The Firmus IPO Died on Delivered Megawatts, Not on Price
The Firmus IPO is off. Firmus withdrew its $5 billion ASX listing after a delivery partner stepped back from a 1.6GW AI buildout that reached 42MW.
Independent reporting on Kubernetes, cloud-native infrastructure, virtualization, storage, and security. For engineers and platform teams.
The Firmus IPO is off. Firmus withdrew its $5 billion ASX listing after a delivery partner stepped back from a 1.6GW AI buildout that reached 42MW.
AI data center grid costs land on the developer in Texas from 8 October, with a $100,000 study fee and $50,000 per megawatt posted as security.
Meta's Hyperion data center power demand is 4,500 megawatts, four times New Orleans, and Louisiana is fast-tracking ten gas units to serve it.
GlobalFoundries will make silicon interposers for TSMC CoWoS packaging in New York, the first US source for an AI hardware bottleneck.
Compute utilization across Kubernetes fleets sits near 5 percent. A new grid says 750 megawatts of that idle hardware is already in sight.
AI data center power is the bottleneck. Builders are buying dead factories for their grid connections, and Sesterce bet 10 billion euros on a paper mill.
AI chip financing now runs through the seller. Broadcom, Oracle and SpaceX are borrowing billions to buy accelerators, and the vendor backstops the debt.
CoreWeave India arrives as a 240 MW lease at AdaniConneX's Taloja campus, with Nvidia Vera Rubin racks due in 2028 and room to double the capacity.
Arista's Etherlink SU-144 puts open scale-up Ethernet inside the AI rack, the layer Nvidia's NVLink owned. Up to 144 accelerators in one hop.
Morgan Stanley puts the US AI data center power shortage at 32 GW by 2028. Nvidia and Broadcom are shielded. Memory and optics are not.
Independent analysis of cloud-native infrastructure, Kubernetes and data center economics. No vendor spin.