Torsten Reil on Building Helsing and Europe’s Defense Technology

Harry Stebbings asks Torsten Reil what he carried from building NaturalMotion to co-founding Helsing. NaturalMotion began as an Oxford spinout simulating creatures controlled by neural networks. Its technology was licensed for films and games, including Grand Theft Auto, but revenue grew more slowly than Reil wanted.

The company took a risk and pivoted from licensing technology to making games itself. Reil says it underestimated the difficulty, but CSR Racing became a major iPhone game and the business took off. He brought the willingness to focus, recognize when a strategy is not working and adapt quickly to Helsing.

Reil also learned how much attention hiring requires. At NaturalMotion, adding people to accelerate work sometimes increased complexity instead, and the company made large changes, including layoffs, to function again. He and co-founders Gundbert and Niklas committed to recruiting exceptional people from the beginning at Helsing.

Helsing had about 350 people at the time of the interview. Stebbings asks whether a growing company must eventually settle for merely good hires. Reil thinks there are enough exceptional people for a company of that size and that top performers can keep teams faster and lighter, though he acknowledges larger scales might require compromises.

In interviews, Reil looks for a sense of effortlessness and mastery of detail, often within 30 minutes. Stebbings looks for exceptional achievement in any field; Reil prefers an unscripted conversation to a fixed list of questions. He says Helsing aims for talent density and would prefer four exceptional people to eight good ones, even at above-market pay.

Rather than running many small bets, Helsing has a handful of small teams assigned to areas it knows must exist. People can partly self-select for teams that communicate well, and teams may move on as others take over. Reil says performance management demands difficult conversations, an early commitment and willingness to help people improve or part amicably.

Reil no longer schedules regular one-to-ones with senior colleagues, though he acknowledges they may work for people who structure them well or need development guidance. He says recurring meetings can crowd calendars. At Helsing, leaders also periodically review and calibrate performance ratings, correcting both excessive harshness and excessive leniency.

A hiring mistake might involve cultural fit, willingness to go further, or technical ability on a new problem. Reil says it is important to notice problems early and work with the person; he has seen performance-improvement plans succeed, so he does not consider the outcome predetermined.

Helsing borrowed the idea of speed from Zynga. Founded in 2021, the company felt geopolitical conditions were changing and raised money early to move quickly. Reil describes accepting temporary inefficiency to build recruiting capacity, then moving from an outside provider to an internal team.

Stebbings warns that founders can use blitzscaling as an excuse for unnecessary spending. Reil says the inefficient phase should be explicitly temporary: Helsing planned six to twelve months. It raised an €8.5 million seed round, followed roughly four months later by €100 million from Daniel Ek’s Prima Materia, according to Reil.

That financing let Helsing recruit early and pursue large official defense programs instead of relying on small innovation grants, Reil says. He believes the company could only have been co-founded: defense combines government relations, security, export law and company building, while the three founders also act as sounding boards for difficult decisions.

Reil identifies Russia’s 2014 annexation of Crimea and what he regards as an inadequate Western response as a turning point. He says Helsing’s March 2021 presentation anticipated two potential galvanizing events: a Russian invasion of Ukraine and a Chinese invasion of Taiwan. The Ukraine invasion came in 2022, earlier than the 2025 timing he says they had expected; the Taiwan scenario remains a forecast, not a fact.

Reil argues European societies did not fully confront the consequences of the Crimea annexation because thinking about conflict and suffering was uncomfortable. He advocates higher defense spending even at the expense of other priorities, while acknowledging budget pressure and the difficulty of persuading voters. His spending figures are approximate estimates offered in the interview.

Asked about a potential Trump administration, he says Europe had already been moving toward greater defense spending and sovereignty. He expects that direction could accelerate as Europe becomes less able to rely indefinitely on partners. He does not claim to know whether Trump will withdraw support for Ukraine.

