Funding 8 min read

Solvera Raised $6.3M to Be the Engineering Firm Rather Than Sell It Software

The Oslo company signs as the engineering firm of record on Norwegian MEP work, bills for the outcome instead of licences or hours, and says it passed $1M in contracted revenue in three months with nine people. Seedcamp led the pre-seed announced on 15 October.

Exposed ventilation ductwork with circular diffusers running beneath a dark commercial ceiling, alongside lighting track and cable containment

Solvera, an Oslo company, announced a $6.3 million pre-seed round led by Seedcamp on 15 October. The money is unremarkable for the category. The structure it is funding is not. Solvera is not selling software to engineers: on its own account it takes the engineering work, signs for it, and charges for the result.

What Was Announced

The round was led by Seedcamp, with participation from Construct Venture, Inception Fund, Momentum, Obos Ventures, PSV Tech, Sondo and The Nordic Web Ventures. Solvera’s release names as angels the Spacemaker founders Anders Kvåle and Håvard Haukeland; Seedcamp’s own post additionally names former OBOS chief executive Daniel Kjørberg Siraj among the angel investors. Neither document gives a valuation, and no valuation figure has appeared in coverage of the round that this publication could find.

The company was founded by Arif Mirza (chief executive), Pål Rasmussen (chief technology officer) and Thorvald Thorsnes (chief commercial officer). The release says Mirza and Rasmussen “spent much of the past decade in leadership roles at Spacemaker and Autodesk” — Spacemaker being the Oslo generative-design company Autodesk announced in November 2020 that it was acquiring for approximately $240 million, net of cash — and that Thorsnes previously co-founded the Y Combinator-backed energy infrastructure company Enode.

Solvera says it began operations in July 2026, passed $1 million in contracted revenue in its first three months, and has grown to a team of nine. The new money goes to building out the engineering and technology team in Oslo, opening a London office, and further development of the company’s operating system.

The Model Is the Story

The release states the arrangement without hedging: “Solvera delivers the full end-to-end service as the engineering firm of record, with clients paying for the outcome, not software licenses or consulting hours.”

Mirza’s explanation of why that requires a new firm rather than a new product is the clearest articulation of the thesis:

“We’ve had twenty years of better building engineering software, and now AI, but productivity and the fundamental way engineering gets delivered has barely changed. Technology is only part of the answer. To capture its full potential, you also have to rethink how the firm operates and how it makes money. That is incredibly difficult to do inside organizations built around thousands of people, active projects and billable hours.”

In Seedcamp’s write-up, Mirza puts a number on the gap he is describing: he “estimates that 85% of tasks in architecture and engineering could be done at least twice as fast, yet AI is used in fewer than 5% of them.” Those are his estimates, presented as such in the investor’s post, and no methodology accompanies them.

Tom Wilson, a partner at Seedcamp, frames the bet in terms of what it is not:

“We invested in Solvera because Arif, Thorvald and Pål are not just building software for engineers, they are rebuilding the firm itself. That is a much bigger swing, and a much better one.”

Why “Firm of Record” Is Load-Bearing in Norway

“Engineering firm of record” could be marketing anywhere. In Norway it maps onto a specific statutory object, and that is worth spelling out because it constrains what Solvera can automate.

Norwegian building rules assign responsibility for design to a firm, not an individual. The building regulator’s guidance to § 12-3 of the building application regulations, on the responsibilities of the ansvarlig prosjekterende — the responsible designer — opens: “All prosjektering skal gjøres av ansvarlig prosjekterende foretak” (“all design shall be done by a responsible designing firm”). The regulation’s own text sets out what that firm is answerable for; the guidance explains the principle. That firm must quality-assure and document that the technical building regulations are satisfied, produce the site plan, and coordinate interfaces with the other designing firms on the project. A firm takes that on for a given project by declaring ansvarsrett, and the regulator’s guidance defines the declaration as meaning “at foretaket tar på seg ansvar som ansvarlig i det konkrete byggetiltaket fordi det oppfyller krav til organisasjon, kvalitetssikringsrutiner og faglige kvalifikasjoner” — that the firm accepts responsibility on the specific project because it meets requirements on organisation, quality-assurance routines and professional qualifications.

