Skip to main content
OnTheRiceSEE THE WORLD EARLY.
OnTheRice’s Emerging DataInvestment

Actual chip sales face a fresh profit test

By OnTheRice · Sources and limitations · Report a correction

BULLISHHorizon1 week, adjusted to the trading calendar
Probability 56.0%Subjective estimate
Research referenceUSD 453.31 per ADR. Latest verified observation: 10 October 2026, 04:00 SGT / 9 October, 16:00 EDT. StockAnalysis / S&P Global regular-session raw USD close. The regular market is closed at preparation time; this is a closing reference, not a fresh Saturday trade.
24-hour change

Research reference · source time USD 453.31 per ADR. Latest verified observation: 10 October 2026, 04:00 SGT / 9 October, 16:00 EDT. StockAnalysis / S&P Global regular-session raw USD close. The regular market is closed at preparation time; this is a closing reference, not a fresh Saturday trade. · not a current quote or execution price · Reference source

Source confidenceModerate for the combined inference. Named company and official disclosures support facts; market providers support prices. Source confidence does not measure the chance of a price move. Missing independent clusters and current minute data limit the forecast.

First published
1 week, adjusted to the trading calendar

What’s changing?

We see a gap between real chip sales and the price investors are willing to pay for them. Across OTR Database Data's 06: 00, 07: 00 and 08: 10 reports, strong foundry output sits beside worries about AI customers' future revenue and financing. We checked the original TSMC filing: September sales rose 54.6% from a year earlier, while falling 0.6% from August. January-September sales rose 41.1%. These are sales already reported, not a new funding promise. The next test comes at TSMC's 15 October earnings call. We favour a modest recovery if profits and guidance show that customers still convert plans into shipments. We choose one week because the call falls inside that period; 24 hours would mostly cover a closed market.

Key points

  • Physical supplier sales remain strong; software monetisation and project financing are separate risks; 15 October results can test how much demand turns into profit. We are not treating repeated reporting of the same sales release as several independent facts.

What could change this view?

  • Weak forward guidance, margins or utilisation would outweigh the reported growth. A renewed broad technology selloff could beat company evidence. A good sales number can still disappoint investors who expected more.

Bottom line

We lean BULLISH over 1 week, adjusted to the trading calendar, measured from USD 453.31 to the matching source price at 17 October 2026, 04:00 SGT / 16 October US closing auction. The estimate is modest and uncalibrated.

Its source facts and prices still need independent checks.

How likely is this call? · Three frameworks

Taiwan Semiconductor Manufacturing ADR (TSM) · BULLISH · Subjective estimate

Directional BULLISH estimate 56.0%Rise scenario > +1% 55.0%Decline scenario < −1% 30.0%Neutral within inclusive ±1% 15.0%

Subjective and uncalibrated; not recalculated by the OTR probability engine. Directional sign probability and thresholded return-scenario weights are separate estimates. These percentages are not observed outcomes or a verified current price.

We apply the shared method in subjective mode. We have no trained posterior, fitted likelihood ratios or validated exact-pattern success rate. The following scenario weights are our current research judgement, not measurements.

Bayesian · what changed the estimate?

We weigh the supporting facts in Evidence behind this view against the contrary facts in Evidence against this view. Already-priced news, stale data and source overlap reduce the estimate. No numerical historical prior or likelihood ratio was fitted.

de Finetti · are the comparisons fair?

We compare the same instrument and price convention at the chosen horizon. Company reports, sector prices and fund flows are different measurements. Reprints and several reports about one announcement are dependent. The six researchers are reviewers, not six independent market experiments.

Kolmogorov · do the numbers agree?

55 + 30 + 15 = 100%. A rise above +1%, a fall below -1%, and a return inside inclusive ±1% form three separate, complete outcomes. The 56% sign estimate lies within its matching directional scenario and neutral-band bounds. Missing evidence is never assigned to neutral.

Read the shared method · Read the public probability method

Evidence behind this view

Checks whether separate pieces of evidence support the same view.

OTR Database Data connects manufacturing revenue, AI monetisation doubt and financing conditions. Primary TSMC figures confirm the manufacturing strand. Nvidia and Broadcom's latest issued results confirm strong realised infrastructure sales as older context, not breaking demand data. A separate GlobalFoundries announcement confirms a multi-year 2 billion-dollar agreement for US silicon interposers used in TSMC packaging. Taiwan's Finance Ministry reports September exports of 87.22 billion dollars, up 60.9% from a year earlier. These support capacity planning and wider trade activity, although national exports are not TSMC-only sales and planned 2028 capacity does not arrive next week. The earnings calendar supplies the near-term test.

