My guess: too many DC resources dedicated to AI, leaving insufficient free capacity to handle traffic spikes

I laughed then realized this is plausible

	▲	avisser 11 hours ago \| parent \| next [-]
		Seems unlikely to me. Different hardware requirements. AI workloads aren't running on your standard 2U server.
	▲	cozzyd a day ago \| parent \| prev [-]
		And then they're probably trying to debug with their AI which isn't helping.