Cloud providers are out of B200 GPUs after the Inkling release (twitter.com)

I'm not sure I buy the direct causation? Why didn't GLM 5.2 have that effect? It's a better model, by some measures, and a little easier to run as it's quite a bit smaller.