Gimlet Labs Raises $300M Series B at $3 Billion Valuation
Gimlet Labs has raised $300 million in a Series B round that values the AI inference infrastructure startup at $3 billion. Andreessen Horowitz led the round, with Arm Holdings, Microsoft's M12, and Samsung Ventures also participating. Total funding now stands at $392 million.
The company is not developing another AI chip. Instead, it has built software and infrastructure designed to make several different types of processors work together. Its platform currently supports hardware from six semiconductor manufacturers: Nvidia, AMD, Intel, Arm, Cerebras, and d-Matrix.
Gimlet Labs describes its core product as a "multi-silicon inference cloud." The technology breaks an AI inference workload into stages, routing each to a GPU, CPU, or dedicated accelerator based on which hardware is better suited to the task.
The company claims this heterogeneous approach can improve the performance of frontier AI workloads by between three and 10 times under the same cost and power conditions, though independent benchmarks were not provided.
What sets the company's current trajectory apart is its move beyond software orchestration into physical data center infrastructure. Running different classes of silicon inside the same facility presents challenges around power density, networking, and cooling that purpose-built data centers were not necessarily designed to handle.
Newsletter
The corridor, every morning.
Funding rounds and cross-border capital moves, one email, five minutes.
Since March, Gimlet Labs has secured multi-billion-dollar contracts and expanded its managed capacity to several hundred megawatts.
Why It Matters
Gimlet Labs is betting that future AI infrastructure will not belong to a single chip architecture, and it is building at both the software and physical-infrastructure layers to make that case.
The multi-hundred-megawatt buildout signals that large-scale AI deployment is encountering real engineering bottlenecks beyond model training, and solving them at the orchestration and facilities level is becoming a venture-scale opportunity in its own right. There are still unknowns.
Gimlet Labs has not disclosed revenue or customer count, making it difficult to assess the $3 billion valuation against commercial performance. There is also a structural question: chipmakers in its ecosystem could develop their own orchestration capabilities over time. For now, the company's six-manufacturer footprint, fresh capital, and infrastructure scale represent a concrete thesis that AI inference does not have to be tied to one type of processor.
More in Funding