What past price moves show

Compares today’s price moves with recorded past patterns.

Verified closes fell from 485.80 on 5 October to 482.30, 472.20, 457.99, and 453.31 on 9 October. Four consecutive falling closes challenge our recovery case. They do not establish a statistically reliable bounce. Complete current-24 h minute bars and calibrated matching analogues were not available in the pattern report.

Evidence against this view

Shows evidence and risks that could challenge this view.

The market has sold the shares despite the good revenue release. The monthly sales decline and missing margin detail matter. Buyers' funding constraints can hit future orders before past sales show weakness.

Why the direction changed

Explains recorded changes from expecting a rise to expecting a fall, or the reverse.

Learning review before publication

The supplied direction was retained: No investible directional call.

The supplied synthesis was retained.

Reason: The supplied draft was retained.

No individual outcome lesson was cited in this review.

This review belongs to the new draft before publication. Earlier published calls and their original deadlines remain unchanged.

What people say vs the facts

Compares recorded public discussion with checked market facts.

AI demand-collapse claims and permanent-growth claims both go beyond one release. We can verify sales and the calendar; we cannot verify future margins before the results.

How past forecasts performed

Shows recorded forecast results and what they can teach us.

The latest pattern audit provides no calibrated matching forecast sample for this instrument and horizon. We do not use the site's overall claimed hit rate as TSM's probability.

No applied outcome lesson is claimed for this card.

PENDINGPRICE UNVERIFIEDHistorical record
Sources, original wording and price checks

Author-submitted research. The underlying facts and market prices have not been independently verified.

Original research record

How likely is this call? · Three frameworks

We apply the shared method in subjective mode. We have no trained posterior, fitted likelihood ratios or validated exact-pattern success rate. The following scenario weights are our current research judgement, not measurements.

Bayesian · what changed the estimate?

We weigh the supporting facts in Evidence behind this view against the contrary facts in Evidence against this view. Already-priced news, stale data and source overlap reduce the estimate. No numerical historical prior or likelihood ratio was fitted.

de Finetti · are the comparisons fair?

We compare the same instrument and price convention at the chosen horizon. Company reports, sector prices and fund flows are different measurements. Reprints and several reports about one announcement are dependent. The six researchers are reviewers, not six independent market experiments.

Kolmogorov · do the numbers agree?

55 + 30 + 15 = 100%. A rise above +1%, a fall below -1%, and a return inside inclusive ±1% form three separate, complete outcomes. The 56% sign estimate lies within its matching directional scenario and neutral-band bounds. Missing evidence is never assigned to neutral.

Evidence behind this view

OTR Database Data connects manufacturing revenue, AI monetisation doubt and financing conditions. Primary TSMC figures confirm the manufacturing strand. Nvidia and Broadcom's latest issued results confirm strong realised infrastructure sales as older context, not breaking demand data. A separate GlobalFoundries announcement confirms a multi-year 2 billion-dollar agreement for US silicon interposers used in TSMC packaging. Taiwan's Finance Ministry reports September exports of 87.22 billion dollars, up 60.9% from a year earlier. These support capacity planning and wider trade activity, although national exports are not TSMC-only sales and planned 2028 capacity does not arrive next week. The earnings calendar supplies the near-term test.

What past price moves show

Verified closes fell from 485.80 on 5 October to 482.30, 472.20, 457.99, and 453.31 on 9 October. Four consecutive falling closes challenge our recovery case. They do not establish a statistically reliable bounce. Complete current-24 h minute bars and calibrated matching analogues were not available in the pattern report.

Evidence against this view

The market has sold the shares despite the good revenue release. The monthly sales decline and missing margin detail matter. Buyers' funding constraints can hit future orders before past sales show weakness.

Why the direction changed

This is a new research source call. We have not verified a prior locked TSM call to claim a reversal. Our view connects current supplier evidence with the forthcoming earnings test.

What people say vs the facts

AI demand-collapse claims and permanent-growth claims both go beyond one release. We can verify sales and the calendar; we cannot verify future margins before the results.

How past forecasts performed

The latest pattern audit provides no calibrated matching forecast sample for this instrument and horizon. We do not use the site's overall claimed hit rate as TSM's probability.

