RUNWAYVC: when LPs become startups’ first customers

With a €40 million first close for its second fund, RUNWAYVC could look like just another fund dedicated to industrial AI, robotics and automation. Yet its distinctive feature lies less in its size than in the composition of its investor base.

AKER, HALLIBURTON, AKER BP and AKER SOLUTIONS are not merely potential sources of capital: they own the factories, infrastructure and equipment, and control the budgets, within which the technologies backed by the fund will have to prove their value.

RUNWAYVC is therefore making its LPs part of its investment infrastructure.

Forty million euros is a respectable first close for a Nordic early-stage fund. Yet it does not explain why RUNWAYVC’s second fund deserves closer attention.

The Norwegian fund plans to make around 20 investments over the next three to five years, primarily in Norway and across the Nordic region, with a handful of deals elsewhere in Europe and the United States. It will invest from pre-seed to Series A in industrial artificial intelligence, software, robotics, automation and autonomous systems, with initial tickets ranging from NOK 5 million to NOK 10 million.

One more specialist fund in a European market that now has plenty of them. The difference emerges when the focus shifts from what RUNWAYVC finances to who finances RUNWAYVC.

AKER, the sole investor in the first fund, remains the cornerstone investor in Fund II. It is now joined by HALLIBURTON, AKER BP and AKER SOLUTIONS, as well as KLP, Norway’s largest pension fund; state-owned INVESTINOR; Norwegian industrial families; and several investors from the technology and finance sectors.

All of them provide capital. But some can offer something far harder for an industrial startup to secure: a place where its technology can actually operate.

In Physical AI, access to the field becomes a form of capital

A SaaS startup can build its product in the cloud, distribute it over the Internet and begin acquiring customers without seeking permission from a Fortune 500 company every time it wants to test a new feature.

Physical AI operates according to a different economic logic. A robot designed to work on an offshore platform must go to an offshore platform. A system intended to make an industrial vehicle autonomous must take control of an industrial vehicle. Predictive maintenance technology must gain access to machines, their data and their operating history.

Between the prototype and the commercial product lies a stage that is particularly difficult to finance with capital alone: the real world.

The startup must find an industrial company willing to open up its factory, warehouse or infrastructure. It must convince operational, technical, cybersecurity and safety teams. It must connect its technology to existing systems, some of them decades old. It must demonstrate that the product still works when confronted with dust, vibrations, bad weather, shift changes or simply a machine that does not behave exactly as it did in the laboratory.

And once everything works, procurement still has to be persuaded. None of this is a small undertaking.

In this economy, access to the field becomes almost a form of capital. And that is precisely what several of RUNWAYVC’s investors possess.

HALLIBURTON operates at the heart of the global oil and gas industry. AKER BP manages energy assets in the North Sea. AKER SOLUTIONS supplies equipment, technologies and services to energy infrastructure.

For a startup developing robots or autonomous systems for heavy industry, these groups are not merely prestigious names on a list of LPs. They are places where its products could become useful.

The LP sits at both ends of the chain

Traditional venture capital separates roles fairly clearly.

The LP provides capital to the fund. The fund selects startups. The startups use the money to build their products and find customers.

The chain is relatively linear:

LP → VC → startup → customer.

RUNWAYVC is attempting to turn it into a loop.

Industrial LP → RUNWAYVC → startup → pilot with the LP → contract → commercial reference.

The same industrial group can therefore sit at both ends of the chain. It indirectly finances a technology through the fund and may then become one of its first users.

RUNWAYVC is explicit about this model: several of its LPs are themselves industrial operators and therefore potential buyers of the technologies it backs. That materially changes a fund’s value proposition.

When an entrepreneur raises a few hundred thousand euros from a traditional VC, what they primarily gain is time: a few additional months to recruit, develop the product and find a market. When they also gain access to a major industrial company capable of testing that product, the fund may also shorten their route to market.

The best LP is no longer necessarily the one that contributes the most money. It may be the one capable of signing the first purchase order.

MINERVA and HIVE already show how the model works

The first two investments from Fund II help explain this logic.

MINERVA HUMANOIDS is developing humanoid robots designed to perform dangerous and demanding industrial operations. Its challenge will obviously not be limited to building a humanoid capable of walking, manipulating objects or completing a sequence of tasks. The company will also have to determine how many hours the robot can operate without human intervention, which operations can genuinely be entrusted to it, how it responds when its environment changes, how it fits into safety procedures and, above all, whether its cost is lower than the economic value of the work it automates.

A demonstration can prove that a robot works. Only an industrial site can prove that it is useful.

The second investment, HIVE AUTONOMY, makes the mechanics of the model even clearer. The Norwegian company is developing a Physical AI layer capable of making existing industrial machinery autonomous. Its technology is already being used at several sites across Scandinavia, and the company recently announced a $15 million pre-Series A round led by SUPERSEED.

HIVE does not necessarily need industry to purchase an entirely new generation of machinery. It aims to add intelligence to an installed base already representing billions of euros in assets. To prove that proposition, it needs machines, sites and operators—exactly what the industrial ecosystem surrounding RUNWAYVC possesses.

