AI Infrastructure
-
Business
SpaceX lines up about $40 billion from an Apollo-led lease to pay for NVIDIA Vera Rubin GPUs
SpaceX has signed a binding commitment for a master finance lease under which an Apollo-managed entity will buy GPU pods and lease them to the company, with principal payments of roughly $40 billion. The deal is expected…
-
Infrastructure & Chips
Tata Communications commits up to $450 million to a 10,000-km fibre 'AI Super Highway' across India
Tata Communications plans to spend $400 million to $450 million on roughly 10,000 km of new long-haul fibre that will tie India's main data centre clusters to one another and to its subsea cable landing stations, a bet t…
-
Infrastructure & Chips
Eagle releases EAGLE-4, a speculative-decoding drafter that adjusts itself to batch size and hardware, with plugins for SGLang, vLLM and TokenSpeed
The team behind the EAGLE family of speculative-decoding methods published EAGLE-4 on October 20, 2026: inference code, the eagle-harness package and trained drafters that size each speculative draft from measured costs.…
-
Infrastructure & Chips
NVIDIA signs capacity deals with six suppliers to ease AI hardware bottlenecks, bringing Samsung Electro-Mechanics into its substrate supply chain
At GTC Berlin on October 21, NVIDIA said agreements with Amkor, IBIDEN, KYEC, Samsung Electro-Mechanics, SPIL and Unimicron will expand substrate, packaging, testing and optical-engine capacity, including 2.5 times more …
-
Infrastructure & Chips
Google Cloud readies A5X instances on NVIDIA Vera Rubin NVL72 and plans up to 960,000 Rubin GPUs
At GTC Berlin on October 21, 2026, NVIDIA and Google Cloud detailed A5X, Google's first Vera Rubin NVL72 instances, with general availability due next month and a European rollout planned. Google Cloud also made confiden…
-
Business
SpaceX's AI unit brings in $7.6 billion in the third quarter as leasing out compute pushes it to an operating profit
SpaceX reported on October 20, 2026 that its Super Intelligence segment, which houses Grok, Cursor and its data center business, earned $7.56 billion in third-quarter revenue, nearly triple the previous quarter, and $431…
-
Infrastructure & Chips
Meta opens up Prometheus, its gigawatt-scale Ohio supercluster, and says it is scaling toward about 1 million NVIDIA GPUs
Meta published a detailed engineering account of Prometheus on October 19, 2026. The New Albany, Ohio supercluster's first phase, now hundreds of thousands of NVIDIA GPUs across several buildings, trained the Muse models…
-
Infrastructure & Chips
Microsoft pulls out of Nscale's flagship Loughton AI data centre as grid power slips into the 2030s
Microsoft has ended its agreement to anchor Nscale's planned AI supercomputer in Loughton, Essex, the Telegraph reported on October 18, 2026, after grid officials signalled the site may not get enough power until the 203…
-
Infrastructure & Chips
Scaleway plans an AMD MI355X AI supercomputer in Sweden and an NVIDIA B300 GPU Zone in Copenhagen
French cloud provider Scaleway announced at its ai-PULSE event two Nordic AI compute sites: its first supercomputer outside France, built on AMD Instinct MI355X in Sweden, and a Copenhagen cluster of several thousand NVI…
-
Infrastructure & Chips
NTT Data nears deal to build Anthropic's first dedicated data center in Japan, a 500-megawatt site in Osaka
Nikkei and the Asahi Shimbun report that NTT Data Group is close to agreeing a deal to build Anthropic an Osaka data center of up to 500 megawatts, with construction costs of up to about 1 trillion yen. Neither company h…