Solvera’s Norwegian-language site states that it holds ansvarsrett and that “hver leveranse kontrolleres og godkjennes av en erfaren rådgivende ingeniør” — every deliverable is checked and approved by an experienced consulting engineer. It names the engineers who do it, listing a consulting engineer for ventilation, plumbing and sprinkler systems and another for electrical.

The analysis here is this publication’s own. The familiar model in AI for design is to sell a tool and leave the liability where it was, with the licensed firm that pressed the button. Solvera is doing the opposite: it is accepting the statutory position and then automating inside it. That makes the human-review step a legal requirement rather than a reassurance in a FAQ — and it also caps how far the automation can run, because the thing being automated is work for which a named firm has declared responsibility to a municipality. Whether that is a moat or a ceiling is the open question in the business, and three months of trading does not answer it.

What It Is Actually Doing

Solvera starts with mechanical, electrical and plumbing engineering, which the release characterises as work where engineers “must navigate hundreds of requirements and constraints, work across dozens of legacy tools, size and route systems, coordinate amongst themselves and across disciplines, and continuously adapt designs as projects evolve.”

The platform is called SolveraOS. The company’s site describes agents that take the repetitive work — sizing air volumes, routing ducts, updating electrical circuits — and redo it when the project changes, with simulation of energy, air volumes and loads across design alternatives. Rasmussen’s statement of the division of labour:

“The most valuable thing in building engineering has always been the engineer’s judgement, yet much of their day goes to sorting a constant stream of information and changes, manual data entry, and coordination. Our goal is to free up their time so they can focus on high impact decisions, with our agents handling the rest like testing alternatives, coordination, and production work.”

The release says the company is engaged on several Norwegian building projects spanning early-phase engineering through detailed design, “including commercial projects with Betonmast, one of Norway’s leading contractors.” That customer is confirmed by the customer: a named executive at a named contractor, quoted on the record. Kristian Gjestemoen, managing director of Betonmast Buskerud-Vestfold, is quoted in the release:

“As a contractor, we see the consequences of fragmented and inefficient engineering every day: unnecessary cost, slow iterations and problems that are discovered too late. We don’t believe adding more hours will solve this, so we want partners who are willing to challenge established ways of working, move faster, and take responsibility for better outcomes. That is why we chose Solvera for several of our projects.”

A Norwegian version of the same quote appears on Solvera’s own site under Gjestemoen’s name. Betonmast describes itself as one of Norway’s largest building contractors; the boilerplate in Solvera’s release puts it at around 500 employees and annual revenues of approximately NOK 4.1 billion, and part of AF Gruppen.

Note what the $1 million figure is and is not. The release says “contracted revenue” — work signed, not work delivered and recognised — and the company has existed for about a quarter. It is a real commercial signal from a nine-person firm and it is also a very small sample of the model’s economics.

Where This Sits

Two directions have been visible in AEC engineering AI this year, and Solvera is a third.

The first is selling the software to the firms: Endra raised a $50 million Series A led by Andreessen Horowitz to automate MEP engineering, the same discipline Solvera is starting with, on the conventional premise that the engineering firms remain the customer. The second is the incumbent services firm buying the capability: AECOM paid $390 million for the Norwegian AI startup Consigli, and Accenture built a capital-projects AI division to sell advisory and delivery around the same shift.

Solvera’s position is that neither of those changes the unit economics of engineering, because both leave the billing model intact. Wilson’s version: “Construction has absorbed two decades of tooling without shifting how the work actually gets delivered or paid for.” Whether a firm whose revenue is decoupled from hours can beat one whose revenue is hours is, in principle, testable — and MEP is a reasonable place to run the test, because the work is rule-bound enough to automate and specific enough to price.

The scepticism to hold onto is about scale rather than concept. Nine people and three months of trading is a prototype of a firm, not a firm; the London office is announced rather than open; and the comparison that matters — margin per project against a conventional consultancy doing the same scope — cannot be made from outside, and will not be available until Solvera has delivered enough completed projects for anyone to compute it. The company is not claiming otherwise. It is claiming the model is worth building, and $6.3 million is roughly the cost of finding out.