SMH Is Up While Micron and SanDisk Are Down: Just Opening Gaps, or Weakness After the Open?

Short answer: daily performance and performance since the open answer different questions

At generation, Micron (MU) was -1.96% versus its previous close and SanDisk (SNDK) was -0.30%, while SMH was +0.71%. Using each security's opening price instead, MU and SNDK remained negative while SMH remained positive. Their opposing directions were not simply a legacy of gaps between the previous close and the open. [1; our calculations]

The additional finding is that MU and SNDK did not merely open lower. Their open-to-snapshot returns still lagged SMH by 1.37 and 0.50 percentage points, respectively. This supports a narrow description of relative weakness after the open, not a conclusion about capital outflows, deteriorating industry fundamentals or future prices. [1; our calculations]

Generated and quotes retrieved as of 2026-10-01 22:39:22 China Standard Time (UTC+8), equivalent to 2026-10-01 10:39:22 EDT. This is an intraday observation, not a closing assessment. The timestamp records completion of this API retrieval, not a synchronized last-trade time for every security. Prices continue to change.

Which numbers use a comparable interval?

All quotes come from this retrieval of Tiger live data. Each previous-close denominator is the September 30, 2026 US close and was cross-checked against Tiger historical daily bars. Prices are in US dollars. [1]

All returns use unrounded API inputs before display to two decimals. Displayed prices have lower precision, so calculations from the printed table may differ slightly. “Since open” measures two endpoints; it does not claim that prices fell continuously between them. [1; our calculations]

Why not simply add the two percentage changes?

Let C be the previous close, O the opening price and P the snapshot price. The opening gap is g = O/C − 1; the return since the open is s = P/O − 1; and the daily return is d = P/C − 1. Expressed as decimal returns, the exact identity is:

d = (1 + g) × (1 + s) − 1 = g + s + g × s.

When the inputs are expressed in percent units, the interaction term is g × s / 100. For MU at this snapshot:

-1.960% = -1.011% + -0.959% + an interaction of +0.010 percentage points.

This separates return intervals; it does not identify a news catalyst. The opening gap describes the change between the previous close and the open. It cannot, by itself, identify the responsible headline, investor category or order. [1; our calculations]

What does the decomposition add?

1. SMH's daily gain covers two intervals, not just the period since the open

SMH's opening gap was +0.29% and its subsequent return was +0.41%, combining into a daily return of +0.71%. Both intervals were positive in this snapshot, but the daily gain was not the gain since the open. Calculating both intervals identifies where relative differences appeared; the daily figure alone does not. [1; our calculations]

2. MU and SNDK's relative weakness did not occur only before the open

MU's open-to-snapshot return minus SMH's corresponding return was -1.37 percentage points. SNDK's was -0.50 percentage points. Both differences remained negative after excluding each security's opening gap. The observed relative weakness therefore cannot be attributed entirely to lower openings. These are descriptive return spreads, not risk-adjusted alpha or measures of capital flows. [1; our calculations]

3. QQQ illustrates another starting-point effect, not a whole-market verdict

QQQ had an opening gap of +0.37%, a return since the open of -0.28%, and a daily return of +0.09%. Its positive opening gap and negative subsequent return partially offset each other. Four observations cannot establish advance/decline breadth across the US market. [1; our calculations]

What should readers check next?

• Recalculate both intervals after the close. Keep the same securities and definitions, then check whether the signs and relative ordering persist. Do not relabel this intraday assessment as a full-session result.
• Identify the starting point in a headline. Above the previous close, above the open and up from the intraday low are three different comparisons, each requiring its own denominator.
• Expand the evidence before expanding the claim. This article lacks comparable same-interval observations for Samsung Electronics and SK hynix. It is not a global memory-industry assessment, and it does not combine Korean closes or older Hong Kong sessions with US intraday returns.
• Separate price descriptions from causal claims. Today's news catalysts, investor-category net transactions and earnings-estimate revisions were not verified for this article. Missing explanations are not confirmed facts.

Frequently asked questions

Does this establish that the semiconductor sector has turned weaker?

No. It describes returns between specified endpoints for four observations, and the relative performance of two stocks against SMH. It contains no full-constituent breadth count, historical percentile or trend test.

Is this a trading signal?

No. It is a method for testing whether a market description matches its numbers. It contains no holdings, trading instructions or promised returns.

Source and scope

[1] Tiger OpenAPI: Get Stock Briefs and Get Bars. Numerical observations were retrieved through actual API calls for this article; AlphaGBM calculated the returns and spreads. The public documentation explains fields rather than independently reproducing this live snapshot. Raw responses and calculation checks were retained locally.

The daily research log supplied the editorial starting point. Its US observations referred to an earlier settled session and were not reused as current intraday numbers. This article does not claim a premarket-to-intraday reversal; it compares only the explicitly identified price endpoints.