Reflection AI has signed a more than $1 billion deal with European cloud provider Nebius to secure computing capacity, including access to Nvidia's latest GB300 chips, through 2029 Source: thenextweb, cna. This is the startup's second major compute deal in weeks, following a multibillion-dollar agreement with SpaceX for the same class of Nvidia chips, reportedly worth about $150 million per month through 2029 Source: cna.
AI startups are racing to lock in hardware as demand growth from businesses outpaces new data-center supply, driving up chip prices Source: thenextweb. The need for capacity has become a constraint for many firms. Nebius, a neocloud spun off from Yandex and now listed on Nasdaq, has been a key player, backed by Nvidia with investments totaling over $2.7 billion Source: thenextweb. It also holds large deals with Meta (up to $27 billion) and Microsoft (up to $19.4 billion) Source: techcrunch.
Founded in 2024 by two former Google DeepMind researchers, Reflection develops open-source AI models as a cheaper, more customizable alternative to closed systems from OpenAI and Anthropic Source: techcrunch, cna. The startup is valued at about $8 billion and has raised close to $2.6 billion from investors including Nvidia, Sequoia Capital, and Lightspeed Venture Partners Source: thenextweb. It is reportedly in talks to raise another $2.5 billion at a $25 billion valuation.
Interest in open-source models has surged recently. The Trump administration pressured Anthropic and OpenAI to restrict their most powerful models, highlighting the risk of relying on providers that could be cut off overnight Source: techcrunch. Reflection's co-founder and CTO Ioannis Antonoglou stated, "The need for open models is clear, and this additional compute capacity will allow Reflection to continue to build and train frontier AI models at scale" Source: thenextweb, cna.
“Reflection AI has locked in more computing power. The US startup signed an AI compute deal worth more than $1bn with Nebius, the Amsterdam-based cloud provider, according to Bloomberg.”
“AI startups are racing to lock in the computing power needed to train and run their models as demand growth from businesses adopting the technology outpaces new data-center supply.”
“Just last month, the Trump administration pressured Anthropic and OpenAI to restrict their most powerful new models, raising concerns that access to AI models could be taken away overnight.”
“The need for open models is clear, and this additional compute capacity will allow Reflection to continue to build and train frontier AI models at scale.”
“Reflection, launched by two former Google DeepMind researchers, develops open-source models that serve as an alternative to the offerings from OpenAI and Anthropic.”