Loading organizations...
§ Private Profile · 255 Potrero Ave, San Francisco, CA 94103
Builds an inference cloud service optimized for AI agents, running workloads across diverse hardware for 3-10X faster performance.
Gimlet Labs is an artificial intelligence infrastructure company building an inference cloud optimized for AI agents by running workloads across diverse hardware architectures. The software automatically maps agentic tasks to heterogeneous hardware, including GPUs and SRAM-centric silicon, delivering three to ten times faster performance at the same power consumption while making AI processing up to one hundred times more efficient. Operating with eight-figure revenues, the company provides cloud services to a customer base that recently tripled to include a top hyperscaler and a leading frontier model laboratory. The enterprise emerged from stealth in late 2025 and secured $80 million in Series A funding led by Menlo Ventures, with participation from Eclipse, Factory, and Prosperity7. Spun out of a Stanford University research project, Gimlet Labs was founded by Zain Asgar, Michelle Nguyen, Omid Azizi, Natalie Serrino, and James Bartlett.
Gimlet Labs has raised $92.0M across 2 funding rounds.
Gimlet Labs has raised $92.0M in total across 2 funding rounds.
# Gimlet Labs: AI Infrastructure for the Agentic Era
Gimlet Labs is an applied research lab building infrastructure software to make AI workloads 10X more efficient.[3] The company addresses a critical bottleneck in modern AI systems: as agentic AI applications generate 5-15X more tokens than traditional chat models, infrastructure teams struggle with GPU efficiency, cost, and scaling.[5] Gimlet's platform decouples agentic workloads from specific hardware by intelligently orchestrating compute across heterogeneous accelerators—routing compute-bound tasks to high-throughput GPUs, memory-bound tasks to higher-bandwidth accelerators, and network-bound tasks to nodes with fast interconnect.[5]
The company emerged from stealth mode in October 2025 with a $12 million seed round led by Intel CEO Lip-Bu Tan, positioning itself as a key player in AI software efficiency.[2] With early revenues already in the eight-figures, Gimlet is deploying its platform across AI-native startups and Fortune 500 companies.[2] The startup's core offering—Gimlet Cloud—provides serverless inference for AI agents, handling scheduling, orchestration, and optimization so developers can focus on building agentic capabilities.[3]
Gimlet Labs was founded by researchers including Omid Azizi and Natalie Serrino as co-founders, positioning the company as an applied research lab rather than a traditional startup.[4] The founding emerged from a clear observation: the rapid advancement of AI models has outpaced infrastructure capabilities. As agentic systems became more prevalent, the computational demands exploded—yet existing hardware orchestration solutions were rigid, inefficient, and unable to adapt to diverse accelerator types.[5]
The company's early traction validated the problem-solution fit. By the time of its public launch in late 2025, Gimlet had already achieved significant revenue deployment across enterprise customers, suggesting the founders had been quietly building and validating the platform during their stealth phase.[2] This approach—emerging with both funding and proven customer adoption—reflects a research-first mentality focused on solving hard infrastructure problems rather than chasing hype.
Gimlet Labs sits at the intersection of three powerful trends reshaping AI infrastructure. First, the shift toward agentic AI systems is creating unprecedented demand for inference compute—agents that reason, plan, and iterate generate orders of magnitude more tokens than simple chat interfaces, straining existing GPU capacity.[5] Second, GPU scarcity and cost remain acute constraints despite increased supply, making efficiency gains economically critical for both startups and enterprises.[2] Third, the heterogenization of AI hardware—with specialized accelerators from Intel, AMD, Cerebras, and others competing with NVIDIA—creates an opportunity for software that abstracts away hardware specificity.
Gimlet's timing is strategic. As enterprises deploy agentic systems at scale, infrastructure costs become a primary concern. A 10X efficiency gain translates directly to competitive advantage: more tokens per dollar, lower latency, and fuller utilization of existing hardware investments.[5] The company's backing by Intel's CEO signals that even chip manufacturers recognize the value of orchestration software that can distribute workloads across diverse silicon—a hedge against single-vendor lock-in.
By solving the orchestration problem, Gimlet influences the broader ecosystem by democratizing access to efficient AI compute. Smaller companies and startups that cannot afford massive GPU clusters can now run agentic workloads more cost-effectively, potentially accelerating the adoption of AI agents across industries.
Gimlet Labs is positioned to become a critical infrastructure layer in the agentic AI era. The company's research-first approach, combined with early revenue traction and heavyweight backing, suggests it will likely expand its platform offerings—moving from orchestration into adjacent areas like cost optimization, multi-tenant scheduling, and edge deployment.[3][4]
The key question ahead is adoption velocity. If agentic AI becomes as ubiquitous as expected, Gimlet's software could become as foundational as Kubernetes is for containerized workloads. Conversely, if GPU efficiency improves faster than anticipated or if cloud providers (AWS, Google Cloud, Azure) build competing orchestration into their platforms, Gimlet's addressable market could narrow.
What's certain: as long as agentic systems generate exponential compute demand and hardware remains heterogeneous, the infrastructure problem Gimlet solves will only grow more acute. The company's ability to stay ahead of both hardware innovation and agentic AI evolution will determine whether it becomes an essential utility or a specialized tool.
Gimlet Labs has raised $92.0M in total across 2 funding rounds.
Gimlet Labs's investors include Menlo Ventures, Tim Tully, Lip-Bu Tan, Nick McKeown, Raghu Raghuram, W. M. Coughran, Eclipse Ventures, Factory, Abhishek Shukla, Triatomic Capital, Nepenthe Capital, Walden International.
Gimlet Labs has raised $92.0M across 2 funding rounds. Most recently, it raised $80.0M Series A in March 2026.
| Date | Round | Lead Investors | Other Investors | Status |
|---|---|---|---|---|
| Mar 1, 2026 | $80M Series A | Menlo Ventures, TIM Tully, Anthology Fund | LIP BU TAN, Nick Mckeown, Raghu Raghuram, W. M. Coughran, Eclipse Ventures, Factory, Abhishek Shukla, Triatomic Capital | Announced |
| Oct 1, 2025 | $12M Seed | — | Nepenthe Capital, Walden International, Akshay Kothari, Dylan Field | Announced |