Six months from $80M to $300M

The six month gap shows how quickly investors decided Gimlet had moved from an interesting systems startup to a company that could matter at cloud scale. It launched publicly with a seed in October 2025, raised an $80M Series A on March 23, 2026, then closed a $300M Series B on September 4, 2026 at a $3B valuation. That pacing usually means buyers are pulling product into real workloads faster than a normal enterprise sales cycle would allow.1 2 3 4

• Gimlet was not financing a concept in March 2026. At the time of the Series A, the company said it had already emerged from stealth five months earlier and was seeing demand from frontier labs and large AI companies for its inference cloud, which routes AI jobs across different chip types to cut cost and latency.2 5 6
• The investor mix widened sharply in the next round. The Series A was led by Menlo with Eclipse, Factory, Prosperity7, and Triatomic. By September, Andreessen Horowitz led the Series B, with new backers including Arm and M12, while several earlier investors stayed in. That kind of pile on usually signals conviction that the product sits close to core AI infrastructure spend.2 3 4
• The internal financing record also shows multiple Series A and Series B share classes issued around the same dates, with Series A share prices at $1.15, $2.31, and $2.88 in March 2026, and Series B share prices up to $50.28 in late September 2026. That points to a rapidly repriced cap table as demand for the company accelerated.3

The next phase is less about fundraising speed and more about proving Gimlet can become a durable control layer for inference across GPUs and newer AI chips. If it keeps winning workloads from labs, hyperscalers, and large inference operators, the company can grow from a hot infrastructure startup into a standard routing layer inside production AI stacks.4 5 6