Advancements in semiconductor technology, GPU development, and data center infrastructure that power AI computing capabilities.
Rocket One and Virginia Commonwealth University will conduct a one year research program on magnetic memory based on skyrmions, with work beginning November 1.
American Electric Power will buy up to 1.6 GW of Mainspring Linear Generators through 2034, starting with a 190 MW first tranche for delivery from 2027.
NTT DATA is reportedly in final talks to build Anthropic its first dedicated Japan data center in Osaka, at a cost of up to about $6.3 billion, the Nikkei reported.
Ubitium raised $14 million to commercialize a processor that combines sensing, AI inference and machine control. Evaluation units for its UB410 chip are planned for 2027.
Seeq and Applied Computing signed a commercial agreement that combines the Orbital foundation model with Seeq's industrial analytics platform for energy operations.
Supermicro expanded its Open Compute Project portfolio with rack scale systems, centralized power, liquid cooling, networking and management software for AI data centers.
Orbius and Armada have introduced an edge cloud service that combines satellite connectivity with modular data centers for AI processing at remote sites.
Nebius signed a letter of intent to become the first off taker at VOLT planned 600 MW Dutch AI Gigafactory in Rotterdam.
Broadcom CEO Hock Tan says capital may become a bigger constraint on AI growth than shortages of chips, memory or power, while Intel CEO Lip-Bu Tan points to the cost of capital and the need for returns.
Lightmatter unveiled a 204.8 Tbps optical switch tray and a 102.4 Tbps XPU tray at the 2026 OCP Global Summit. Both systems fit in one rack unit and use the company's Passage optical engines and Guide DR laser modules.
Wiwynn expanded its OCP Global Summit showcase with an AMD Helios rack, new liquid cooling designs and a co packaged optics system.
Wistron and Wiwynn will present NVIDIA Rubin compute, storage and networking systems at the 2026 OCP Global Summit in San Jose.
The XA P2N-E2 combines NVIDIA Vera CPUs with support for up to two RTX PRO 6000 Blackwell Server Edition GPUs in a 2U server.
Cisco and NVIDIA plan to develop an open specification for a pluggable optical module supporting 400 gigabits per second per lane. Cisco also introduced new networking silicon and systems for AI infrastructure.
Vuzix will demonstrate its Causeway optical interconnect evaluation platform at two October industry events and offer evaluation units to qualified partners.
Envalior will present engineering materials for AI data center power distribution, connectivity and cooling components at Fakuma 2026 in Germany.
Microsoft will expand its Azure Maia inference system to deployments spanning multiple gigawatts from 2027, when Maia 300 will enter production.
insidepacket's AI Access Platform connects enterprises with participating GPU cloud providers through private operator networks. The platform will first be introduced in Europe with Zayo Europe.
Texas Instruments introduced an analog to digital converter that continuously monitors data center power rails and captures transient faults at nanosecond resolution.
MiTAC is presenting new AI servers, modular data center systems, open source firmware and liquid cooling technology at the OCP Global Summit 2026.
Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.
