Google TPU Veteran Launches MatX, Secures $500 Million to Boost AI Performance 10x
MatX, an AI chip startup established by two ex-Google hardware engineers, has closed a $500 million Series B funding round. The investment was co-led by quantitative trading firm Jane Street and Situational Awareness, a fund established by former OpenAI researcher Leopold Aschenbrenner.
The founding team brings deep expertise to the table. CEO Reiner Pope previously managed AI software development for Google’s TPU and played a pivotal role in optimizing the efficiency of Google’s PaLM model. CTO Mike Gunter, the chief architect behind Google’s TPU hardware, brings nearly three decades of chip design experience.

Challenging NVIDIA with a Tenfold Performance Boost
Unlike NVIDIA GPUs, which are built for versatility across gaming, graphics, and scientific computing, MatX pursues a highly specialized strategy. The company is engineering a processor dedicated exclusively to large language models (LLMs). By eliminating the redundant computing modules inherent in general-purpose GPUs, MatX claims its chips can process Transformer architecture models up to ten times faster than current NVIDIA offerings.
This “less is more” philosophy targets the escalating costs of AI compute. The latest funding round also attracted investors including Marvell Technology, NFDG, Spark Capital, and Stripe co-founders Patrick and John Collison.
Mass Production Scheduled for 2027, Valuation Nears $5 Billion
While MatX has not officially released its latest valuation, it was valued above $300 million during its Series A round in 2024. Given the market success of direct competitor Etched, which recently closed a similarly sized round at a $5 billion valuation, industry sentiment remains optimistic about MatX’s potential.
MatX has secured a manufacturing partnership with TSMC and intends to use this capital to accelerate chip prototyping and mass production, targeting official shipments to global AI labs in 2027
Related article
How to fix Core Web Vitals for better SEO ranking
Build a Free Website with a Complimentary Domain and HostingIntroductionPart 1: Launching your website for freeStep 1: Creating an account in profreehostStep 2: Getting your free domainStep 3: Installing Wordpress on your domainPart 2: Building your
Samsung to Supply OpenAI With HBM4, Challenging NVIDIA’s AI Dominance
To eliminate dependence on NVIDIA’s computing infrastructure, OpenAI is rapidly expanding its proprietary hardware operations, positioning South Korean semiconductor leader Samsung as a key partner.According to recent reports from the Korean Economic
Anthropic reports workforce seniority rising as meeting time increases
Anthropic’s workforce composition and work patterns have shifted dramatically over the last year, according to a report released by the AI company on Friday.Anthropic suggests its internal trends may signal broader economic shifts in hiring, labor de
Related Special Topic Recommendations
Comments (0)
0/500
MatX, an AI chip startup established by two ex-Google hardware engineers, has closed a $500 million Series B funding round. The investment was co-led by quantitative trading firm Jane Street and Situational Awareness, a fund established by former OpenAI researcher Leopold Aschenbrenner.
The founding team brings deep expertise to the table. CEO Reiner Pope previously managed AI software development for Google’s TPU and played a pivotal role in optimizing the efficiency of Google’s PaLM model. CTO Mike Gunter, the chief architect behind Google’s TPU hardware, brings nearly three decades of chip design experience.

Challenging NVIDIA with a Tenfold Performance Boost
Unlike NVIDIA GPUs, which are built for versatility across gaming, graphics, and scientific computing, MatX pursues a highly specialized strategy. The company is engineering a processor dedicated exclusively to large language models (LLMs). By eliminating the redundant computing modules inherent in general-purpose GPUs, MatX claims its chips can process Transformer architecture models up to ten times faster than current NVIDIA offerings.
This “less is more” philosophy targets the escalating costs of AI compute. The latest funding round also attracted investors including Marvell Technology, NFDG, Spark Capital, and Stripe co-founders Patrick and John Collison.
Mass Production Scheduled for 2027, Valuation Nears $5 Billion
While MatX has not officially released its latest valuation, it was valued above $300 million during its Series A round in 2024. Given the market success of direct competitor Etched, which recently closed a similarly sized round at a $5 billion valuation, industry sentiment remains optimistic about MatX’s potential.
MatX has secured a manufacturing partnership with TSMC and intends to use this capital to accelerate chip prototyping and mass production, targeting official shipments to global AI labs in 2027
How to fix Core Web Vitals for better SEO ranking
Build a Free Website with a Complimentary Domain and HostingIntroductionPart 1: Launching your website for freeStep 1: Creating an account in profreehostStep 2: Getting your free domainStep 3: Installing Wordpress on your domainPart 2: Building your
Samsung to Supply OpenAI With HBM4, Challenging NVIDIA’s AI Dominance
To eliminate dependence on NVIDIA’s computing infrastructure, OpenAI is rapidly expanding its proprietary hardware operations, positioning South Korean semiconductor leader Samsung as a key partner.According to recent reports from the Korean Economic
Anthropic reports workforce seniority rising as meeting time increases
Anthropic’s workforce composition and work patterns have shifted dramatically over the last year, according to a report released by the AI company on Friday.Anthropic suggests its internal trends may signal broader economic shifts in hiring, labor de





Home






