The Dow Rose, Semis Fell 2.4%, and Everything Speculative Was Sold

The Dow rose 0.27% on Monday. Semiconductors fell 2.43%. Between those two numbers sits the cleanest single-session sort by speculativeness this book has seen all month: eleven of the least-established names in the tracked universe averaged -4.42%, ten of the most-established averaged +0.81%, and almost nothing landed in between.

01Two indices, opposite directions

The Dow closing higher while the Nasdaq-100 fell 1.00% and semiconductors fell 2.43% is not a risk-off day in the usual sense. In a genuine risk-off session everything falls and Treasuries rally. Here gold rose 0.79%, long Treasuries rose 0.62%, and volatility was flat at +0.05% — the fear gauge barely moved while the speculative end of the market lost the better part of a tenth of its value.

That combination points at a rotation with a specific sorting rule rather than a scare. The market was not reducing exposure. It was changing what kind of company it wanted to own.

Figure 1 · Friday against Monday

Friday's move told you almost nothing about Monday's

Reading this figure

1. The lower band is almost entirely speculative names, regardless of where they sat on Friday.
2. The upper band is almost entirely mega-cap: META, MSFT, GOOGL, AMZN.
3. Tempus AI ran +8.8% Friday and fell 9.0% Monday — the sharpest reversal on the board.
4. A vertical sort with no horizontal pattern means Monday used a different rule from Friday.

02What the sorting rule actually was

“Risk” is too vague to be useful here. The names that fell hardest share something more specific than volatility: the distance between what they earn today and what their price assumes they will earn later.

1. Start with what each name earns now

   Meta, Microsoft, Alphabet and Amazon all generate large current profits. Tempus AI, IonQ, Applied Digital and Rocket Lab are priced on revenue that mostly does not exist yet. That is the axis, and it is not the same as market capitalisation or sector.

2. A leveraged fund doubles whatever happens to its underlying

   NBIL is a 2× daily-reset fund on Nebius. Nebius fell 3.75% and NBIL fell 7.49% — almost exactly double. That is mechanical, not a separate opinion about Nebius.

   NBIS −3.75% → NBIL −7.49%

3. Memory got sorted with the speculative names, not with semis

   This is the detail worth keeping. SanDisk and Micron fell harder than the semiconductor index because the thing holding them up was a policy report nobody had confirmed — a future-earnings story, not a current-earnings one.

   SNDK −6.45%, MU −5.83% vs SMH −2.43%

4. So the Dow rises while the Nasdaq falls

   The Dow is weighted toward companies with current earnings. Nothing about Monday required a view on the economy; it required a preference for profit that already exists.

   DIA +0.27% vs QQQ −1.00%

Figure 2 · The session ranked

The nine worst and five best, and the pattern in the names

TEM−8.97%

IONQ−8.47%

NBIL−7.49%

SNDK−6.45%

RKLB−5.91%

MU−5.83%

SMCI−5.56%

SYM−4.30%

DDOG−4.18%

EL+2.15%

META+1.66%

AMZN+1.33%

CAKE+1.13%

GOOGL+0.94%

Live chart — both ends of the sort

TradingView, three-month daily. Switch symbols with the tabs.

SNDKTEMMETASMH

Live third-party data, independent of the fixed figures above. For context, not recommendations.

03The macro tape

04Positions

Prices via Massive Market Data daily bars. Sleeve membership is my own classification and equal-weighted; it describes the names, not the money. This article is educational and reflects my own analysis; it is not investment advice.