


Expect more from your chips: generating optimized accelerator code from applications to kernels.


Intel walked away from its Gaudi chips. We used AI to make them outperform NVIDIA’s H100. The lesson? Hardware alone doesn’t determine performance or economics. At Standard Kernel, we’re building AI that writes the software to unlock the full potential of more accelerators.


In collaboration with CoreWeave, we show how our autonomous system optimizes LLM inference end to end, from engine configuration to custom GPU kernels: up to 63% more traffic for GLM-5.3, 2x per-user throughput for Qwen3.8-27B, and 10% more throughput for DeepSeek V4.1 Flash. 🔗

good to see we're all innovating 🫡
Startup name matrix filled up fast

If AI can generate CUDA kernels, what does a kernel engineer do now? The work is shifting from hand-writing every kernel toward building the systems that make AI-generated kernels useful in practice — helping agents learn new hardware, improving search and feedback, and building better ways to verify correctness. Generation is getting cheaper, but correctness is becoming a critical problem. A generated kernel can look faster on a benchmark but fail in production because of race conditions, stream interactions, output reuse, or other subtle bugs. That makes verification and deep hardware expertise more central to the job. At Standard Kernel, we’re building that infrastructure. Our cofounder @anneouyang spoke with Business Insider about how kernel engineering is changing. businessinsider.com/cuda-eng…

Full post at standardkernel.com/blog/reim… (5/5)

One exciting application: a universal optimization layer across DSLs. High-level DSLs (Triton, CUTLASS, TileLang, ThunderKittens) are powerful but opaque, as they don’t reveal why one outperforms another. By working at the shared PTX layer, we can compare, learn, and compose their best implementations into kernels that outperform them all. (4/5)

We built a hybrid system that combines traditional program analysis with LLMs. It operates in two phases: first, PTX is condensed into a more tractable representation for the LLM; second, the LLM uses this representation to optimize and learn from the code, building up a growing knowledge base over time (including how to use the latest specialized hardware features!) (3/5)

PTX is a low-level and important layer in the kernel stack to which programs compile. Experts often write inline PTX by hand directly to squeeze out every last bit of performance. PTX is unforgiving and massive –– a simple kernel can span hundreds of thousands of tokens and one wrong token breaks the whole program. The long context and precision required makes this layer essentially out of reach for LLMs for a long time. While progress has helped, LLMs are still not sufficient on their own. (2/5)

We built a system combining program analysis and LLMs to transform and optimize PTX. By operating at this shared layer across DSLs (e.g. Triton, TileLang, ThunderKittens, CUTLASS), our system learns the best ideas from each and generates kernels that outperform all of them (1/5)

We have a utilization problem. GPUs are running <30% capacity. @Standard_Kernel (@anneouyang + @ChrisRinard ) unlocks up to 4x performance. Why we invested: jumpcap.com/insights/why-we-…

AI progress increasingly depends on how efficiently workloads run on hardware. @Standard_Kernel is tackling this challenge at the kernel level, unlocking more performance from modern GPUs. We're proud to lead their seed with @generalcatalyst, @CoreWeave, @felicis, & @ericsson

It’s rare to find founders so perfectly and uniquely suited to solve a problem, let alone a problem of this magnitude and importance. Proud to lead @Standard_Kernel’s seed round.
Excited to share @Standard_Kernel's seed round and some reflections on what we’ve learned about kernel generation and what we believe is next. Grateful to our amazing team, supporters, and the broader community pushing this space forward.

Anne is killing it. Here's my quote from the press release Kernel generation is key for improving performance and efficiency of AI hardware. As fleet sizes for users of AI hardware get larger, and more hardware diversity is introduced, Standard Kernel becomes key to deployment.”
Excited to share @Standard_Kernel's seed round and some reflections on what we’ve learned about kernel generation and what we believe is next. Grateful to our amazing team, supporters, and the broader community pushing this space forward.

Excited to share @Standard_Kernel's seed round and some reflections on what we’ve learned about kernel generation and what we believe is next. Grateful to our amazing team, supporters, and the broader community pushing this space forward.