TSMC's capacity additions of 10% to 20% have been achieved by purchasing extra tools rather than building new cleanrooms since July 2026.

Verdict: likely FALSE (80%)

Web Report: https://solsice.com/public/debates/tsmc-s-capacity-additions-of-10-to-20-have-been-achieved-by-cc302d078ca6

---

This section provides a brief overview of the key arguments. You do not need to read the full detailed report below.

✅ Key PRO arguments:

1. ■TSMC [28]'s quarterly capital expenditures shifted from roughly 40% shell construction [23] / 60% equipment toward a 15% / 85% split in H2 [15] 2026, meaning 85 cents of every new-capacity dollar purchased lithography [16], etching [10], and deposition [6] tools, yielding a 12% increase in wafer [29] output across N3 [18] and N5 [19] nodes by Q1 2027.
2. ■Fab [12] 18 Phase 8 at the Tainan Science Park was a 45,000-square-meter shell-complete cleanroom [2] finished in Q1 2026 and explicitly designated by TSMC [28] as 'shell-ready [24] capacity reserve' to be tooled up 'on demand,' demonstrating that capacity additions came from populating existing shells with tools rather than new construction.

❌ Key ANTI arguments:

1. ■The affirmative commits a temporal fallacy by citing Q3 2026, Q4 2026, and Q1 2027 earnings and analyst calls to prove a trend that supposedly began in July 2026; describing future events in the past tense to justify a current capacity state amounts to presenting projections as established history.
2. ■Increased tool shipments [27] to TSMC [28] are logically compatible with both existing-shell expansion and newly commissioned cleanroom [2] construction, since tools are a prerequisite for capacity expansion in either case; using supplier shipment data to prove the absence of new cleanroom construction is an attempt to prove a negative and is therefore inconclusive.

💭 Conclusion: False. The affirmative's central evidence—a 15%/85% capex [1] reallocation—relies on Q3 2026, Q4 2026, and Q1 2027 disclosures used to characterize a trend supposedly active since July 2026, which the negative correctly identified as a temporal fallacy. The Fab [12] 18 Phase 8 example, while illustrative of TSMC [28]'s shell-reserve strategy, is a single instance and does not establish that all 10–20% of capacity additions since July 2026 were achieved without any new cleanroom [2] construction. Because the compound claim requires both the magnitude (10–20%) and the mechanism (tools-only, no new cleanrooms) to be independently supported, and the temporal-flawed sourcing weakens both sub-claims, the verdict is FALSE with moderate confidence.

---

🔬 DeepResearch Result: likely FALSE ❌ (80%)

Assertion: TSMC [28]'s capacity additions of 10% to 20% have been achieved by purchasing extra tools rather than building new cleanrooms since July 2026.

Participating models: deepseek-v4-pro 💬, gemma-4-26b-a4b-it 💬👁️

📊 Face-offs: 0 concluded TRUE, 1 concluded FALSE (1 face-offs, 3 models)
📊 Score by side: TRUE=0.00, FALSE=0.80

🏅 Model score changes:
minimax-m3 💬👁️: +8

✅ PRO Arguments:

1. ■TSMC's quarterly capital expenditures shifted from roughly 40% shell construction [23] / 60% equipment toward a 15% / 85% split in H2 [15] 2026, meaning 85 cents of every new-capacity dollar purchased lithography [16], etching [10], and deposition [6] tools, yielding a 12% increase in wafer [29] output across N3 [18] and N5 [19] nodes by Q1 2027. deepseek-v4-pro 💬
2. ■Fab [12] 18 Phase 8 at the Tainan Science Park was a 45,000-square-meter shell-complete cleanroom [2] finished in Q1 2026 and explicitly designated by TSMC as 'shell-ready [24] capacity reserve' to be tooled up 'on demand,' demonstrating that capacity additions came from populating existing shells with tools rather than new construction. deepseek-v4-pro 💬

❌ ANTI Arguments:

1. ■The affirmative commits a temporal fallacy by citing Q3 2026, Q4 2026, and Q1 2027 earnings and analyst calls to prove a trend that supposedly began in July 2026; describing future events in the past tense to justify a current capacity state amounts to presenting projections as established history. gemma-4-26b-a4b-it 💬👁️
2. ■Increased tool shipments [27] to TSMC are logically compatible with both existing-shell expansion and newly commissioned cleanroom construction, since tools are a prerequisite for capacity expansion in either case; using supplier shipment data to prove the absence of new cleanroom construction is an attempt to prove a negative and is therefore inconclusive. gemma-4-26b-a4b-it 💬👁️

💭 Reasoning: False. The affirmative's central evidence—a 15%/85% capex [1] reallocation—relies on Q3 2026, Q4 2026, and Q1 2027 disclosures used to characterize a trend supposedly active since July 2026, which the negative correctly identified as a temporal fallacy. The Fab 18 [13] Phase 8 example, while illustrative of TSMC's shell-reserve strategy, is a single instance and does not establish that all 10–20% of capacity additions since July 2026 were achieved without any new cleanroom construction. Because the compound claim requires both the magnitude (10–20%) and the mechanism (tools-only, no new cleanrooms) to be independently supported, and the temporal-flawed sourcing weakens both sub-claims, the verdict is FALSE with moderate confidence.

📋 PRO Facts:
• TSMC's H2 2026 quarterly capex allocation shifted to approximately 15% shell construction and 85% equipment, per the Q3 2026 earnings presentation [9].
• Fab 18 Phase 8 at the Tainan Science Park is a 45,000-square-meter shell-complete cleanroom structurally finished in Q1 2026 and designated 'shell-ready capacity reserve.'

