London and Toronto-based Intrepid Growth Partners has closed its first fund at $525 million, assembling sovereign, state-backed, institutional, and private investors to back growth-stage AI companies trying to redesign established industries rather than add another layer of general-purpose software.

The fund has more than 80 limited partners, with named backers including Singapore’s Temasek, Abu Dhabi Investment Council, the British Business Bank, Export Development Canada, and Business Development Bank of Canada. Its close gives Intrepid a substantial pool of growth capital at a point when policymakers in the UK, Canada, and Europe remain concerned that promising AI companies can raise early-stage funding domestically but turn elsewhere when they need larger cheques to expand internationally.

Intrepid was founded in 2023 by Mark Machin, Mark Shulgan, and University of Toronto economist Ajay Agrawal, and operates from London and Toronto. The firm has already built a portfolio spanning engineering, cybersecurity, healthcare, advertising technology, enterprise software, and other application-layer AI businesses.

One of its UK investments is London-based PhysicsX, which raised a $300 million Series C at a valuation of about $2.4 billion in June. PhysicsX develops AI models for engineering simulation, allowing customers in industries including aerospace, semiconductors, automotive, energy, and materials to evaluate physical designs far more quickly than conventional high-fidelity simulation alone.

Growth capital remains a policy problem

The presence of several public or state-linked institutions in Intrepid’s limited-partner base illustrates how governments are attempting to influence the financing system without directly choosing every company that receives investment. Instead of making grants to individual startups, institutions such as the British Business Bank can commit capital to professional fund managers, which then make commercial investment decisions across a portfolio.

That model is familiar in venture capital, but AI has given it additional political weight because governments increasingly treat access to growth finance as part of technology strategy. Early-stage ecosystems can produce strong research spinouts and startups without necessarily retaining them as independent companies once the capital required for international sales, acquisitions, infrastructure, and recruitment rises into the tens or hundreds of millions.

Intrepid is deliberately operating in that later part of the market. Its portfolio already includes companies such as PhysicsX, healthcare AI business Iterative Health, cybersecurity company Exein, mechanical-engineering collaboration platform CoLab, advertising software group StackAdapt, and Skin Analytics, which applies AI to skin-cancer detection pathways.

The firm’s investment thesis is built around what it calls system-level change. Rather than backing companies simply because they attach AI to an existing task, it looks for businesses capable of redesigning a larger workflow or operating model around cheaper prediction and machine intelligence.

That distinction is commercially important because many AI features are likely to become difficult to defend as underlying models improve and software incumbents add similar capabilities. A company whose value rests solely on generating a document or answering a question may find its function absorbed into a larger platform, whereas businesses embedded deeply in engineering, clinical, industrial, or regulated workflows can build additional barriers around proprietary data, integrations, specialist knowledge, and customer processes.

The UK–Canada corridor has practical logic

Intrepid’s dual base is unusual but not arbitrary. Britain and Canada both have strong university AI research communities and a record of producing technically sophisticated companies, while neither offers a domestic pool of late-stage technology capital comparable with the US.

The fund is positioning itself around that gap while investing more broadly across North America and Europe. Its advisers and partners include researchers and operators with backgrounds in machine learning, economics, institutional investment, defence, and technology companies, giving it a network intended to support businesses crossing from technical product development into larger-scale commercial deployment.

PhysicsX illustrates the type of UK company the strategy is targeting. Its software sits within industrial engineering, where faster simulation can affect aircraft components, semiconductor equipment, energy technology, automotive design, and advanced manufacturing. Those applications have longer sales cycles and higher integration requirements, but they also offer a clearer route to measurable economic value if the technology reduces engineering time or improves physical products.

Public money does not remove venture risk

The involvement of state-backed investors should not be mistaken for an endorsement of individual portfolio companies or a guarantee of returns. Limited partners provide capital to a fund whose managers retain responsibility for investment selection, while growth-stage AI valuations remain exposed to shifts in technology, competition, capital markets, and customer adoption.

The $525 million vehicle is also modest beside the multi-billion-dollar funds raised by some large US venture firms, meaning Intrepid is competing through specialisation and access rather than balance-sheet scale. Its own portfolio suggests it is willing to syndicate deals with much larger investors, as it did in PhysicsX’s Temasek-led financing.

Where the fund becomes more interesting from a European perspective is the combination of private investment with public institutions trying to keep a route open between domestic research and global-scale companies. Britain has spent years debating the shortage of later-stage capital available to technology businesses, while pension and institutional investors have been encouraged to direct more money towards private growth assets.

Intrepid’s close does not resolve that structural gap, but it adds another specialist investor capable of writing growth cheques from a London base and bringing international state-backed capital into the same pool. The harder test will arrive when the performance of its portfolio shows whether concentrating on industry-specific AI produces durable companies rather than a well-capitalised collection of applications built during an investment boom.