Reil is unsure whether to call the expected increase a defense-spending boom: exposed countries may increase budgets more rapidly than others, and some additional money will repair existing deficiencies rather than buy new technology. His company stands to benefit from investment in new defense equipment.

Helsing says its decisions about customers go beyond export law. Reil describes employee ethics workshops that examine technology use and potential sales, with the founders making final decisions. He says the process is intended to involve people with different views, although not everyone necessarily agrees with the outcome.

He says the Economist democracy index was an initial reference: countries at either end are easier to assess, while countries in the middle and their alliances present harder judgments. Stebbings raises the changing debate over Israel and Gaza as an example of how a customer’s status may be contested. Reil says conflicts involve difficult trade-offs and a company must develop a view beyond static export rules.

The company’s recently announced strike drones raised further questions about technology and use, Reil says, without detailing all the issues. Asked whether revenue opportunities could compete with European sovereignty, he states that Helsing’s purpose is to make Europe safer and that scale and revenue serve that goal. Those are the defense company founder’s own account of its priorities.

Helsing remains software-first, Reil says, citing its work with Saab on a Eurofighter electronic-warfare upgrade. He claims software can substantially multiply capability. He argues drones require hardware designed for large-scale production: in his view, many existing Western products are built in small numbers and are too expensive for the quantities he envisages.

Helsing announced its HX-2 drone after deciding it could move complexity into software and simplify hardware, Reil says. Hardware still requires expertise in components and supply chains, but the company uses simulation and software testing to iterate before physical tests. He cautions that real deployment, including in Ukraine, can expose conditions a simulation missed.

Stebbings asks whether compliance inhibits government purchases of single-use drones. Reil agrees it can be an obstacle but says procurement rules often exist for good reasons. He argues the larger system can favor established relationships even though individual officials and politicians want innovation.

His proposal is to reserve 20% of defense equipment spending for new defense companies, excluding maintenance from that denominator. He estimates their current share at less than 1%, without supplying a source. Stebbings points out the proposal could favor Helsing; Reil later acknowledges it sounds self-serving, and his company has a direct financial interest.

Reil says established defense primes remain important, but warns of a potential “Tesla moment” if they do not adapt to autonomy, software and AI. Saab is a close partner and a Helsing shareholder, he says. He wants established companies and newer entrants to work together rather than assuming primes will disappear.

He points to Iranian and Russian drone development as evidence of a move toward larger numbers of autonomous systems. His view is that Europe must connect its systems and develop its own capacity at scale to deter adversaries. These are his strategic claims, not independently verified performance comparisons.

Discussing Ukraine, Reil estimates that Ukraine produces 1.5 million first-person-view drones a year and Russia roughly as many. He expects operators soon to control multiple drones rather than one each. Both the production estimates and the timing of that transition are his assertions from the interview.

Reil defines drone success partly by whether it can hit a target despite electronic interference. He describes communications and navigation jamming as major constraints and argues that lower-cost drones can threaten expensive tanks and artillery. His examples of numbers required and the effects of saturation are illustrative judgments, not universal results.

He predicts future battlefields will become increasingly dangerous for people and ultimately more contactless, while recognizing that conflict still concerns physical places such as ports and routes. Stebbings says autonomous swarms are frightening; Reil agrees they should concern everyone and argues societies must decide how to retain safety and ethical standards.

Reil believes public support for defense spending has increased but says people may disagree once concrete trade-offs are presented. As a hypothetical defense minister he repeats the 20% proposal. He also argues that primes have insufficient incentives for research and development, whereas Helsing pre-finances development with capital raised from investors.

He expects a future market with a small number of primes and many suppliers. Reil cautions venture investors that funding a defense technology team does not guarantee an exit: acquisition prices may not support venture valuations, and becoming a public prime requires much more than promising technology. His view favors larger, established programs and is informed by his own company’s position.

A program of record is an official, often multiyear government program, he explains, contrasting it with a small innovation project. A startup that earns, for example, €500,000 on an innovation project may still lack the government relations, compliance, security and sales infrastructure needed to compete for a major contract.