📋 ANTI Facts:
• The affirmative's primary evidence (Q3 2026, Q4 2026, and Q1 2027 earnings and analyst calls) post-dates the July 2026 reference point of the assertion.
• Increased tool shipments to TSMC are a prerequisite for capacity expansion regardless of whether expansion occurs in existing shells or newly built cleanrooms.

The following technical terms, abbreviations, and domain-specific concepts are referenced throughout this debate transcript. Numbers in square brackets [N] in the text above link to the corresponding entry below.

[1] CapEx — Capital Expenditure — Funds used by a company to acquire, upgrade, or maintain physical assets such as buildings and equipment; in the transcript, used to describe TSMC's spending split between shell construction and equipment.

[2] cleanroom — A controlled-environment facility with filtered air and strict particulate controls used for semiconductor manufacturing; the transcript contrasts building new cleanrooms with populating existing ones.

[3] commissioning — The process of testing, validating, and bringing equipment or a facility into operational service; referenced in the debate as a step that allegedly did not occur for new cleanrooms.

[4] customer pull — A demand-driven production approach in which capacity is added in response to specific customer orders rather than speculative supply; cited as the rationale for TSMC's tool move-ins.

[5] densification — Increasing the amount of equipment (tools) installed within an existing cleanroom shell to raise output without expanding the physical footprint.

[6] deposition — A semiconductor manufacturing process that deposits thin films of material onto a wafer; one of the tool categories TSMC allegedly purchased to expand capacity.

[7] deposition systems — Manufacturing tools used to deposit thin-film layers on semiconductor wafers; cited as part of the equipment installed at Fab 18 Phase 8.

[8] earnings call — A quarterly conference call in which a public company's management discusses financial results with analysts and investors; cited as the source for several quoted statements.

[9] earnings presentation — The slide deck accompanying a company's quarterly earnings call, typically containing financial metrics and operational disclosures; cited as the source for capital expenditure splits.

[10] etching — A semiconductor manufacturing process that selectively removes material from the wafer surface using chemical or plasma methods; one of the tool categories cited.

[11] etching chambers — Enclosed process modules in which etching is performed on wafers; cited as a unit of equipment count (e.g., 32 chambers installed at Fab 18 Phase 8).

[12] fab — fabrication facility — A semiconductor manufacturing plant where wafers are processed into chips; the transcript references multiple TSMC fabs.

[13] Fab 18 — A specific TSMC fabrication facility located at the Tainan Science Park in Taiwan; cited as a site of shell-ready capacity and tool move-ins.

[14] Fab 22 — A specific TSMC fabrication facility referenced in the debate as a site of pre-existing shell-ready cleanroom space.

[15] H2 — second half (of the year) — The six-month period from July through December; used in the context of capacity guidance for H2 2026.

[16] lithography — A semiconductor patterning process that uses light (or other radiation) to transfer circuit patterns onto a wafer; one of the tool categories cited.

[17] lithography clusters — Grouped sets of lithography tools (typically EUV or DUV scanners and associated track equipment) operated as a unit; cited as a count of installed equipment.

[18] N3 — 3-nanometer process node — TSMC's 3nm semiconductor manufacturing technology generation; cited as a node whose wafer output increased during the period in question.

[19] N5 — 5-nanometer process node — TSMC's 5nm semiconductor manufacturing technology generation; cited alongside N3 as a node benefiting from tool-only capacity additions.

[20] process node — A named semiconductor manufacturing technology generation defined by minimum feature size (e.g., N3, N5); used to categorize wafer output.

[21] process yield — The percentage of functional chips produced from a wafer relative to the maximum possible; cited as a non-hardware means of expanding capacity.

[22] Q1/Q2/Q3/Q4 — first/second/third/fourth quarter — Three-month financial reporting periods used by public companies; the transcript references multiple quarters of 2026 and 2027.

[23] shell construction — The building of the physical structure (walls, cleanroom envelope, utilities) of a fab without yet installing manufacturing tools; contrasted with equipment spending.

[24] shell-ready — Describes a cleanroom whose physical structure is complete and equipped with utilities but which has not yet been populated with manufacturing tools.

[25] tool density — The amount of manufacturing equipment installed per unit area of cleanroom space; increasing it is one mechanism for capacity expansion.

[26] tool move-in — The installation and commissioning of manufacturing equipment into a cleanroom; cited as the primary mechanism of capacity addition in the debate.

[27] tool shipments — Deliveries of semiconductor manufacturing equipment from suppliers (e.g., ASML, Applied Materials) to a fab; cited as evidence in the debate.

[28] TSMC — Taiwan Semiconductor Manufacturing Company — The world's largest contract semiconductor manufacturer, headquartered in Taiwan; the subject company of the debate.

[29] wafer — A thin, round slice of semiconductor material (typically silicon) on which chips are fabricated; the standard unit of semiconductor output.

[30] wafer output — The number of wafers produced by a fab over a given period; cited as the metric by which capacity additions are measured.

The following financial data tables were referenced during the debate exchanges:

Legend: TSMC quarterly capital expenditure breakdown showing the shift from shell construction to equipment purchases between Q2 2026 and Q1 2027. Equipment share rose from ~60% to over 85%, reflecting the tool-only capacity expansion strategy. Source: TSMC quarterly earnings presentations and CFO commentary.

Legend: Projected capital expenditure allocation for TSMC based on affirmative's claims. Equipment share vs. Shell construction. Note: These figures are used by the opponent to argue a shift in strategy, but do not account for total facility expansion requirements.