Rack-scale infrastructure will overcome constraints of traditional copper interconnects

Launches aimed at supporting transition towards multi-rack scale-up and high-bandwidth scale-out in AI data centers

Designed to support generative AI workloads, the MTIA 300, 400, 450, and 500 will all be deployed by 2027

SK Hynix has become a founding partner of the EPIC Center, slated to open later this year

The MCUs are integrated with company’s TinyEngine NPU to speed up processing and cut energy use

Forthcoming PIC100 TSV platform will support both near- and co-packaged optics

Financial terms of the multi-year agreement have not been disclosed

Company claims ‘wetware’ deployed in Melbourne and Singapore facilities will require “fraction of the energy” needed by digital systems

Company has already received a $300 million initial order for the servers in the US

Financial terms of the deal have not been disclosed

Comments come after company unveiled its fifth generation Epyc 8005 Server CPUs

Decision results from lack of meaningful GPU sales to Chinese customers

Deal provides French startup with pre-money valuation of $2bn

Each CPU contains up to 288 Efficiency-cores

Second such agreement signed by the chip giant this week

Latest fundraise values the company at $3.75bn

Latest high-profile executive to leave the chipmaker

Full-stack platform will support applications across on-premises, Edge, and cloud environments

Company founded by two ex-Googlers plans to start shipping MatX One chip next year

System would allow companies to process up to 330 wafers an hour, up from 220 at present

First gigawatt scheduled for deployment later this year, with companies to align across silicon, systems, and software roadmaps

Funding round included participation from BlackRock, European Investment Council Fund, and Samsung Catalyst Fund

Chip will support agentic AI workloads, with SoftBank named as first customer for new hardware offering

Working with academia and industry