Blab AI
Vienna Startup Ora Computing Raises 3.5 Million Seed Round to Accelerate AI Model Compression
When Vienna’s Ora Computing closed its €3.5 million seed round on June 24, 2026, it signaled a shift toward a new breed of AI—compact, efficient, and ready for the edge.
The round, led by Constructor Capital and Greencode Ventures, also drew continued support from XISTA Science Ventures, the firm that helped launch Ora. The capital will be earmarked for expanding the team, extending the compression technology to the biggest frontier models, and rolling out a commercial product aimed at cloud inference providers and enterprises that deploy AI.
"We were created to challenge the idea that massive scale is required for useful intelligence," CEO and co‑founder Stefan Sack told reporters. "The next wave of AI adoption will be driven by more compact models that are highly efficient and optimised for specific use cases rather than large general‑purpose cloud models."
Ora’s seed round sits on the lower end of 2026’s European AI funding landscape. Larger deals that year went to compute capacity and data‑centre infrastructure—Mistral AI, Nscale and Verda, for example, raised significant sums—while smaller seed and pre‑seed rounds focused on software layers such as AI memory, agent governance, licensed data access and compression.
The company’s focus on reducing model size and inference cost dovetails with a broader trend: capital increasingly flows into technologies that make AI cheaper, more deployable and more efficient to run. "AI’s energy appetite is growing faster than the world can build the infrastructure to feed it," said Terhi Vapola, founder and managing partner of Greencode Ventures. "One key approach is to make AI itself more efficient, and that is exactly what Ora does."
Founded in 2024 by Sack and Raimel Medina—both quantum‑computing researchers from the Serbyn group at the Institute of Science and Technology Austria (ISTA)—Ora claims its compression stack can shrink the memory footprint of large AI models by up to 80 % and accelerate their execution by up to four times. The technology works across different hardware types and can be dropped into standard inference frameworks without custom software layers, retraining or changes to existing infrastructure.
According to the company, AI inference—the process of running a model to generate outputs—has become a significant and fast‑growing cost for any organisation deploying AI at scale. Large deployments can now cost tens of millions of euros per month in compute alone, and the cost rises as models grow.
For organisations that want to run AI locally on devices such as cars or industrial equipment, the models are often too large to fit. By compressing models, Ora says the required compute power is reduced, which in turn lowers energy consumption and carbon emissions. At a 1 % market penetration, the company estimates its technology could eliminate more than 50,000 tonnes of CO₂ annually.
Unlike some compression tools that force a binary choice between compression levels, Ora’s algorithm continuously maps the trade‑off between model size and accuracy, allowing companies to optimise for their specific hardware and cost constraints.
The startup has tested its approach on a 70‑billion‑parameter model, compressing it in hours at a compute cost of under $1,000. Industry figures for comparable work are in the hundreds of thousands of dollars.
Ora’s announcement comes as cloud providers and enterprises look for ways to reduce inference costs while maintaining performance. The company’s commercial product will target both cloud inference providers—who can offer smaller, cheaper models to their customers—and enterprises that need to run AI workloads on edge hardware or in data centres.
The seed round gives Ora the resources to accelerate product development, expand its engineering team, and pursue partnerships with cloud and edge hardware vendors. The company has not yet announced a specific launch date for its commercial offering.
In summary, Ora Computing’s €3.5 million seed round reflects a growing investment focus on AI efficiency layers. The startup’s compression technology promises significant reductions in memory usage, inference speed, compute cost, and carbon footprint, positioning it to serve both cloud and edge AI markets as the demand for large foundation models continues to rise.
The round, led by Constructor Capital and Greencode Ventures, also drew continued support from XISTA Science Ventures, the firm that helped launch Ora. The capital will be earmarked for expanding the team, extending the compression technology to the biggest frontier models, and rolling out a commercial product aimed at cloud inference providers and enterprises that deploy AI.
"We were created to challenge the idea that massive scale is required for useful intelligence," CEO and co‑founder Stefan Sack told reporters. "The next wave of AI adoption will be driven by more compact models that are highly efficient and optimised for specific use cases rather than large general‑purpose cloud models."
Ora’s seed round sits on the lower end of 2026’s European AI funding landscape. Larger deals that year went to compute capacity and data‑centre infrastructure—Mistral AI, Nscale and Verda, for example, raised significant sums—while smaller seed and pre‑seed rounds focused on software layers such as AI memory, agent governance, licensed data access and compression.
The company’s focus on reducing model size and inference cost dovetails with a broader trend: capital increasingly flows into technologies that make AI cheaper, more deployable and more efficient to run. "AI’s energy appetite is growing faster than the world can build the infrastructure to feed it," said Terhi Vapola, founder and managing partner of Greencode Ventures. "One key approach is to make AI itself more efficient, and that is exactly what Ora does."
Founded in 2024 by Sack and Raimel Medina—both quantum‑computing researchers from the Serbyn group at the Institute of Science and Technology Austria (ISTA)—Ora claims its compression stack can shrink the memory footprint of large AI models by up to 80 % and accelerate their execution by up to four times. The technology works across different hardware types and can be dropped into standard inference frameworks without custom software layers, retraining or changes to existing infrastructure.
According to the company, AI inference—the process of running a model to generate outputs—has become a significant and fast‑growing cost for any organisation deploying AI at scale. Large deployments can now cost tens of millions of euros per month in compute alone, and the cost rises as models grow.
For organisations that want to run AI locally on devices such as cars or industrial equipment, the models are often too large to fit. By compressing models, Ora says the required compute power is reduced, which in turn lowers energy consumption and carbon emissions. At a 1 % market penetration, the company estimates its technology could eliminate more than 50,000 tonnes of CO₂ annually.
Unlike some compression tools that force a binary choice between compression levels, Ora’s algorithm continuously maps the trade‑off between model size and accuracy, allowing companies to optimise for their specific hardware and cost constraints.
The startup has tested its approach on a 70‑billion‑parameter model, compressing it in hours at a compute cost of under $1,000. Industry figures for comparable work are in the hundreds of thousands of dollars.
Ora’s announcement comes as cloud providers and enterprises look for ways to reduce inference costs while maintaining performance. The company’s commercial product will target both cloud inference providers—who can offer smaller, cheaper models to their customers—and enterprises that need to run AI workloads on edge hardware or in data centres.
The seed round gives Ora the resources to accelerate product development, expand its engineering team, and pursue partnerships with cloud and edge hardware vendors. The company has not yet announced a specific launch date for its commercial offering.
In summary, Ora Computing’s €3.5 million seed round reflects a growing investment focus on AI efficiency layers. The startup’s compression technology promises significant reductions in memory usage, inference speed, compute cost, and carbon footprint, positioning it to serve both cloud and edge AI markets as the demand for large foundation models continues to rise.
