NuAxi

Company and supply-chain hub

TSMC

NuAxi groups public briefings that mention TSMC: product launches, capacity and pricing signals, and supply-chain changes. The page updates continuously.

527 briefings trackedLatest briefings

Latest briefings

View more

Fri, Oct 30

  1. Impact score 0

    TSMC’s Patent and the Harsh Realities of 3D DRAM Competition

    TSMC's patent disclosures involving vertical DRAM raise the question of whether the foundry could eventually challenge established memory makers. The source notes that US20220359529A1, published in November 2022, describes a vertical-channel transistor integrated with a hybrid trench-and-stacked capacitor, while another application filed in May 2022 and published as US20230240066A1 places a capacitor directly above a vertical gate-all-around transistor. The article stresses these are technical explorations that should not be merged into a single architecture or treated as product announcements, and do not establish TSMC's production readiness, investment commitment, or intention to sell commodity DRAM; plausible uses (embedded memory, logic-memory integration, or future customer requirements) remain an inference rather than a disclosed commercial strategy. SK hynix has publicly identified both a 4F² vertical-gate platform and 3D DRAM as future directions, though these are related but distinct concepts. The article lists three obstacles for 3D DRAM: electrical control, capacitor integration, and manufacturing economics, concluding that competition will be decided by turning promising structures into dependable products at competitive cost. The source states no price movement or supply change.

    DRAM3D DRAMHBMTSMC, SK hynix and 4 totalSemiWiki
  2. Impact score 58

    Intel Confirms "Nova Lake" Uses 18A-P Internal Node, More Details at CES

    During the Q3 2026 earnings call, Intel CEO Lip-Bu Tan confirmed that the company's upcoming flagship client CPU, codenamed "Nova Lake," will be produced by Intel Foundry using its 18A-P node. Initially, Intel had planned to outsource the main compute tile production for "Nova Lake" to TSMC using its 2 nm N2 node, with TSMC expected to handle about 60-70% of the production and the remainder produced in-house by Intel Foundry on the 18A node, meaning a dual-sourcing arrangement for the next-generation chip. Later supply chain information suggested the company intended to bring the majority of "Nova Lake" compute tile production in-house, with these compute tiles made on the Intel 18A-P node. The CEO stated: "Nova Lake," our first Intel 18A-P client product, is tracking ahead of schedule, with design wins expected to be twice those of "Panther Lake" at launch, and more details will be shared at CES in January. The source does not state the cause for the shift of production in-house. This production arrangement remains planned; no completed production is reported.

    Nova Lake client CPU (18A-P compute tile)Intel, Intel Foundry and 3 totalTechPowerUpCapacity Change
  3. Impact score 45

    RT NEWS: #TSMC is buying #AUO’s Taichung plants for TWD 38.8B (US$1.2B), gaining existing cleanrooms and infrastructure, while the display maker scra...

    According to a repost from TrendForce, TSMC is buying AUO's Taichung plants for TWD 38.8B (US$1.2B), gaining existing cleanrooms and infrastructure, while the display maker AUO scraps a separate factory sale to Quanta. This is announcement information and does not by itself prove that the transaction has closed or taken effect. The source does not state the cause.

    Taichung plantsexisting cleanrooms and infrastructureTSMC, AUO and 3 totalTrendForce Official (Market Data)Partnership/M&A
  4. Impact score 30

    Chip Industry Week In Review

    This is a weekly industry roundup listing multiple semiconductor developments: HBM shortages threaten AI expansion; TSMC weighs price hikes (under consideration only, not confirmed); advanced packaging ramps; domain-specialized AI models; Infineon and UMC boost fab investments; Intel's quantum move; Lam and imec advance 3D integration; Google expands TPUs; AI-driven earnings; and CPU benchmarking methodology for agentic AI. The source provides only headline-level mentions, with no specific figures, dates or causes, and does not state implementation status of these items.

    HBMTSMC, Infineon and 7 totalSemiconductor Engineering
  5. Impact score 55

    Amazon 3Q26 earnings call: TSMC capacity behind AWS's choice to start selling its own AI chips

    On Amazon's third-quarter 2026 earnings call, AWS reported revenue growth of 39.3% year-over-year in the third quarter of 2026, the cloud unit's fastest pace in nearly five years and its sixth straight quarter of acceleration. Segment sales reached US$46.0 billion, putting AWS at a US$184 billion annualized run rate, and operating income rose 55% to US$17.8 billion. The same call attributed AWS's choice to start selling its own AI chips to TSMC capacity. The source does not state the timing or implementation status of this decision.

