OODA AI AB (publ) announces that the company has implemented GPT-OSS, OpenAI’s newly released open-weight model, in close collaboration with Phala Network. The solution is the result of OODA AI’s own infrastructure development to enable open, transparent and GDPR-compliant AI computation. It marks a strategic step in the company’s commitment to decentralized AI for business-critical applications.

The model was released by OpenAI on August 5, 2025 and was deployed into production by OODA AI within 24 hours of launch. This demonstrates the company’s ability to rapidly translate new technology into operational solutions.

As part of OODA AI’s decentralized AI infrastructure, GPT-OSS has been integrated into the company’s platform with technical support from Phala Network. By combining OODA AI’s production environment with Phala’s Trusted Execution Environments (TEE), secure AI computation is achieved without exposing user data.

"There is a continuous demand for powerful AI solutions that do not compromise on transparency, integrity or data protection. The combination of GPT-OSS, our own infrastructure and Phala’s technical framework provides exactly that. It is an open and secure foundation for the next generation of AI applications. We are always committed to staying at the forefront of implementing the most advanced AI technologies." says Arli Charles Mujkic, CEO of OODA AI AB.

"OODA AI has demonstrated both speed and vision by being among the first to integrate GPT‑OSS‑120B in a secure, decentralized production environment. This achievement is powered by OODA AI & Phala’s TEE‑based computing Network & Resources, ensuring the deployment is both secure and transparent. By combining OODA’s enterprise‑grade execution with Phala’s privacy‑preserving infrastructure, we’re showcasing the practical power of confidential AI computation. Together, we’re proving that the next generation of AI will be open, verifiable, and built on user‑controlled infrastructure." says Marvin Tong, CEO of Phala Network.

OpenAI’s decision to release open weights for its models confirms the direction OODA AI has long advocated. The future of AI will be open, verifiable and decentralized. This reinforces our conviction that trust and transparency will be fundamental pillars of the next-generation AI ecosystem.

About the implementation:

GPT-OSS-120b is a 117 billion parameter Mixture-of-Experts model with 128 experts per layer, of which 4 are activated per token.
The model is now active in OODA AI’s decentralized production environment, with verified computation via Phala Cloud and TEE-based nodes.
AI computation is performed within a closed, verifiable infrastructure with full data control and without the use of commercial cloud platforms or opaque components.

For further information, please contact:
Arli Charles Mujkic, CEO, OODA AI AB
Email: charlie@ooda.ai
Telephone: +49 176 626 999 64

Måns Pontén Söderlind, Deputy CEO, OODA AI AB
Email: mans@ooda.ai
Telephone: +46 70 766 47 71

About OODA AI AB
OODA AI is a group of companies with a focus on artificial intelligence (AI) and software-as-a-service (SaaS). The group of companies provides AI services in the form of decentralized and distributed AI computing, AI tools as well as AI consulting, training and implementation. The group of companies operates on a global market, the head offices are located in Stockholm (Sweden) and Munich (Germany). Listed on Nasdaq Stockholm First North, also traded on Börse Frankfurt, Börse Stuttgart and Börse München.

For more information, www.ooda.ai

The company's Certified Adviser is Eminova Fondkommission AB | 08-684 211 00 | info@eminova.se