ByteDance, Alibaba and Tencent order $35bn of Nvidia chips, Reuters reports
THE SO WHAT
ByteDance, Alibaba, and Tencent reportedly ordering over 3 million RTX PRO 5500s—$35 billion of GPUs—shows Chinese hyperscalers are front-loading compute while they can. If you rely on the same hardware, expect tighter supply, longer lead times, and more pressure to optimize utilization per watt and per dollar.
READ THE SOURCE
MORE FROM THE WIRE
AI rivals are thirsty for SpaceX's compute, Elon Musk says
If hyperscalers are “hounding” SpaceX for compute, orbital and remote sites are becoming part of the AI capacity map, not a sideshow. For teams planning multi-year training roadmaps, power, geography, and non-traditional providers now belong in the sourcing conversation.
Applied AI'Faster code is not the same as better software delivered sooner': New report claims AI isn't solving productivity across the entire software development lifecycle - yet
If AI is only speeding up coding but not the full lifecycle, the constraint has moved to design, integration, and verification. Treat AI coding tools as accelerators inside a reworked SDLC — not as a magic lever on “developer productivity.”
Applied AIHarvey adds Australian startup Mary’s case-fact tools to its legal AI
Harvey wiring in Mary’s case-fact system — plus Autologyx — is a move toward AI that sits on top of a firm’s actual evidentiary graph, not generic text. If you’re buying legal AI, the key question is now: how cleanly does it plug into your fact base and workflows, not how clever the model is.
Applied AIDell put Nvidia DGX Spark silicon into its Pro Precision workstation — and undercut it by $1,000
High-end AI silicon is leaking out of rarefied datacenter boxes into relatively accessible workstations—budget lines that assumed “cloud-only” training and inference need a refresh. If your teams are GPU-constrained, it’s time to re-run the math on owning a few of these boxes versus renting everything from the cloud.