Chris McGuire(@ChrisRMcGuire):What I said was it would triple the amount of AI compute China will bring online this year (not including illegally smuggled chips). The math on this checks out: Per public reports, China will make 750,000 Huawei Ascend 950PRs this year. Each 950PR is 500 TFLOPS FP16 in dense performance, so 750,000 Ascend 950PRs is 375 EFLOPS - that’s how much AI compute China will make without US chips (I’m assuming the vast majority of their AI chip production is Huawei, obviously there will be some from other companies but likely a much smaller number). As you noted, each Nvidia H200 is 989 TFLOPS FP16 in dense performance, so 750,000 H200s is 742 EFLOPS - that’s how much AI compute China would add from this US sale alone. 742 + 375 = 1,117 EFLOPs - that’s how much compute China would bring online total with domestic production and this US sale. 1,117 / 375 = 2.979x - that’s the multiple that this sale would increase the amount of AI compute that China brings online each year. I rounded up from 2.979 to 3. Next time before you make personal accusations about how much I have read and studied these issues, please at least go to the effort to read and listen to what I said a little more closely, to make sure you’re actually understanding my argument.

2026.05.16 19:34

What I said was it would triple the amount of AI compute China will bring online this year (not including illegally smuggled chips). The math on this checks out: Per public reports, China will make 750,000 Huawei Ascend 950PRs this year. Each 950PR is 500 TFLOPS FP16 in dense performance, so 750,000 Ascend 950PRs is 375 EFLOPS - that’s how much AI compute China will make without US chips (I’m assuming the vast majority of their AI chip production is Huawei, obviously there will be some from other companies but likely a much smaller number). As you noted, each Nvidia H200 is 989 TFLOPS FP16 in dense performance, so 750,000 H200s is 742 EFLOPS - that’s how much AI compute China would add from this US sale alone. 742 + 375 = 1,117 EFLOPs - that’s how much compute China would bring online total with domestic production and this US sale. 1,117 / 375 = 2.979x - that’s the multiple that this sale would increase the amount of AI compute that China brings online each year. I rounded up from 2.979 to 3. Next time before you make personal accusations about how much I have read and studied these issues, please at least go to the effort to read and listen to what I said a little more closely, to make sure you’re actually understanding my argument.

tphuang@tphuang

2026.05.16 00:13

Reminder that Chris McGuire thinks "Ascend950 is worse than 910C" Just using pure compute #: Each H200 is 989 TFLOPS FP16 in dense performance. 750k -> 742 EFLOPS At end of 2025, China admitted to having 1590 EFLOPS compute in operation. There were 117 operating AIDC & 430 under construction + 480 in advanced planning as of 2026/2, so China's smart compute capacity will easily 2x over this yr & likely more than 2x in 2027. I'm likely underestimating China here. So, there is a reason they have been hesitating on buy just 750k H200s. But according to Chris, these H200 will triple China's AI computing power capacity. I doubt he has even looked into & read how much compute China has. Because if he did, he would have read what I posted by now. If China was desperate for H200, it would have allowed AI labs to buy them by now. it's taking a cautious approach bc it's not desperate. It can wait & make sure the AI labs fully support domestic chip options before taking Jensen's leftover stash.