Gimlet raises $80M
Gimlet Labs closed an $80M Series A to build a multi‑silicon inference cloud that runs across NVIDIA, AMD, Intel, ARM, Cerebras and d‑Matrix — a bet that inference will be heterogeneous and multi‑vendor. The raise underscores demand for multi‑silicon inference orchestration. (techcrunch.com)
Menlo Ventures led Gimlet’s Series A, with participation from Eclipse Ventures, Factory, Prosperity7 and Triatomic. (gimletlabs.ai) The company’s founding team is led by Zain Asgar (co‑founder and CEO) alongside Michelle Nguyen, Omid Azizi and Natalie Serrino. (crunchbase.com) Asgar previously founded Pixie (acquired by New Relic) and serves as a Stanford adjunct, credentials Menlo’s partner highlighted in its investment note. (menlovc.com) Gimlet emerged from stealth in October 2025 claiming eight‑figure revenues at launch and says its customer base has rapidly expanded to include a “top frontier lab” and a hyperscaler. (finance.yahoo.com) The product, branded Gimlet Cloud, orchestrates and “slices” agentic inference across diverse silicon — citing integrations with NVIDIA, AMD, Intel, ARM, Cerebras and d‑Matrix — and the company reports 3–10× speedups on >1T‑parameter models. (techcrunch.com) Gimlet has announced a technical partnership to deploy d‑Matrix Corsair accelerators inside its cloud to target memory‑bound phases of inference and claims order‑of‑magnitude gains in latency and throughput per watt versus GPU‑only stacks. (marketwatch.com) The company says new capital will fund expanded engineering hires and build‑out of multi‑silicon data‑center infrastructure that connects heterogeneous hardware with low‑latency interconnects. (gimletlabs.ai) Menlo framed the investment as a bet on software that decouples agentic workloads from single‑vendor hardware, arguing that heterogeneous orchestration is necessary as inference scales and silicon supply strains grow. (menlovc.com)
