Ebenbuild: EIC Accelerator Bet on Digital Twins of the Lung

For an inhaled medicine, the dose that matters is the one that reaches the diseased part of the lung, and that is the one nobody can easily measure. It depends on the shape of each patient's airways, on disease, on how the patient breathes and on the aerosol itself. Ebenbuild, a Munich company spun out of the Technical University of Munich, builds a simulated copy of a patient's lungs from a standard CT scan and tracks each aerosol particle through the airways and into the alveoli. Its product for drug and inhaler developers is called Twinhale.

The EIC Accelerator selected Twinhale at its March 2025 cut-off for blended finance. CORDIS records an EU contribution of €2,330,125 for a project running from 1 October 2025 to 30 September 2027, with the agreement signed on 26 September 2025. A year into the project, the model behind the product has passed peer review. Who checked it, and on how much data, is not stated in the company's announcement of the paper.

A small Munich software company, still active

The Munich commercial register, as shown by North Data, records Ebenbuild GmbH as an active company with three managing directors and a corporate purpose of developing and marketing medical software. It filed its annual report for 2025 on 30 July 2026, and no insolvency or liquidation entry appears. The register lists balance sheets for 2023 to 2025, but North Data does not display their figures for free, and no revenue or loss figure has been published by the company.

The team is still small. A Startbase report of February 2022, citing Handelsblatt, put the staff at eight, and the company's team page names 16 people in October 2026, among them a head of quality management and a medical affairs manager.

How the lung model works

Established ways to estimate where an inhaled drug lands are either physical, such as cascade impactors and simplified physical lung models, or imaging studies in which volunteers inhale a radiolabelled aerosol and are scanned. Computer models exist, but the paper says they have so far failed because of fundamental limitations in method. Ebenbuild's model, as described in its Communications Medicine paper, builds the airways from CT data and adds the alveolar region, then simulates airflow, tissue movement and particle transport through a whole breathing cycle. The paper says model generation and calibration are automated and that the model is ready to be applied at scale.

The company's pitch to pharma, set out in its EIC announcement of 2 July 2025, is that Twinhale lets developers test devices, doses and patient groups on screen before trials, cutting failures and cost. The paper's introduction gives the stakes it sees: an average development cost of $115 million for an inhaled therapy and a bench-to-bedside success rate of 16%.

What the validation shows, and who did it

Ebenbuild announced on 15 April 2026 that the study had been published, and its CEO called it "rigorous, independent, clinical validation" of the modelling engine. The paper compares the model's predictions with SPECT/CT deposition images from an earlier clinical study run by others and designed for model validation. Across 10 inhalation experiments in six healthy subjects, the lobe-by-lobe comparison of 50 data points gave a correlation coefficient of 0.95. The authors also simulated one patient with idiopathic pulmonary fibrosis to show the model can represent diseased tissue.

A second study, in the Journal of Applied Physiology in 2025, tested the company's ventilation model against bedside electrical impedance tomography in patients with acute respiratory distress syndrome (ARDS); its authors come from Ebenbuild, the Technical University of Munich and University Hospital Schleswig-Holstein. That study concerns the intensive care product, not Twinhale.

From the intensive care unit to drug developers

Ebenbuild did not start as a pharma supplier. When it raised its seed round in March 2022, High-Tech Gründerfonds and Munich Startup described a product for doctors ventilating ARDS patients, with the money meant to bring the prototype to market readiness and approval as a medical device. Startbase reported in February 2022, citing Handelsblatt, that the round was meant to be enough for a revenue-generating product within two years at the latest.

That ventilation product, now called Aerogram, was still described as "in development" in the company's April 2026 release. Public funding for the ventilation work has continued: North Data's listing of federal funding shows €299,040 for the BEATE ventilation project from August 2020, €871,280 as Ebenbuild's share of the UBIC consortium on digital lung twins for acute respiratory failure from August 2023, and €583,999 for a follow-up, BEATE2, from August 2026. The company's own news page gives €1.8 million for the whole UBIC consortium with five university hospitals. Twinhale, the drug delivery product, is the one the EIC chose to fund, and the one the company presents as its market launch.

Pilots, not contracts

The July 2025 release and Munich Startup both say Twinhale has been used in pilot studies with PARI Pharma and Pieris Pharmaceuticals. The Pieris work is public: Ebenbuild staff are co-authors of a Nature Communications paper of April 2025 on PRS-220, an inhaled protein for pulmonary fibrosis, and the company's release on it says Twinhale simulated deposition for two patient cases taken from the CT database of the Open Source Imaging Consortium. No commercial terms were disclosed for either pilot, and no other customer has been announced.

The market figures in the EIC release are the company's and are not consistent with each other: it describes a "Euro >5 billion in silico trials market" in the text and a "Euro 15 billion" market in a bullet point of the same release.

1. Validated whole-lung deposition modelPublishedPeer-reviewed in April 2026; analysis by company authors against outside data from six healthy subjects.
2. Diseased lungsShown in examplesOne IPF patient in the validation paper; two imaging-database cases in the PRS-220 study.
3. Twinhale sold to biotech, pharma and medtechNot shownTwo pilots named; no contract or revenue disclosed.
4. EIC project resultsToo earlyCORDIS shows no report; the project runs to September 2027.

The money

The only equity round Ebenbuild has announced is its €2.5 million seed in March 2022, led by High-Tech Gründerfonds and Bayern Kapital. Since then, the disclosed money has been grants. The EIC release says the EIC Fund may invest up to €10 million in future equity rounds; the company had announced no such round by October 2026, and the EIC Fund's investment page does not name Ebenbuild. TUM's chair of computational mechanics, from which the company spun out, noted that Ebenbuild was one of 40 companies selected from 959 applicants.

Why the EU would fund it

EU law on animal experiments, Directive 2010/63/EU, requires that a scientifically satisfactory method not using live animals be used wherever possible. The validation paper makes a different claim of its own: that the model could become an alternative to nuclear imaging of aerosol deposition and reduce the need for radiation-based imaging in clinical studies. Whether Twinhale has replaced any animal or human study in a customer's programme is not public. The project's own case, in the CORDIS objective, is that the respiratory system has been "a sort of black box" in research and clinical practice, and that patient-specific lung models can open it.

The reporting is based on publicly available sources, listed below.