AI Hardware Archives - https://krolmarc.com

Google launched eighth‑generation TPUs at the same price as the previous generation, promising 2.8× training performance and 80% better inference. Google announced a dedicated training chip and a separate inference chip, each slated for availability later this year, and highlighted a 384 MB SRAM memory on the TPU 8i inference chip, triple the prior generation's capacity, per [CNBC] reported on 2026-04-22. ##Amin Vahdat on Specialized Chips## "With the rise of AI agents, we determined the community would benefit from chips individually specialized to the needs of training and serving," said Amin Vahdat, senior vice president and chief technologist for AI and infrastructure, underscoring the strategic split between training and inference workloads. ##Google vs. Nvidia: Memory and Throughput Compared## Google’s TPU 8i inference chip matches Nvidia’s focus on large memory for rapid responses, while the training chip delivers 2.8× the performance of its predecessor at identical pricing, though Nvidia’s exact figures remain undisclosed; Citadel Securities and all 17 U.S. Energy Department national laboratories are already leveraging the new silicon, with Anthropic committing gigawatts of TPU capacity. ##Amazon and Meta Parallel Custom Chip Strategies## Amazon previously introduced separate training and inference chips in 2018 and 2020, and Meta is collaborating with Broadcom on multiple AI processor versions, showing a broader industry move toward specialized AI silicon. Google expects both TPU chips to be available later this year, expanding the hardware portfolio for Google Cloud customers and reinforcing its AI infrastructure roadmap.

Google TPU 8 Launch: Same Price, 2.8× Faster Training vs Nvidia

mohssin essaddiki3 months ago5 days ago02 mins

Google launched eighth‑generation TPUs at the same price as the previous generation, promising 2.8× training performance and 80% better inference. Google announced a dedicated training chip and a separate inference chip, each slated for availability later this year, and highlighted a 384 MB SRAM memory on the TPU 8i inference chip, triple the prior generation’s capacity,…

Cerebras Systems filed for an IPO at an $8.1 billion valuation in the infrastructure sector. G42 led the round with Amazon Web Services and OpenAI as co-investors, per TechCrunch reported on 2026-04-19.##Andrew Feldman Cerebras Systems CEO##Cerebras Systems, building what CEO Andrew Feldman describes as the fastest AI hardware for training and inference, brought in $510 million in revenue in 2025 with a net income of $237.8 million excluding certain one-time items, though it was a non-GAAP net loss of $75.7 million. G42 Investment Scale: Abu Dhabi Backs##G42's decision to lead rather than co-invest suggests higher conviction, with the company holding $10 billion in strategic deals reported by the Wall Street Journal. ##AWS OpenAI Strategic Deals##Amazon Web Services agreed to use Cerebras chips in Amazon data centers, while OpenAI has a deal reportedly worth more than $10 billion, with Cerebras taking business from Nvidia at inference according to Feldman. ##Nvidia $8.1B Valuation: How Cerebras Compares##Cerebras raised a $1.1 billion Series G at an $8.1 billion valuation last year, facing comparison to established players in AI infrastructure despite being unprofitable on a GAAP basis. Cerebras has not disclosed how much it hopes to raise in the IPO, with a spokesperson indicating the offering is planned for mid-May as the filing process moves forward.##Future IPO Revenue Projections##Cerebras Systems outlined plans for the IPO, though specific raise targets remain undisclosed as the company proceeds with the public offering.

Cerebras Files IPO at $8.1B: AI Hardware Context

mouhcineessadiki@gmail.com4 months ago5 days ago02 mins

Cerebras raised $1.1 billion in Series G at an $8.1 billion valuation. The round closed after the company filed to go public following a delayed 2024 IPO attempt. G42 led the round with Amazon Web Services and OpenAI as co‑investors. The filing did not disclose how the funds will be used, and [TechCrunch] reported on 2026-04-19 that the…