    AI chipsAmazon Web Services (AWS), TSMCDigiTimes Asia
  6. Impact score 30

    SEMICON West 2026: AI Made the Steps Faster. The Time Between Them Is Next.

    Conference report from SEMICON West 2026: SEMI President and CEO Ajit Manocha told the CEO Summit that the semiconductor industry crossed $1 trillion in annual revenue this year, almost four years ahead of earlier projections; a year ago the forecast for 2030 was $1 trillion, now it is $2 trillion, with $3 trillion possible by 2035; the industry has 102 new fabs coming online through 2029 and needs more than 50 beyond that by 2035. Synopsys CEO Sassine Ghazi named the shortage of engineers as the industry's challenge, and Brooks Automation CEO Dave Jarzynka noted a wafer spends only about 25% of its time in the fab actually being processed. In the design session "Bridging Design and Manufacturing with AI" on Tuesday, October 13 in Room 20 at Moscone North, speakers from TSMC, Intel, Applied Materials, Lam Research, Cadence, Silvaco, PDF Solutions, McKinsey, Cerebras, SanDisk and VerifAIX covered seven themes. Specific figures include: Intel's DTCO decisions dropped from weeks to one or two hours; Applied Materials cut a 3D gate-all-around device simulation from 20.5 days to 19.7 hours on GPUs; McKinsey showed AI-assisted programs reaching tape-out nine months after design starts and cutting debug closure time by 40%; TSMC's agentic place-and-route runs around the clock at 3 to 5x productivity; Rhines showed NAND process steps rising from about 400 to as many as 1,800, wafer lot cycle time stretching from 8 to 10 weeks at 28nm to 16 weeks or more, and mask sets climbing from $1 to 3 million to $20 to 40 million or more; VerifAIX cited first-silicon success at 14% and the best AI pass rate on hardware code generation at 34%; PDF Solutions' Kimon Michaels cited Qualcomm's engineering and manufacturing data growing 600% since 2022, while less than 5% of manufacturing data gets analyzed; Lam's research published in Nature found engineers and algorithms working together, with the handoff in the right place, cut the cost of reaching a process target in half. The session also covered co-optimization (DTCO/STCO) replacing sequential handoffs, trust issues in autonomous design, and digital twins changing development. This is a conference recap and industry background; the source does not state the cause (no single event cause).

    NAND3D gate-all-around deviceCFETTSMC, Intel and 16 totalSemiWiki

Thu, Oct 29

  1. Impact score 60

    [News] Samsung Sees 8nm-and-Below Utilization Topping 100% in 2027; 2nm Projects to More Than Double

    According to Financial News, Samsung expects demand for its 8nm-and-below processes to exceed capacity, with utilization projected to surpass 100% in 2027; per Yonhap News, the number of 2nm projects secured in 2026 is expected to more than double YoY. According to Yonhap News, Samsung secured additional 2nm projects from major cloud service providers and AI/HPC customers in 3Q26, building on wins in the previous quarter. Per Seoul Economic Daily, Samsung plans to begin mass production of its second-generation 2nm mobile chip, the Exynos 2700, in 4Q26, while ramping up 4nm NPUs for a major customer. Per Central News Agency, Taiwanese ASIC firm Faraday said on October 27 that it had secured a U.S. cloud AI chip project using 2nm gate-all-around (GAA) technology, valued at tens of millions of dollars; Samsung will handle wafer fabrication, with Taiwanese foundry UMC and memory manufacturers also participating. Per Yonhap News, Samsung expects HPC to account for over 50% of its foundry revenue in 2027; as background, TSMC's HPC segment contributed 68% of total revenue in 3Q26. With advanced-node demand outpacing capacity expansion, Samsung is considering investments in new fabs in the U.S. and South Korea; per Green Economic Daily, Samsung plans to continue expanding 8nm, 4nm, and 2nm capacity, and the first Taylor foundry fab in the U.S. remains on schedule, with cleanroom operations underway and equipment installation and wafer runs in progress; per The Guru, the company is reportedly preparing to break ground on its second Taylor fab in November, with mass production targeted for 2030. Samsung expects double-digit foundry revenue growth and further earnings improvement in 4Q26. Per Digital Today citing Kiwoom Securities, the combined operating loss of Samsung's foundry and System LSI businesses is projected to shrink from KRW 6.742 trillion in 2025 to KRW 3.921 trillion in 2026, with quarterly losses estimated to have eased to KRW 777 billion in 3Q26.