The first customer generates more than revenue

In software, the first customer generally provides an initial measure of product-market fit. In industry, that customer serves several additional purposes. First, it validates the technology: a machine that operates for several months in a real industrial environment provides evidence that no demonstration can fully replace. It then validates integration. Industrial technology rarely operates on its own; it must communicate with machinery, software, procedures and organisations already in place.

Finally, it validates the product’s economics. An industrial group does not pay because a robot is impressive or an algorithm sophisticated. It pays because the technology can increase output, reduce costs, prevent an accident, improve equipment availability or replace an operation that has proved difficult to automate.

Above all, the first major customer provides a reference. An unknown startup telling an industrial company that it can automate part of its operations is essentially asking that company to take a risk. The same startup, once it can demonstrate that a group such as AKER or HALLIBURTON already uses its technology, is no longer selling quite the same proposition.

The first deployment turns a hypothesis into a precedent. That precedent makes the second contract—and the next funding round—easier to secure.

The first customer therefore does more than generate revenue. It gradually turns technological risk into execution risk.

RUNWAYVC pools part of the corporate venture function

The model also offers a different way to understand the relationship between large companies and venture capital. Until recently, an industrial group seeking access to startups had three main options.

It could create its own corporate venture fund, establish open-innovation programmes or wait until the technologies were mature enough to acquire them.

Each approach has its limits. Building a CVC requires a team, an investment strategy and enough deal flow to justify the organisation. Open-innovation programmes often generate plenty of pilots but relatively few deployments. Acquisitions, meanwhile, tend to occur after most of the technological value has already been created.

RUNWAYVC offers an intermediate architecture: several industrial groups become LPs in a specialist fund. RUNWAYVC handles sourcing, selection, investment and portfolio support. In return, the industrial companies gain privileged access to a portfolio of technologies that directly address some of their operational challenges.

The fund therefore becomes a form of pooled corporate venture capital without being confined to the strategy of a single group. For the industrial company, the investment creates optionality: it can monitor technologies, work with selected startups, potentially become a customer and, over time, consider more substantial partnerships.

For the startup, it potentially shortens the distance between investor and buyer.

AKER first tested the model on itself

This architecture did not begin with Fund II. RUNWAYVC’s first fund had an unusual feature: it had a single LP, AKER.

Since 2022, the team says it has invested in 24 companies, completed 23 follow-on investments and recorded two exits. Its portfolio companies have reportedly raised more than NOK 2 billion from external investors.

Above all, the portfolio reveals the industrial coherence the fund is seeking.

SONAIR develops 3D ultrasonic sensors, particularly for autonomous robot perception. OTEE works on industrial automation. WSENSE develops underwater communication technologies. Other portfolio companies focus on the maintenance and operation of industrial equipment.

Fund I can therefore be read as an initial experiment that used the AKER ecosystem as an interface between venture capital and industry. Fund II is now attempting to broaden that interface.

HALLIBURTON, AKER BP and AKER SOLUTIONS are not merely increasing the vehicle’s financial capacity. They are potentially increasing the number of industrial problems to which the startups can be exposed, the number of sites they can access and the number of prospective first customers they can convince.

RUNWAYVC is therefore gradually evolving from a fund backed by the AKER ecosystem into a platform connecting several parts of Norwegian and international industry.

A €40 million fund is easy to copy; its environment is not

RUNWAYVC is based at Aker Tech House near Oslo, within a broader complex that includes RUNWAYFBU, more than 60 industrial and technology companies, an AI and robotics laboratory, an incubator and a range of industrial partnerships. This infrastructure helps reveal where the fund’s real barrier to entry may lie.

Raising €40 million is not trivial, but dozens of European funds now have that capacity. Recreating a network of major industrial companies willing to share their problems, open their sites, mobilise their engineers, test still-imperfect technologies and eventually purchase them is far more difficult.

Capital can be raised. Industrial trust must be built. For a fund specialising in Physical AI, the latter may become a greater competitive advantage than the size of the vehicle itself.

The LP must not become the market

The loop nevertheless carries a risk: a major industrial company can be an excellent design partner and a poor market.

A startup that gradually adapts its product to the demands of a single group may turn a generic technology into a bespoke development project. Each new requirement adds another feature, each site demands a different integration, and what was supposed to become a scalable product ends up looking like a services business.

The relationship between LPs and portfolio companies must therefore retain a boundary. The industrial group’s role is to help the startup find its market, not to become its market.

The first customer must demonstrate that the technology works well enough to be sold to the second, the tenth and eventually the hundredth. Only under that condition will RUNWAYVC’s model be able to extend beyond the Norwegian ecosystem.

RUNWAYVC’s real asset may not be its fund

The €40 million gives RUNWAYVC the means to invest, but in this configuration it will not be enough to determine Fund II’s performance. The true indicator lies elsewhere: how many portfolio companies will the fund manage to move from prototype to pilot, from pilot to contract and from a first contract to the international market?

The question extends well beyond RUNWAYVC. As venture capital turns towards robotics, autonomous systems and Physical AI, funds will have to learn how to finance companies whose primary constraint is not always access to capital.

They need machines on which to experiment, factories in which to deploy, engineers capable of confronting their technologies with operational constraints and companies large enough to accept the risk of becoming their first customers.