Why Greencode invested in Ora Computing

^{Image credit: Ausstellung Im Erweiterten Raum, aus der Sammlung STRABAG ART, Künstlerhaus Wien, 2026}

In June, we announced our investment in Ora Computing, an Austrian startup that compresses AI foundation models. Co-leading their €3.5 million seed round, we are excited to back Ora's mission to become the efficiency layer of the AI stack.

"AI's energy appetite is growing faster than the world can build the infrastructure to feed it. One key approach is to make AI itself more efficient, and that is exactly what Ora does. Compressing models radically without sacrificing accuracy makes a tremendous difference to their customers," comments Terhi Vapola, Founder and Managing Partner of Greencode Ventures.

‍

About Ora Computing

Ora's software shrinks AI foundation models by up to 80% and makes them run up to four times faster, while keeping accuracy loss between 0 and 5%. Compressed models need far less compute, memory, and energy to serve the same workload. This lowers costs for anyone running AI at scale in the cloud, and it makes capable models small enough to run locally on cars, industrial equipment, and other edge devices where today's models simply do not fit.

Their solution works across different hardware types and drops directly into standard inference frameworks, which means there is no need for custom software layers and capital-intensive retraining, and no changes to existing infrastructure. Where competing approaches force a binary choice between fixed compression levels, Ora's algorithm continuously maps the full tradeoff between model size and accuracy.

Ora has demonstrated this by compressing a 70-billion-parameter model in hours at a compute cost of under $1,000, against industry figures of hundreds of thousands of dollars for comparable work. The company has validated its solution with players in the automotive and edge silicon sectors, and its compressed models have been downloaded more than 15,000 times on Hugging Face.

‍

The opportunity

Ora focuses on the challenge of inference – running an AI model to get results – which has become one of the fastest-growing costs for businesses. It is estimated that the global AI inference market could reach $254.98 billion by 2030, growing at 19.2% each year.

AI models are now so large that running them in the cloud is expensive, whereas using them directly on devices such as cars or industrial equipment is often not feasible. This is where edge AI comes in: running models locally, on the device itself. It is one of the fastest-growing corners of the AI market. The global edge AI market is projected to climb from around $25 billion in 2025 to over $118 billion by 2033, growing roughly 22% a year. As AI shifts from data centres onto phones, vehicles, and industrial hardware, the ability to fit a capable model onto that hardware becomes a gating constraint for the whole market. Ora's compression is what clears it.

‍

The team

Ora was founded by two former quantum physicists, Stefan Sack (CEO) and Raimel Medina (CTO). They left quantum computing research to focus on building Ora on the conviction that the next wave of AI adoption will be driven by more compact models that are highly efficient and optimized for specific use cases rather than large general purpose cloud models.

Ora’s team has deep technical expertise forged at top-tier research institutions like the Institute of Science and Technology Austria (ISTA), Harvard and ETH Zurich, as well as industry players like Apple, IBM and Pasqal. Technical depth of this kind is rare, and it is the profile suited to solving the hard problems at the frontier of AI.

‍

The impact

AI is hungry for energy. Ora addresses the energy intensity of AI itself, sitting at the convergence of digital and green transitions. A compressed model requires proportionally fewer GPU cycles, less memory bandwidth, and less cooling to serve the same inference workload, so every efficiency gain translates into lower energy consumption and reduced carbon emissions. At just 1% market penetration, Ora estimates its technology could eliminate more than 50,000 tonnes of CO₂ annually.

The impact runs through two channels. In the cloud, compressed models let data centers serve more inference per unit of electricity, lowering the energy intensity of AI directly. At the edge, models become small enough to run on local devices, eliminating the data center energy cost associated with those queries altogether.

‍

What’s next

The recent seed funding will grow Ora's team, extend its compression to the largest frontier models, and support the launch of a commercial product for cloud inference providers and companies deploying AI at the edge. The company has come out of stealth with a validated technology and strong early traction, and we are proud to back Stefan, Raimel, and the team as they build the efficiency layer of the AI stack.

‍

Building something transformational? Get in touch: hello@greencode.vc

‍