First published through Sites

1 week, adjusted to the trading calendar

  • Taiwan Semiconductor Manufacturing ADR (TSM)BULLISH ◷ PENDING
    Research reference: USD 453.31 per ADR. Latest verified observation: 10 October 2026, 04:00 SGT / 9 October, 16:00 EDT. StockAnalysis / S&P Global regular-session raw USD close. The regular market is closed at preparation time; this is a closing reference, not a fresh Saturday trade. · source observation 10 Oct 2026, 04:00:00 SGT. Not a fresh quote or execution price.First published through Sites: 10 Oct 2026, 11:57:49 SGT
    Call and evidence

    BULLISH

    The source-authored research source estimate is retained without engine recalculation. A comparable source-matched endpoint assessment is required at its explicit calendar-adjusted expiry.

    Research reference: USD 453.31 per ADR. Latest verified observation: 10 October 2026, 04:00 SGT / 9 October, 16:00 EDT. StockAnalysis / S&P Global regular-session raw USD close. The regular market is closed at preparation time; this is a closing reference, not a fresh Saturday trade. · Taiwan Semiconductor Manufacturing ADR (TSM) · Not specified in research source.

    The retained quote is a historical reference from the recorded source material at 10 Oct 2026, 04:00:00 SGT. It is not a fresh current quote, nor an issue-time execution price. Source-authored estimates are displayed separately and were not recalculated by the OTR probability engine.

    Open the cited market reference ↗

    Source expiry: 17 Oct 2026, 04:00:00 SGT · 1 week, adjusted to the trading calendar. A comparable source-matched endpoint is required; this imported forecast is not scored by the generic sign-only ledger.

What changed

Recorded public history

We see a gap between real chip sales and the price investors are willing to pay for them. Across OTR Database Data's 06: 00, 07: 00 and…

This is the first recorded public view. Further public updates will appear here.

  1. Actual chip sales face a fresh profit test

    We see a gap between real chip sales and the price investors are willing to pay for them. Across OTR Database Data's 06: 00, 07: 00 and 08: 10 reports, strong foundry output sits beside worries about AI customers' future revenue and financing. We checked the original TSMC filing: September sales rose 54.6% from a year earlier, while falling 0.6% from August. January-September sales rose 41.1%. These are sales already reported, not a new funding promise. The next test comes at TSMC's 15 October earnings call. We favour a modest recovery if profits and guidance show that customers still convert plans into shipments. We choose one week because the call falls inside that period; 24 hours would mostly cover a closed market.

    Forecast prepared: 10 Oct, 11:00 SGTPublished by OTR: 10 Oct 2026, 11:57 SGT

Why the view changed

This is the first recorded view. Future source-backed updates will appear here.

Evidence, memory & review 6 linked sources

Named assets or actors: Taiwan Semiconductor Manufacturing ADR (TSM) (TSM)

Directional takeaway: BULLISH, 1 week, adjusted to the trading calendar

OTR reasoning: We see a gap between real chip sales and the price investors are willing to pay for them. Across OTR Database Data's 06: 00, 07: 00 and 08: 10 reports, strong foundry output sits beside worries about AI customers' future revenue and financing. We checked the original TSMC filing: September sales rose 54.6% from a year earlier, while falling 0.6% from August. January-September sales rose 41.1%. These are sales already reported, not a new funding promise. The next test comes at TSMC's 15 October earnings call. We favour a modest recovery if profits and guidance show that customers still convert plans into shipments. We choose one week because the call falls inside that period; 24 hours would mostly cover a closed market.

Counterpoint: Weak forward guidance, margins or utilisation would outweigh the reported growth. A renewed broad technology selloff could beat company evidence. A good sales number can still disappoint investors who expected more.

Source trail

  • pr.cld.tsmc.com ↗The research source reports this source; source-clock verification is not claimed.
  • www.sec.gov ↗The research source reports this source; source-clock verification is not claimed.
  • investor.tsmc.com ↗The research source reports this source; source-clock verification is not claimed.
  • stockanalysis.com ↗The research source reports this source; source-clock verification is not claimed.
  • gf.com ↗The research source reports this source; source-clock verification is not claimed.
  • www.mof.gov.tw ↗The research source reports this source; source-clock verification is not claimed.

Review status: LEARNING REVIEWED WITH DISCLOSED LIMITS

No score or probability is presented.

These records are in the same category and cite at least one of the same source documents. Shared evidence is not independent confirmation; each record keeps its own date and conclusion.