    8nm-and-below processes2nm4nmSamsung, Faraday Technology and 4 totalTrendForceCapacity Change
  2. Impact score 55

    📡 TSMC Targets 6.4Tbps COUPE by 2027. At SEMICON Taiwan 2026, TSMC set a COUPE roadmap to double per-COUPE bandwidth from 3.2Tbps to 6.4Tbps by 202...

    At SEMICON Taiwan 2026, TSMC set a COUPE roadmap targeting a doubling of per-COUPE bandwidth from 3.2Tbps to 6.4Tbps by 2027, with 51.2Tbps+ stated as a future goal. This is a roadmap target, i.e., a plan rather than an achieved result. Equipment makers such as GMT Global and Suruga Seiki updated their CPO coupling and alignment progress, gearing up for mass production. The post states that FAU coupling equipment is now the new CPO battleground. The source does not state the cause.

    COUPECPOFAU coupling equipmentTSMC, GMT Global and 3 totalTrendForce Official (Market Data)
  3. Impact score 30

    #Apple unveils its first OLED touchscreen MacBook Pro with #TSMC-made M5 Pro/Max chips and #Samsung Display panels reportedly on board. The launch cou...

    Apple has unveiled its first OLED touchscreen MacBook Pro, reportedly built with TSMC-made M5 Pro/Max chips and Samsung Display panels. The source states the launch could boost orders for Taiwanese ODMs Foxconn and Quanta; this is stated as a possibility ('could boost'), and the Samsung Display panel supply is qualified as 'reportedly', not a confirmed fact.

    OLED touchscreen MacBook ProM5 Pro/Max chipsSamsung Display panelsApple, TSMC and 5 totalTrendForce Official (Market Data)
  4. Impact score 45

    TSMC buys AUO Taichung plant for NT$38.8B; Quanta deal scrapped

    AUO has approved the sale of its Taichung Plant Phase II facilities and related infrastructure to TSMC for NT$38.8 billion (US$1.2 billion), excluding taxes; AUO expects the transaction to generate approximately NT$18.5 billion in disposal gains after related transaction costs and estimated taxes. AUO has also terminated a planned NT$19.7 billion sale of its Huaya facilities to Quanta, after the two companies failed to agree on usage specifications and delivery schedules. The Taichung transaction is approved but the source does not state that it has been completed; the Huaya sale is terminated. The source does not state the cause beyond the stated failure to agree on usage specifications and delivery schedules.

    Taichung Plant Phase II facilitiesHuaya facilitiesAUO, TSMC and 3 totalDigiTimes AsiaPartnership/M&A
  5. Impact score 0

    Intel CTO: process gap with TSMC down to a few points, but scale still lags

    Technology podcast host Dwarkesh Patel released a special episode after spending two days touring Intel's fabs and research laboratories and speaking with several technology leaders. Intel CTO Pushkar Ranade said the company's process technology is now only a few percentage points behind TSMC, although catching up in manufacturing scale will take considerably longer.

    process technologyIntel, TSMCDigiTimes Asia
  6. Impact score 60

    [Opinion] Elon Musk Is at It Again

    Tesla and SpaceX are building Terafab in Texas, with Musk aiming to establish a facility capable of processing 1 million wafers per month within 10 years. Intel has joined as a technology partner, and TSMC has confirmed discussions on potential cooperation. Procurement officials from Tesla and SpaceX visited Korean suppliers at SEMICON West. Whether Terafab will materialize as planned remains uncertain.

    Tesla, SpaceX and 6 totalThe ElecCapacity Change
  7. Impact score 55

    [News] Apple Debuts First Touchscreen OLED MacBook Pro; TSMC, Foxconn, Samsung in Supply Chain Spotlight

    Apple has unveiled a series of new products, including an OLED touchscreen MacBook Pro, an M6 MacBook Pro, an M6 iMac, and an OLED iPad mini; the OLED MacBook Pro is Apple's first Mac with an OLED touchscreen. According to MacRumors, the 14-inch and 16-inch models start at US$2,899 and US$3,399, respectively, representing a US$400 premium over comparable mini-LED versions, and both models are scheduled to go on sale November 6. Apple says the M6 marks its transition to 2nm and is expected to use TSMC's N2 process, delivering up to 30% faster graphics performance and 25% faster AI workloads than the M5, while the OLED touchscreen MacBook Pro uses the earlier M5 Pro and M5 Max chips built on TSMC's third-generation 3nm process. According to The Korea Herald, Samsung Display has reportedly secured exclusive OLED panel supply for Apple's 14-inch and 16-inch models, with industry estimates suggesting orders could reach approximately 2.5 million panels through 2027. Per Commercial Times, institutional investors expect the new lineup to drive high-end replacement demand and potentially boost fourth-quarter orders and shipments for Taiwanese ODM partners Foxconn and Quanta, while major PC ODMs remain cautious about the second-half market outlook, with persistently high memory and storage costs putting upward pressure on notebook ASPs and constraining overall PC shipment growth.

