Loading organizations...
§ Private Profile · Santa Clara, CA, USA
Lemurian Labs is a technology company.
Lemurian Labs provides Tachyon, a unified software stack that liberates AI workloads from specific hardware limitations. This platform enables organizations to develop AI applications once and deploy them across diverse hardware environments, ensuring consistent, high performance and true portability. Its approach integrates compiler technology and runtime orchestration, establishing a universal solution that operates efficiently irrespective of underlying infrastructure.
The company was co-founded by Jay Dawani, who serves as CEO, drawing on his background in robotics and foundational work at OpenAI. Dawani's insight centered on the need for a radical new approach to AI software development, driven by the belief that AI should not be constrained by hardware dependencies. This vision aims to streamline the deployment and scalability of complex AI models.
Lemurian Labs’ offering targets organizations seeking to simplify AI development and deployment by eliminating hardware-specific optimizations. The company’s long-term vision focuses on creating an ecosystem where AI runs faster and more efficiently across any hardware, empowering developers with unparalleled flexibility and enabling broader accessibility for advanced AI applications.
Lemurian Labs has raised $37.0M across 2 funding rounds.
Lemurian Labs has raised $37.0M in total across 2 funding rounds.
Lemurian Labs is a Toronto-based technology company founded in 2021 that builds hardware-agnostic AI infrastructure to make AI development fast, affordable, and scalable across any hardware.[1][3][4][5] Its core products include Tachyon, a software stack that ingests PyTorch models and runs AI workloads on CPUs, GPUs, and custom accelerators with performance matching or exceeding hand-tuned kernels, and the Spatial Processing Unit (SPU), a general-purpose AI accelerator under development offering up to 20x greater throughput and 1/10th the cost of legacy GPUs.[2][5] Targeting AI developers, researchers, startups, and companies, Lemurian Labs solves key pain points like bloated costs, hardware lock-in, kernel fragmentation, and infrastructure complexity that hinder innovation.[3][4][5]
The company serves the AI development industry by enabling portability, higher productivity, and efficiency, allowing users to focus on models rather than hardware constraints—pivoting from initial edge AI for robotics to cloud-scale accelerated computing.[1][2]
Lemurian Labs emerged in early 2018 when co-founders Jay Dawani (CEO) and his partner tackled compute shortages while building a foundation model for general-purpose autonomous robotics and generative multiphysics simulation.[1][2] Frustrated by unattainable compute resources and the lack of suitable edge processors balancing throughput, latency, energy efficiency, and programmability, they shifted from a RoboOps platform to designing better AI hardware and software solutions.[1][2]
Incorporated in 2021 and headquartered in Toronto, Ontario, the team draws from decades at NVIDIA, Intel, Google, AMD, Uber, and others in AI systems, compilers, CPU/GPU design, and numerical algorithms.[1][4] Early traction came from rapid iteration on hard problems, de-risking strategies, and pivoting from edge AI to cloud-focused accelerators amid robotics compute bottlenecks.[1]
Lemurian Labs rides the AI infrastructure explosion, addressing compute scarcity, skyrocketing costs, and hardware silos amid foundation models' rise—enabling robotics, multiphysics sims, and scalable AI without Big Tech dominance.[1][2][4] Timing aligns with xPU proliferation (beyond GPUs) and cloud flexibility demands, as legacy systems fail modern workloads' scale.[3][5]
Market forces like hyperscaler shortages and energy crises favor its efficient, portable stack, democratizing AI for startups/scientists versus elite players.[1][2][4] It influences the ecosystem by liberating developers from "post-kernel era" barriers, accelerating innovation in autonomy, science, and multi-agent AI—potentially unlocking platform shifts like robotics if scaled.[1][5]
Lemurian Labs is poised to disrupt AI compute with Tachyon beta in Summer 2026, scaling SPU for production and onboarding "conspirators" to rewrite performance norms.[5] Trends like heterogeneous computing, edge-to-cloud AI, and cost pressures will amplify its edge, evolving it from enabler to foundational layer for accessible, high-impact AI.[4][5]
As hardware wars intensify, its agnostic approach could redefine accessibility, empowering broader humanity-scale breakthroughs while outpacing rigid incumbents—turning today's infrastructure chokehold into tomorrow's unlimited possibilities.[2][3][5]
Lemurian Labs has raised $37.0M in total across 2 funding rounds.
Lemurian Labs's investors include Hexagon, Keith Adams, 1Flourish Capital, Animal Capital, Blackhorn Ventures, Origin Ventures, Oval Park Capital, Planetary Impact Ventures, Silicon Catalyst Ventures, Stepchange Ventures, Salil Deshpande, Untapped Ventures.
Lemurian Labs has raised $37.0M across 2 funding rounds. Most recently, it raised $28.0M Series A in December 2025.
| Date | Round | Lead Investors | Other Investors | Status |
|---|---|---|---|---|
| Dec 4, 2025 | $28M Series A | Hexagon, Keith Adams | 1Flourish Capital, Animal Capital, Blackhorn Ventures, Origin Ventures, Oval Park Capital, Planetary Impact Ventures, Silicon Catalyst Ventures, Stepchange Ventures, Salil Deshpande, Untapped Ventures | Announced |
| Sep 1, 2023 | $9M Seed | Justin Wright Eakes | Aleph VC, Alumni Ventures, Amplify Partners, Bascom Ventures, Dimension Capital, Draper Associates, Humba Ventures, Modern Venture Partners, Streamlined Ventures, Susa Ventures, Upside Partnership, Vzvc, Y Combinator, Alain Hanover, Charlie Songhurst, Erick Miller, Farzad Nazem, IAN Mcnish, Richard Cooperstein, Good Growth Capital, Laura Rippy, Raptor Group | Announced |