Fractile backer @zebulgar said on November 11, 2026, “We’re excited to double down on Fractile’s Series C.” In a separate post that day, @asenkut praised the company’s “latest raise” and said they were doubling down on the round. Source Source
@zebulgar’s chip pitch centers on waiting for model weights, the numerical values a model learns during training. “Right now there's lots of silicon waiting for weights. You can’t FLOPS your way out of that bottleneck,” the backer wrote, arguing that more arithmetic power alone won’t fix the wait. The post describes Fractile as doing “the math where the weights live.” Give the math less of a commute. Source
The backer also forecast that inference, or running a trained model to produce results, would account for a larger share of AI compute over the next few years. The post includes no analysis behind that forecast, technical documentation or chip benchmarks. More enthusiasm is not a performance test. Source
@asenkut also described Fractile’s UK team as expanding into Silicon Valley and cited “super impressive commitments from the largest AI labs.” The post doesn’t name the labs, explain what those commitments involve or give operational details about the expansion. Source
Neither backer gives a funding amount, valuation or closing date, or explicitly confirms that a Series C has closed. @asenkut doesn’t label the raise a Series C, so whether the posts refer to the same financing remains unconfirmed. Source Source
Comments