    OLED touchscreen panelM6 chipM5 Pro chipApple, TSMC and 5 totalTrendForce

Wed, Oct 28

  1. Impact score 60

    TSMC Could Charge Up to 20% More for CoWoS Packaging Amid AI Crunch

    According to supply chain sources cited by Taiwan's Commercial Times, TSMC is reportedly considering raising prices for its CoWoS (Chip on Wafer on Substrate) advanced packaging by as much as 20%; TSMC has not confirmed anything yet. Foundries and outsourced packaging firms are apparently weighing another round of 10% to 20% increases, as CoWoS demand has spread beyond AI GPUs to server CPUs and custom ASICs, leaving a growing backlog of unfilled orders. Testing is turning into a bottleneck of its own: Nomura estimates, via Seoul Economic Daily, that NVIDIA's Blackwell takes roughly four times longer than Hopper in final testing, while Rubin could need seven times as long. Packages keep growing, with TSMC already mass producing CoWoS-L at 5.5 times the reticle limit. Tool lead times have reportedly stretched from three to six months to as long as eight months, and TSMC, which usually negotiates for discounts or extra features, apparently placed this year's equipment orders at existing prices.

    CoWoSCoWoS-LTSMC, NVIDIATechPowerUpPrice Movement
  2. Impact score 45

    Skymizer partners with Egis to develop 12nm and 6nm AI accelerators

    Taiwanese startup Skymizer has formally announced a strategic partnership with Egis Technology to develop Flow (T12), a new-generation AI inference chip. Fabricated on TSMC's 12nm node and paired with LPDDR5 memory, the solution requires just three PCIe cards and 750W of power to run 3-trillion-parameter large language models (LLMs), representing roughly one-tenth the power consumption of a comparable Nvidia H100 deployment. Chip development is part of the announced partnership plan; the source does not state a production or delivery timeline.

    AI inference chip (Flow T12, 12nm)Skymizer, Egis Technology and 3 totalDigiTimes AsiaPartnership/M&A
  3. Impact score 50

    Rosenblatt names top semiconductor picks amid data-center strength

    According to a report dated October 27, 2026, Rosenblatt Securities named AMD its top semiconductor pick heading into earnings season, citing strengthening data-center fundamentals and broadening demand despite recent share price declines across the sector. The firm noted that semiconductor stocks with significant data-center revenue exposure declined approximately 25% on average over the past month, while citing strong earnings reports from Penguin Solutions, Micron, TSMC, and Texas Instruments, along with Supermicro's positive preannouncement, as evidence that data-center fundamentals continue to strengthen. Rosenblatt expects TSMC's 2nm capacity to double in 2027, which will support the ramp of AMD's Venice server CPU family; the firm views Venice as the market's highest-performing server CPU, featuring 16 memory channels compared to 12 in the prior generation. AMD CEO Lisa Su recently increased the projected 2030 server CPU total addressable market to $220 billion from $120 billion. The firm said Meta's Muse and other emerging personal AI agents are driving server CPU demand. Rosenblatt's primary concern for data-center growth in 2027 and beyond is supply-chain capacity rather than demand weakness: Lam Research cited near-term supply constraints and Texas Instruments highlighted extended lead times for back-end equipment, but the firm characterizes these as tactical challenges the industry can address rather than evidence of severe product shortages or weakening demand.

    semiconductorserver CPU2nm capacityAMD, TSMC and 7 totalYahoo Finance — TXN

Tue, Oct 27

  1. Impact score 65

    TSMC reportedly weighs 20% CoWoS price increase as Nvidia Blackwell and Rubin chips put pressure on packaging capacity

