Groq Seeks a Series B at $8 Billion, More Than Doubling Its Valuation in Three Months on a Debt-Free Pitch

Groq is seeking a new Series B at an $8 billion valuation, more than double the $3.15 billion pre-money mark it took in August, according to Bloomberg, which cited investor documents. The target excludes the capital being raised.

Executive chairman Alex Davis has been approaching investors directly, three months after Groq raised $350 million. His pitch, laid out in a letter to investors, is that Groq can produce growing margins without the leverage that listed inference providers depend on. Davis argued that rivals leaning on debt end up acting as enablers for the industry while carrying large unfulfilled obligations, and that the arrangement weakens their ability to build a durable business.

The timing is deliberately contrarian. Shares of CoreWeave and Nebius have been under pressure since the US midterms, hit by worries about the borrowing that funds AI capacity and by the prospect of state-level resistance to data center construction. Elon Musk, whose SpaceX competes in the same market, has predicted that more than 200 neoclouds will die and consolidate. Groq is asking for a higher price into that selloff.

Its own history explains the shape of the business. Groq was founded in 2016 as one of a wave of chipmakers trying to take inference workloads from Nvidia. Late last year Nvidia signed a $20 billion licensing agreement that brought founder and chief executive Jonathan Ross and a group of employees across. What remained pivoted to operating data centers, selling inference capacity with an emphasis on cybersecurity workloads.

The financial projections are aggressive and unaudited: Davis expects revenue to grow eightfold next year and EBITDA to reach $2 billion by 2028. He said the new money would fund growth against short- and medium-term customer contracts rather than long-term take-or-pay commitments, part of the same argument that Groq is carrying less structural risk than its peers. Groq and Davis's investment firm Disruptive declined to comment on the fundraising.

The claim worth testing is the one about capital structure. Debt-light is a real advantage if demand holds, because it removes the refinancing cliff that makes leveraged neoclouds fragile when a customer walks. It is a weaker advantage if the debt-funded rivals are winning the largest training and inference contracts, which is where the scale economics sit. Groq's answer is that its customers are buying capacity in months, not years, and that inference for security operations is a recurring need rather than a speculative build.

Either way, the round is a test of whether private markets will keep marking AI infrastructure up while public markets mark it down. An $8 billion valuation for a company that gave up its chip franchise and its founder would have been hard to defend in October. In the middle of a neocloud rout, it is a bet that the debt-free version of the business is the one that survives the cycle.