Nvidia answered the crypto bust by pushing data-center AI and posted record revenue

In 2018 Nvidia looked unstoppable, then the crypto bubble burst and flooded the market with used GPUs just as data-center buying slowed. Its stock lost more than half its value and it logged several quarters of year-over-year sales and profit declines. Jensen Huang's answer was not to wait for the next mining wave but to push the same silicon into a durable market: accelerated computing and AI.

Nvidia kept building the data-center and AI story, from recommendation engines for Alibaba and Baidu to GPU-accelerated Arm servers and tools like TensorRT for inference. By the fourth quarter of fiscal 2020 (ending January 2020) the bet showed up: revenue grew 41% to $3.11bn, gaming chips rose 56% to $1.49bn, and data-center revenue set a record at $968m, up 43%.

It ended a run of four consecutive quarters of year-over-year declines and the stock more than doubled over the following year. The crypto revenue was gone, but Nvidia now had an AI-driven engine instead.

Q4 FY20 revenue $3.1B up 41%, gaming chips up 56% and data-center at a record $968M, up 43%.

Crypto mining revenue was volatile and had already flooded the GPU market, so Nvidia stopped depending on it.

The same GPUs were ideal for AI training and inference, a growing and non-speculative workload.

Nvidia invested in software and system solutions, not just chips, to lock in data-center buyers.

When hyperscale spending returned, Nvidia was the default platform and set records.

A platform company that rides a fad survives by converting the same silicon into a durable, non-speculative market — accelerated computing.

The data-center and AI franchise became Nvidia's dominant business through the 2020s, driving the company to become one of the world's most valuable.