- Essays··9 min read
China's exascale push against GPUs · what LineShine's interconnect says about post-NVLink fragmentation
LineShine, which debuted atop the June 2026 TOP500, is the first system to exceed two exaflops of sustained double-precision performance without GPU accelerators. It was built under US export controls from indigenous Arm processors, on-package HBM, and a custom fat-tree fabric called LingQi. The ranking is the least of it: China has closed the interconnect gap that sovereign exascale programmes could not paper over, and the scale-up fabric market is now genuinely fragmented.
supercomputinginterconnectai-infrastructurechina-techRead - Essays··7 min read
The AI Rulebook Is Now a Customs Form
The last forty days of AI regulation were barely about AI. They were about ports, subsidiary structures, export taxes and chips that triangulate their own location. The rulebook is now a customs form.
ai-regulationexport-controlseu-ai-actai-sovereigntyRead - Essays··8 min read
Geopatriation in Plain Sight: A Field Report from Brussels, Delhi, and Washington
Geopatriation, the localisation of AI workloads under sovereign pressure, is the dominant pattern in 2026. The compute sits under local governance; the silicon supply chain has not been localised at all. The truth is in the clauses.
ai-sovereigntyexport-controlseu-ai-actgeopatriationRead