital. Cloud infrastructure, AI-generated software development, no-code tools and easier distribution allow small teams to build and test products much faster.

A company that might once have raised its first round with a presentation and prototype may now have a functioning product, several customers and real-life revenue.

This is positive in several respects. It provides investors with more evidence and can reduce product market fit risk. It also raises an important question: if companies can reach revenue with less external capital, will they ultimately need fewer funding rounds before reaching Series A, profitability or an exit?

That could produce better outcomes for founders and early investors through less dilution and simpler capital structures. However, that theory is not yet proven and it will take at least 2-3 years to be validated by the market.

Series A expectations have also increased, and many companies are using larger pre-seed rounds to fund more ambitious milestones. Higher entry valuations will only benefit investors if eventual exit values increase correspondingly. If they do not, paying 25% more at entry could have a negative effect on long-term fund returns.

Discipline matters more than ever

Innovation cycles are clearly accelerating and a category that appears highly attractive today can be commoditised quickly by a new model, platform feature or open-source product.

This is one reason we believe a generalist strategy remains particularly valuable at pre-seed. It allows us to compare opportunities across sectors and avoid assuming that the most heavily promoted theme will necessarily generate the best returns.

Across AI, robotics, university spinouts and the wider pre-seed market, the fundamentals remain largely unchanged: does the product solve an important problem? Does the team have the ability to build the right product and commercialise it? Is there credible evidence that customers are ready to pay? Can the company build something difficult to replicate? And does the valuation leave sufficient room to generate long-term returns?

The tools available to founders are changing fast and are making it easier to launch a business and generate real-life traction with limited capital. However, this does not mean that all investment opportunities become automatically more attractive and as an investor it’s essential not to lose sight of the fundamentals across all the noise.

Figures relating to SFC’s applications, investments, round sizes, valuations and portfolio revenue are based on internal SFC Capital data as of August 2026. This article is provided for information only and does not constitute investment advice.

* Year-over-year analysis of the first tax year cohort. 18 companies from Angel Fund SEIS 24/25 T1 vs. 16 companies Angel Fund SEIS 26/27 T1.

Capital at risk. Past performance is not indicative of future performance.