Key Takeaways

  • Megaspeed bought $4.6 billion in Nvidia hardware to become the supplier's biggest buyer in Southeast Asia, four days after incorporating in Singapore via a British Virgin Islands shell.
  • The neocloud channeled AI compute directly to Alibaba, bypassing strict hardware export bans by selling cloud access instead of physical silicon.
  • Cross-referencing Malaysian and Indonesian trade manifests against physical data center capacity revealed roughly 50,000 advanced GPUs missing from declared sites.
  • When questioned by reporters, Nvidia claimed the missing hardware sat in newly discovered Malaysian warehouses, but local data center operators contradicted that story on the record.
  • Huawei is actively exploiting these supply chain bottlenecks to sell domestic AI hardware across the Global South despite its own chip manufacturing limits.

The Four-Day Partner

Paper trails move faster than physical silicon. Bloomberg reporter Mackenzie Hawkins tracked how Megaspeed transformed from an obscure entity into an export powerhouse. The company “purchased $4.6 billion worth of Nvidia hardware after being spun off from a Chinese gaming company, passing through a BVI shell, and incorporating in Singapore, then becoming an official Nvidia partner 4 days later.” (play from 6:11)

That speed allowed Nvidia to move massive volume into Southeast Asia while maintaining technical compliance on paper. But Megaspeed was not building a retail cloud for regional startups. Hawkins noted that “Megaspeed is the single largest Southeast Asian buyer of Nvidia chips. It's a major supplier of AI compute to Alibaba, which is one of the most important tech companies and frontier AI leaders in China.” (play from 5:59)

By routing compute to Alibaba through offshore infrastructure, Chinese tech firms can run massive model experiments without importing a single restricted board across their physical borders.

Fifty Thousand Missing Processors

The compliance story unravels once you compare shipping manifests to real estate. Hawkins and her reporting team pulled import filings across the region to see where billions of dollars of hardware actually went.

“We compared that number to Malaysian and Indonesian trade records that we had gotten and found that there were about 50,000 chips unaccounted for, which is like not a small amount of compute,” (play from 20:20) Hawkins said.

Fifty thousand advanced processors can power frontier training runs. When reporters presented the gap, the vendor scrambled for an explanation. Hawkins recounted: “And after we asked about that, they went back to Malaysia, and they said, 'Oh, we found these warehouses that contain what's missing.'” (play from 20:48)

That explanation failed to satisfy regional partners. Hawkins contacted local facility managers to verify Nvidia's account. “Two of the others directly contradicted to us, on the record, Nvidia's account of the hardware as described to the US government.” (play from 22:09)

The gap between what hardware vendors report to Washington and what sits inside Southeast Asian data centers points to an open secret: once chips leave the dock, vendor tracking dissolves into paper declarations. Meanwhile, Huawei is stepping into the vacuum across developing markets, pitching domestic alternatives while American firms face tightening cloud rules.

What to Do With This

Audit your compute supply chain before signing long-term capacity contracts. If you rent clusters from secondary neoclouds or overseas brokers, require the exact street address, data center operator name, and cage number in the master services agreement. If a provider cannot certify physical rack residency within 48 hours, cancel the contract to protect your workloads from sudden export enforcement seizures.