Side-by-side comparison of AI visibility scores, market position, and capabilities
Transformer-specific ASIC startup raised $500M at $5B valuation in Jan 2026; Sohu chip claims 20x Nvidia H100 inference speed for transformer workloads; fabricated on TSMC 4nm process alongside Apple and Nvidia silicon.
Etched is a semiconductor startup founded in 2022 that is building application-specific integrated circuits (ASICs) optimized exclusively for transformer-based neural network inference. Unlike general-purpose GPUs that must support a broad range of workloads, Etched's Sohu chip is hardwired at the silicon level to execute the transformer architecture — the mathematical backbone of virtually every major AI model including GPT, Gemini, and Claude. By eliminating the flexibility overhead of general-purpose hardware, Etched claims inference speeds up to 20x faster than Nvidia's H100 for transformer workloads, with corresponding reductions in cost per token.\n\nThe Sohu chip is fabricated on TSMC's 4nm process node, the same cutting-edge manufacturing technology used by Apple and Nvidia for their flagship chips. Etched targets large-scale inference deployments — hyperscalers, AI cloud providers, and enterprises running high-volume language model workloads where inference cost is the dominant operational expense. The chip is designed to slot into existing data center infrastructure and provide dramatic efficiency gains for organizations serving billions of AI queries daily.\n\nEtched raised $500M at a $5B valuation in January 2026, a financing round that placed it among the most highly valued AI chip startups globally. The raise reflects investor conviction that transformer inference will remain a dominant workload for years to come and that purpose-built silicon can capture significant market share from Nvidia in this specific segment. Etched is competing in the AI chip market alongside Google's TPUs, Amazon's Trainium/Inferentia, and startups like Groq and Cerebras.
AI chip startup by ex-Google TPU engineers raised $500M+ Series B in Feb 2026 led by Jane Street; chips target 10x Nvidia for LLM training; shipping 2027 via TSMC
MatX is a Silicon Valley AI chip startup founded by former Google engineers who led development of the Tensor Processing Unit (TPU), Google's proprietary chip for large-scale AI workloads. The company was founded on the thesis that the AI infrastructure market requires purpose-built silicon optimized specifically for large language model inference and training — a different design philosophy from Nvidia's general-purpose GPU architecture. MatX's founding team brings direct experience designing the chips that power Google's internal AI at scale, giving it deep technical credibility in a capital-intensive field.\n\nMatX is building chips that target a 10x performance advantage over Nvidia hardware for LLM training and inference workloads, by stripping away general-purpose compute features and maximizing memory bandwidth and interconnect efficiency for transformer model architectures. The chips are designed to serve hyperscalers, AI labs, and large enterprises that run inference at scale, where per-token cost and throughput determine economic viability. MatX plans to begin shipping hardware in 2026, moving from design into commercial production after closing its Series B.\n\nMatX raised over $500 million in a Series B round in February 2026 led by Jane Street, one of the most sophisticated quantitative trading firms in the world — a signal that sophisticated capital views MatX's technical claims as credible and its market timing as right. The round values MatX as a serious contender in the AI chip market that has so far been dominated by Nvidia. As AI inference costs become a primary competitive variable for AI product companies, purpose-built chips from startups with proven TPU pedigrees represent a credible alternative to the incumbent.
Monitor how your brand performs across ChatGPT, Gemini, Perplexity, Claude, and Grok daily.