I am curious what workloads Citadel Securities is running on these TPUs? Are you telling me they need the latest TPUs for market insights?

▲

vibe42 5 hours ago | parent [-]

Training their own, closed, internal models on their own data sets? Probably a good way to squeeze out some market trading signals.

▲

nickandbro 5 hours ago | parent | next [-]

Reminds me of when hedge funds started laying increasingly shorter fiber-optic cable lines to achieve the lowest possible latency for high-frequency trading.

▲

written-beyond 5 hours ago | parent | prev [-]

I thought these TPUs were primarily used for inference?

	▲	vlovich123 4 hours ago \| parent \| next [-]
		TPU8t is for training. But even still, once you’ve trained, you need to run the model too. And these kinds of models already have a huge latency hit so there’s not much hurting running it away from the trading switches.
	▲	knowaveragejoe 4 hours ago \| parent \| prev [-]
		As the article states, there's both training and inference dedicated chips.