Summarize this content to 2000 words in 6 paragraphs
Oumi co-founders Manos Koukoumidis, left, and Oussama Elachqar. (Oumi Photo)
A new startup out of Seattle wants to open up the “black box” of foundational AI models in a bid to help spur more innovation for the booming field of artificial intelligence.
Oumi came out of stealth mode this week, announcing a $10 million seed round and an ambitious plan to build an open source AI research and development platform.
The AI sector needs a “Linux moment,” according to Manos Koukoumidis, a former senior engineering manager at Google Cloud who is CEO and co-founder of Oumi.
Much like the way Linux helped shape the modern computing landscape through open source technology, Oumi wants to make it easier for researchers and developers to collaborate and advance AI techniques.
Its platform can run anywhere and supports various AI-related development, from pretraining to inference to evaluation. Users can work with open models and commercial APIs, and also use prebuilt workflows from Oumi.
“There are so many things that an open community can do much better than any closed model provider,” Koukoumidis said.
Existing proprietary models from companies such as OpenAI are “very restrictive,” and there are few true open source options across the AI landscape that are easy to use, according to Koukoumidis.
Oumi plans to generate revenue by offering open-source solutions tailored for enterprises, such as hosted versions of different training and inference workflows.
Open source models are a hot topic in the tech world given the attention on a new model from a Chinese startup called DeepSeek that uses far less computing power but could match the capabilities of AI software from leading U.S. tech companies.
Koukoumidis said the emergence of DeepSeek represents a validation for Oumi’s strategy, in part because it was built on top of other open source efforts such as PyTorch and Meta’s Llama. It’s a sentiment also shared by Meta AI leader Yann LeCuen.
“Open source models, tools and related infrastructure won’t be trying anymore to close the gap — they will be the ones defining the state-of-the-art,” Koukoumidis said. “On a platform like Oumi, the community can co-build the next DeepSeek.”
However, Koukoumidis noted potential roadblocks with a model like DeepSeek.
“Despite its impressive performance in areas like math and coding, DeepSeek’s biases and censorship indicate what is at stake for the future of AI development,” he said. “For the U.S. to continue to lead in AI, we need open source and open collaboration to develop trustworthy and explainable models. U.S. researchers are at a disadvantage if the U.S. is less open than China.”
Oumi is partnering with 15 top AI research universities in the U.S. and the U.K., including the University of Washington, along with other AI researchers.
“We’re building the world’s biggest AI research lab,” said Oumi co-founder Oussama Elachqar.
Koukoumidis helped build generative AI models at Google, and also spent time at Meta and Microsoft. Elachquar was a machine learning engineer at Apple, Twitter, and Microsoft.
Venrock and Obvious Ventures led the seed round. Other backers include Plug & Play, which recently expanded to the Seattle area, along with Seattle-based firm Ascend.
“There hasn’t been enough focus on open experimentation to collectively push the boundaries of what AI can do,” Ganesh Srinivasan, partner at Venrock, said in a statement. “With an unconditionally open source platform, we won’t cap our talent, we will foster experimentation, and society will reap the benefits.”
Oumi is a public benefit corporation, which describes for-profit companies that aim to make a positive impact on society.
