# matx.com

> AI-optimized mirror of matx.com containing 13 pages totalling 10,476 words of clean markdown content, structured data, and semantic HTML. Original source: https://matx.com/. Last updated: 2026-05-17T19:21:26.663Z.

Each page is available as HTML (with JSON-LD structured data) and Markdown (text-only, ideal for LLMs and RAG).

## Homepage

- [High-throughput chips for LLMs](/site-root.html): We make the best chips physically possible for the large model needs of frontier labs. (179 words)

## Articles & Blog Posts

- [research/rules_derive/index.html](/research/rules_derive/index.html) (1 words)
- [research/smva/index.html](/research/smva/index.html) (1 words)
- [Large Language Models and Their Cost Dynamics](/research/lifetime_llm_cost/index.html): We discuss the importance of and strategies to balance training and inference costs when selecting LLM architectures. (1,503 words)
- [Speculative Decoding in Language Models](/research/sd/index.html): Speculative decoding (SD) has been shown to reduce the latency of autoregressive decoding (AD) by 2-3× for small batch sizes. However, increasing throughput and (4,022 words)
- [Jobs](/jobs/index.html): We're building optimized chips and systems for running large language models. Join our team in Mountain View, California. (916 words)
- [Quantizing Attention Improves Model Efficiency](/research/leaky_quantization/index.html): Quantizing attention improves efficiency on two fronts: the model has higher compute throughput, and loads fewer bytes per key/value. However, training with blo (985 words)
- [404: NOT_FOUND](/about/index.html) (14 words)
- [Introduction](/research/seqax/index.html): We're excited to announce seqax, a research-focused LLM codebase that is simple, efficient, and performs well on up to 100 GPUs or TPUs. Everything you need to  (560 words)
- [Speculative Decoding and Blockwise Sparse Attention](/research/sd_nsa/index.html): Speculative decoding (SD) and blockwise sparse attention both accelerate LLM decoding, but when combined naively, the KV cache may lose sparsity during the veri (1,842 words)
- [February 24, 2026](/research/series_b/index.html) (314 words)
- [sitemap-xml.html](/sitemap-xml.html) (44 words)
- [Announcements](/research/index.html): Research, updates, and news from the MatX team. (95 words)

## Resources

- [Full Page Index](/index.html): Browse all cached pages with rich metadata
- [About This Cache](/content/about.html): Methodology, technical details, and usage guidelines
- [XML Sitemap](/content/sitemap.xml): Machine-readable sitemap for crawler discovery
- [Robots.txt](/content/robots.txt): Crawler directives