Silicon Nexus

TSMC reported Q3 2026 EPS of NT$29.43, exceeding analyst expectations, with both USD-denominated revenue and operating margin surpassing guidance. The dual beat on top- and bottom-line metrics confirms robust AI-driven wafer demand and sustained pricing power at advanced nodes. Results reinforce TSMC's dominant position and may pressure Samsung Foundry's narrative heading into its own earnings.

Why it matters: Quarterly earnings beat on EPS, USD revenue, and operating margin is a direct stock-moving event for TSMC and has read-across implications for key customers and foundry competitors.

TSMC reported Q3 EPS of NT$29.43, surpassing analyst expectations, with both USD-denominated revenue and operating margin exceeding consensus targets in a double beat. The results signal sustained demand strength at leading-edge nodes, driven primarily by AI accelerator and HPC customers. The outperformance reinforces TSMC's pricing power and high-utilization outlook heading into Q4.

Why it matters: TSMC's Q3 earnings beat on EPS, USD revenue, and operating margin simultaneously is a direct stock-moving event for 2330 and a bullish demand signal for the broader AI-driven semiconductor supply chain.

TSMC reported record Q3 revenue and net profit in its earnings call, with EPS of NT$29.43 (~$0.91 USD). The 2nm node's contribution to revenue climbed to 8%, signaling a faster-than-expected ramp of its leading-edge process. The beat underscores continued strength in AI-driven advanced node demand, reinforcing TSMC's pricing power and capacity utilization.

Why it matters: Formal earnings call with record revenue/profit figures and a quantified 2nm ramp milestone are direct stock-moving disclosures for TSMC and its AI-supply-chain peers.

KB Securities analyst Kim Dong-won argues Samsung Electronics' DRAM technology leads China's CXMT (ChangXin Memory) by up to seven years on a mass-production timeline: Samsung ramped 1z-node DRAM in Sep 2019, while CXMT only enters equivalent mass production now; the 1a gap stands at five years. Samsung shares have dropped ~16% in October on China competition fears, but KB Securities views the selloff as excessive given Samsung's continued lead in advanced nodes and HBM. The bank sees the pullback as a buying opportunity, citing potential upside from higher dividends and share buybacks.

Why it matters: An analyst note rebutting market fears about CXMT's progress with concrete process-node benchmarks — directly relevant to Samsung's near-term stock narrative, but not a hard capex, contract, or earnings event.

TSMC (2330) shares dropped approximately NT$70 ahead of its upcoming earnings call, with foreign media highlighting that the stock failed to rally after 3 of its past 4 quarterly reports. The pattern raises caution among investors positioning ahead of the call, suggesting post-earnings selling pressure could persist regardless of headline beat. No new guidance or capex figures were disclosed in this report.

Why it matters: Pre-earnings sentiment and historical post-report price pattern for TSMC is relevant for timing, but the article contains no new fundamental data, guidance, or capex disclosure.

Gartner projects the AI industry will pour roughly $1T into infrastructure in 2026 and an additional $6T from 2027–2030, but attributable AI revenue may reach only $3.6T against an $8.5T breakeven threshold (12% ROI hurdle) — a ~$4.9T gap that points to large-scale asset write-downs. Rapid GPU depreciation as newer hardware displaces older equipment will accelerate impairments for GPU-cloud operators. Gartner estimates ~80% of new cloud providers focused on GPU-as-a-service will not survive as independent companies beyond 2030, while diversified hyperscalers are better insulated.

Why it matters: A sector-level Gartner forecast on AI infrastructure ROI risk and consolidation — no specific company named or imminent contract/earnings event, but directly shapes demand outlook for AI memory and foundry customers.

ASML reported Q3 net sales of €11.7B with a 57.5% gross margin and raised its full-year guidance to €44–45B, citing demand exceeding near-term supply. CEO Fouquet framed growth as dual-speed: 2027 low-NA EUV and immersion DUV capacity each expanding ~30% YoY to support near-term revenue, while the larger High-NA EUV wave depends on Intel's layer ramp, Samsung's 2028 mass-production target, and TSMC's ~2030 advanced-logic adoption. TSMC and ASML are jointly developing a 12-inch mask infrastructure—pilot line 2031, volume production 2033—to ease reticle-stitching constraints for large AI chips.

Why it matters: ASML earnings beat and guidance raise are clear events, but the stock-moving implications for tracked names (TSMC, Samsung) are roadmap-horizon rather than imminent capex commitments.

Four major CSPs have collectively raised 2027 capex guidance to $1 trillion, sustaining Taiwan AI supply chain growth through Q4 as NVIDIA Vera Rubin, Google TPU, and AWS Trainium3 ramp toward mass production. Nomura Investment Trust is overweight thermal management, PCB, CCL, and optical communications/CPO suppliers, noting Vera Rubin enters full ramp in late Q3/early Q4 and their internal stock universe logged a ninth consecutive month of upward EPS revision to +77.3% YoY for 2026. The key risk is U.S. Treasury yields breaking 5.3% amid a resumed Fed rate hike cycle, which Nomura warns may suppress near-term Taiwan equity valuations.

Why it matters: Fund manager strategy note containing a material CSP capex datapoint ($1T 2027 guidance) and Vera Rubin ramp timeline, but lacks company-specific contract or earnings disclosure that would directly move individual stocks.

Nomura Investment Trust recommends rotating into upstream AI supply-chain names — thermal management, rails, test interfaces, and IC design services — as component shortages and price hikes widen the return gap between high-margin upstream players and margin-squeezed assemblers. Four major CSPs have raised combined 2027 capex guidance to $1 trillion, and Nvidia's Vera Rubin GPU is expected to enter mass production late Q3 to early Q4, driving visible monthly revenue inflections for select suppliers. TSMC's Q3 earnings commentary on advanced-process and advanced-packaging demand is flagged as the single most important near-term catalyst.

Why it matters: Actionable sector rotation call backed by concrete data points (CSP $1T capex, Vera Rubin timing, 9-month EPS revision streak), but it is an institutional fund manager's strategy note rather than a primary news event.

TSMC shares dropped sharply in early trading on Oct 15 as investors positioned ahead of the company's afternoon institutional investor conference, pulling Taiwan's TAIEX index down roughly 1,400 points and threatening the key 47,000 psychological support level. The broad selloff reflects pre-earnings nervousness around TSMC's guidance, given its bellwether status for the global semiconductor supply chain. The afternoon call is expected to reset near-term expectations for AI-driven foundry demand.

Why it matters: TSMC earnings/investor conference is a direct stock-moving catalyst with bellwether implications for the entire AI and foundry supply chain.