Google in talks with Marvell Technology to build new AI inference chips alongside Broadcom TPU programme
Briefly

Google in talks with Marvell Technology to build new AI inference chips alongside Broadcom TPU programme
"Google is in talks with Marvell Technology to develop two new chips for running AI models, including a memory processing unit and a new TPU specifically for inference."
"The discussions have not yet produced a signed contract, indicating that while negotiations are ongoing, formal agreements are still pending."
"Google's seventh-generation TPU, Ironwood, debuted this month as the first TPU designed for the age of inference, delivering ten times the peak performance of its predecessor."
"The custom ASIC market is projected to grow 45% in 2026, reaching $118 billion by 2033, highlighting the increasing importance of inference in AI processing."
Google is negotiating with Marvell Technology to create two AI chips: a memory processing unit and an inference-optimized TPU. This partnership adds Marvell as a third design partner alongside Broadcom and MediaTek. The discussions follow Broadcom's recent long-term TPU agreement and indicate Google's focus on inference as a key computing cost. The custom ASIC market is expected to grow significantly, with Google's latest TPU, Ironwood, designed for inference, showcasing substantial performance improvements and plans for mass production this year.
Read at TNW | Artificial-Intelligence
Unable to calculate read time
[
|
]