Retail traders' retreat from individual stocks is running into a very different message from Wall Street's analysts: some of the hardest-hit AI companies are starting to look cheap. Goldman Sachs analyst James Schneider identified Nvidia, Broadcom and AMD as buying opportunities in research reported on November 10.
That call follows a sharp break in the usual retail habit of buying market declines. Vanda Research's preliminary estimate put net selling of individual U.S. stocks at $338 million during the November 5 rout. Later reporting, citing updated figures, put the closing outflow at $151 million after buyers returned near the bell.
A selling wave that changed before the close
The distinction matters because the larger figure became the headline measure of retail capitulation. The initial estimate pointed toward the heaviest net selling since the pandemic crash, when individual investors sold $502 million in stocks on March 17, 2020.
But more than $100 million in buying during the final 10 minutes reduced the November 5 outflow, according to updated figures reported the following day. The $338 million reading describes the selling wave earlier in the session, rather than the final closing total in that account.
These are net flows: stock purchases are subtracted from sales. They measure the balance of trading in individual names, rather than the total value of shares sold or losses in investors' accounts.
The provider tracks retail flows in U.S. equities and exchange-traded funds through its VandaTrack service. Those categories also help explain why selling individual companies didn't mean every retail trader had abandoned bullish bets.
Some continued buying leveraged funds, including ProShares UltraPro QQQ and Direxion Daily Semiconductor Bull 3X Shares, during the selloff. That left a split within retail itself: money was leaving individual stocks while some buyers still wanted amplified exposure to technology and chips.
Wall Street starts looking at valuations
Schneider's November 10 assessment offers the bargain-hunting argument. In the research, reported by Yahoo Finance, the analyst wrote:
We believe stocks are now clearly in oversold territory.
His case rests on what investors are paying for future earnings after the decline. Even assuming no growth in 2028, a scenario he considered relatively unlikely, he put Nvidia and Broadcom at about 12 and 13 times the earnings baseline used in his analysis.
Those multiples belong to that scenario, rather than a claim about the companies' current reported earnings. He also favored AMD's growth prospects, while describing its valuation as less compelling.
The bank's research identifies two practical questions behind the semiconductor selloff. One is how large cloud companies and AI developers will finance spending that exceeds their free cash flow. The other is whether enough land, electricity and completed data centers will be available to support expected chip demand.
Schneider linked greater uncertainty about construction, including in Texas, to the outcome of the U.S. elections. He also acknowledged that attractive valuations might not produce an immediate rebound:
[Although] we recognize there may be few catalysts for stocks to move higher before year end, we think risk/reward is compelling for select names at current levels.
The bullish bets still carry leverage
The ProShares fund targets three times the Nasdaq-100's daily performance before fees and expenses. Its objective resets daily, so holding it for longer can produce returns substantially above or below three times the index's change over the full holding period.
Shares can be bought and sold through a brokerage account. The issuer says volatility and the length of the holding period can increase that difference. That makes the choice between buying a semiconductor stock and buying a leveraged technology fund consequential, even when both trades express a bullish view.