top of page
Exponential Logo.png

The Tech Selling Came Back Friday. The Power Bid Did Not Move.

6 days ago
3 min read

Thursday's tape — the flattest Nvidia session of the month on the day Jensen Huang's resignation landed — now reads as a pause, not a turn. Friday institutions came back to selling: −$2.90 billion net of Nvidia (buys minus sells), on top of −$17.5 billion of S&P 500 Information Technology distribution for the week. And the other side of the AI trade held again: S&P 500 Utilities were net bought +$1.04 billion on the same close, the fifth consecutive session of institutional accumulation. The AI-power rotation did not blink.

The Week in One Chart


XTech US Equity Flow classifies every trade in S&P 500 names into institutional buyers and sellers at each daily close — direction measured from the transactions themselves, not inferred from the price bar. Information Technology versus Utilities, institutional net flow, $ billions:

  • Monday Nov 9: IT −$5.47 · Utilities +$0.45


  • Tuesday Nov 10: IT −$4.60 · Utilities +$1.27


  • Wednesday Nov 11: IT −$2.67 · Utilities +$1.14


  • Thursday Nov 12: IT −$0.17 · Utilities +$0.72


  • Friday Nov 13: IT −$4.58 · Utilities +$1.04


The week totals: Information Technology −$17.51 billion institutional net; Utilities +$4.61 billion net bought. Thursday — the resignation announcement, the $116 billion Nvidia tape — was the week's quietest distribution day, and Friday resumed the pattern. Nvidia closed below a $4 trillion market cap Friday (WSJ) as chip stocks sold off; the averages still notched a winning week (CNBC, Yahoo Finance).

S&P 500 institutional net flow, Information Technology versus Utilities, five sessions through Friday November 13, $ billions

The single names inside the IT complex show the same shape the sector totals do. Nvidia's institutional net flow: −$1.89 billion Monday, −$3.22 billion Tuesday, −$1.31 billion Wednesday, flat Thursday, −$2.90 billion Friday. AMD: net bought again Friday (+$0.43 billion, its third net-buy day of the week's five) and Broadcom +$0.32 billion, its second straight. The money leaving the leader keeps landing in the two rivals the press has named rotation winners — three sessions of AVGO buying and an AMD majority across the week, through the leader's worst tape of the year.

The Rotation Is a Two-Name Story on the Chip Side


WSJ's Friday live blog frames next week's tests: Nvidia's earnings and Anthropic's IPO. The flow record through Friday says institutions have not waited for either.

Meanwhile, the Power Complex: Five Straight Sessions


The debate over whether AI data centers actually get built keeps being answered from the money side. Utilities have been net bought every session this week — +$4.61 billion across five days — while the tech sector distributed $17.5 billion. These are the same institutions, on the same closes, expressing opposite conviction in the same theme: fund the power the data centers run on, trim the chips. Nobody on a price chart can see that split; it takes classified buyers and sellers at every close.

Why Classified Daily Flow Was Necessary to See This


The press narrative for the week is a leadership story with a $4 trillion milestone. The transaction record shows a different shape: a distribution-and-accumulation pattern that held through every session of it, a pause on announcement day that Friday's tape proved was a pause, and a five-session utilities bid that never wavered. None of that is visible on a price chart, because price cannot say who traded. Group attribution — who did what, each close — is only measurable when buyers and sellers are classified separately every session.

The next section explains where these numbers come from.

Data Source: XTech US Equity Flow (Sector and Single-Name)


This analysis draws on XTech classified flow — institutional and retail order flow measured directly from the tape, not inferred from price. [Form block intentionally omitted: no campaign form id is registered for an equity daily post. To be added when the campaign white paper lands.]

Comments


Read the research,
then test the data.

Request a white paper from our research library, or book a call with the research team behind it.

bottom of page