The talent needed for those functions is scarce, Reil says, making it unlikely that every startup can become a prime. He compares product-market fit to only a fraction of the work. Helsing and Anduril are building the rest of that infrastructure, he says, while acknowledging that Helsing still has a way to go.

Asked whether naivety helps founders, Reil says he would still have started Helsing knowing how hard it would be, partly because three founders can share the burden. Startup setbacks can arrive when a team is already exhausted; the ability to rely on co-founders matters when the next unexpected problem appears.

At NaturalMotion, Rockstar once told him its engine for Grand Theft Auto IV needed to run ten times faster or be removed. Reil thought the team had exhausted its optimization options. Engineers went to Edinburgh, worked with Rockstar and found the performance improvement; the experience strengthened his respect for exceptional engineers.

Reil argues that Europe’s main startup problem is ambition rather than early-stage capital. He says more founders should aim to build large companies or return for a second or third venture, rather than sell at €100 million and retire. This is his judgment about Europe, not a universal assessment of founders’ choices.

He says Europe has strong engineering talent and that Helsing recruits almost exclusively in Europe. US investors can still add growth capital and useful ambition: at NaturalMotion, he credits Benchmark partner Mitch Lasky with encouraging him to think bigger. Helsing began with European investors, later added US capital, and remained Europe-dominated in ownership, he says.

Stebbings worries Europe is becoming less competitive through dependence on subsidies and public employment. Reil agrees there is a negative trend but argues founders and political participants have agency to change it. He thinks founder quality and ambition matter more than discussions of tax and regulation, a debatable assessment rather than a settled causal finding.

Asked where Europe needs more sovereignty, Reil names energy, including future technologies such as fusion, and semiconductors as well as defense. He doubts sufficient capacity can be built in two or three years but believes it can be achieved over five to ten years. That is his forecast, not a commitment or established outcome.

On Russia and Ukraine, Reil says the future is hard to predict but calls a threat to NATO’s eastern flank plausible within a few years. On China and Taiwan, he considers conflict plausible and, if it happens, potentially less than ten years away, while stressing uncertainty. These are geopolitical judgments made in December 2024, not verified predictions.

Reil says people are reluctant to contemplate conflict and its suffering, himself included. Money, he adds, gave him security after NaturalMotion and made it easier to take risks with another company; it is helpful but not sufficient for success. A long daily walk is one small thing that makes him happy.

In the quick-fire round, he repeats his view that attracting strong founders matters more than taxes or regulation. He says a fast-growing company must repeatedly revisit its structure, and quarterly founder offsites give Helsing space to discuss what works and what needs to change.

Reil disputes the idea that every decision is better than no decision. Consciously waiting for necessary information can itself be a sound decision, he says. He also rejects “work smart, not hard,” arguing that both matter, and says Helsing’s founders provide the vision and direction for the company.

He has changed his mind about how much process exceptional teams need at the start of a rapidly changing project. In his experience they can sometimes make progress without standups or deadline predictions, though structure becomes necessary later.

Asked which company he would run for a day, Reil says he would shut down TikTok, calling its engagement mechanisms dangerous. Stebbings disagrees, describing how he chooses useful fitness and food content. Reil argues the platform can encourage excessive use, political manipulation and extreme content, while acknowledging he does not know whether manipulation is occurring.

On X under Elon Musk, Reil says the experience feels different but reserves judgment on whether it is better or worth investing in at the quoted valuation. These are personal impressions, not a study of platform effects.

Looking back, Reil wishes Helsing had known how soon Russia would invade Ukraine. He says the company would have concentrated earlier on systems addressing the immediate problems, instead of also building other systems and a technology foundation. He believes Helsing eventually contributed, but needed time to develop its capabilities.

Finally, Reil reflects on how an equal founder group works despite ambition and ego. His answer is mutual respect and appreciation of each co-founder’s contribution. Their emphases differ and roles shift, but all three participate in managing and scaling the company.