Institutional investors — the so-called smart money — appear to be nibbling on the beat-up stock price of Nvidia (NVDA).

Retail purchases of the AI market darling have plunged this month, according to new data from Vanda Research. On Monday, retail investors only purchased a net $34 million in Nvidia's stock, based on Vanda's research.

Retail investors have also begun "dumping" positions in the 2X Bull Nvidia ETF.

Yet, Nvidia shares have started to come back up — they are up 5% in February compared to a slight decline for the S&P 500 (^GSPC). The stock is up 13% from its Feb. 3 closing lows.

Shares are still down 2.5% year to date.

"By process of elimination, if retail's stepping aside, the 'smart money' crowd must be behind the latest recovery [in Nvidia]. However, institutional option flows in the name have yet to rebound strongly, which is something that would give us greater confidence around this view," Vanda Research said.

"Still, we believe we can take the above as a vote of confidence by parts of the institutional world in the stock and the sector now that analysts have had time to better digest the potential ramifications of DeepSeek's launch. The verdict from these latest developments? DeepSeek doesn't appear to threaten AI's growth trajectory."

The Nvidia bulls have been dealing with negative sentiment as of late, somewhat of a rarity for one of the most popular stocks in the world.

Evercore analyst Mark Lipacis said in a recent note there are three reasons for the weakness: 1) DeepSeek lowering AI demand in aggregate, 2) DeepSeek shifting AI compute cycles away from Nvidia GPUs and to ASICs [custom chips], and 3) Blackwell chip delays.

China-based DeepSeek surprised markets in late January after unveiling RI, its AI model that gives a ChatGPT-esque performance at a cheaper price tag. RI cost a reported $5.6 million to build a base model, compared with the hundreds of millions of dollars incurred at US-based companies such as OpenAI and Anthropic.

Fears mounted instantly that US companies are overspending on AI infrastructure, which includes Nvidia chips.

"Conventional wisdom all of last year was that training amazing models was going to be possible for only a handful of companies," Snowflake CEO Sridhar Ramaswamy told me on Yahoo Finance's Opening Bid podcast (listen below). "What DeepSeek has done over the past few weeks is shatter that belief by saying they can train a model for $6 million."

The Street has moved to defend Nvidia heading into earnings on Feb. 26, underscoring the view of the smart money rotating into the name.