    According to Trendforce, citing supply chain sources and local media, TSMC is weighing raising prices for its Chip on Wafer on Substrate (CoWoS) advanced packaging technology by as much as 20%; TSMC has yet to confirm the report. The reported cause is strained capacity amid AI demand and longer delivery times for chipmaking tools: equipment lead times have reportedly stretched from three to six months to as long as eight months, and TSMC has reportedly placed this year's equipment orders at existing prices instead of negotiating discounts. CoWoS combines logic dies and HBM stacks in a single package; Nvidia's Blackwell already pairs two large GPU dies with eight HBM3E stacks, while the upcoming Rubin and Rubin Ultra are expected to push integration further and increase pressure on packaging. Testing is another bottleneck: Blackwell reportedly takes three to four times longer to test than Hopper, and Rubin could need as much as seven times longer; server CPUs and custom ASICs also compete for advanced packaging capacity. Background: TSMC deputy co-COO Cliff Hou said that by July the company's expected equipment needs had reached 1.9 times its original estimate; earlier this month GlobalFoundries agreed to produce silicon interposers for TSMC's CoWoS in the U.S. under a five-year contract worth $2 billion, with production expected to ramp up in the first half of 2028; TSMC is reportedly also considering broader price increases for 2027, with advanced-process wafer prices expected to rise 5% to 10%, steepest for 2nm and 3nm-class node customers, and customers seeking capacity beyond original allocations facing premiums that could push some costs up by as much as 25%. Industry estimates cited by Trendforce put TSMC's monthly capacity at over 106,000 wafers in 2026, with a target of 227,000 wafers in 2027. If both packaging and wafer price increases materialize, AI hardware costs will likely climb further.

    CoWoS先进封装先进制程晶圆硅中介层TSMC, Nvidia and 3 totalTom's HardwarePrice Movement
  2. Impact score 95

    RT NEWS: Another price hike in the works? #TSMC is reportedly eyeing a 20% jump in #CoWoS prices, on top of planned foundry price increases, as AI GPU...

    According to TrendForce, TSMC is reportedly eyeing a 20% jump in CoWoS (advanced packaging) prices, on top of planned foundry price increases. The source states the cause is a fresh surge in advanced packaging demand fueled by AI GPUs, server CPUs, and custom ASICs. The report also mentions TSMC skipping equipment price renegotiations (per the report title). Note: this is a 'reportedly' claim — a plan/intent, not a confirmed or effective price increase.

    CoWoSAI GPUserver CPUTSMCTrendForce Official (Market Data)Price Movement
  3. Impact score 20

    Anthropic's IPO filing puts TSMC, HBM, direct leasing at the center of AI compute risk

    Anthropic filed a preliminary IPO registration statement with the US Securities and Exchange Commission, in which it describes how much of its capacity depends on a small group of chip and memory suppliers, most of them in East Asia. The filing puts TSMC, HBM and direct leasing at the center of its AI compute risk disclosure. This is a risk statement in a preliminary registration document; it does not report any capacity change, halt or price movement.

    HBMTSMCDigiTimes Asia
  4. Impact score 65

    [News] TSMC Reportedly Eyes 20% CoWoS Price Hike, Skips Equipment Price Renegotiations Amid AI Demand Surge

    According to supply chain sources cited by Commercial Times, as increasingly large AI accelerators strain advanced packaging capacity, TSMC's CoWoS supply remains tight, and TSMC is reportedly weighing price hikes of up to 20%, with some packaging lead times already stretching; Commercial Times reports that foundries and OSAT providers are considering another round of CoWoS price increases of 10%–20%. The stated backdrop is that advanced packaging demand is expanding beyond AI GPUs to server CPUs and custom ASICs, adding to a growing backlog of unfilled orders; longer AI chip testing times are an additional bottleneck, with Seoul Economic Daily citing Nomura Securities that NVIDIA Blackwell's final testing takes an estimated four times as long as Hopper's, while Rubin could require seven times as long. On equipment, Commercial Times reports equipment lead times have lengthened to six to eight months, up from three to six months, as customers prioritize securing deliveries over negotiating prices; TSMC, which typically negotiates price cuts or additional features at no extra cost, reportedly placed equipment orders at existing prices this year without seeking discounts or upgrades. Separately, according to Liberty Times, TSMC has finalized plans to raise wafer-out prices by 3%–6% starting January 2027, depending on the process node, with advanced nodes such as 2nm and 3nm seeing the steepest increases. On capacity, analysts cited by Commercial Times estimate monthly CoWoS capacity will exceed 130,000 wafers in 2026 and target 200,000 in 2027; Goldman Sachs projects TSMC's CoWoS capacity to reach 1.275 million wafers in 2026 (roughly 106,000 per month), before surging 114% YoY to 2.73 million in 2027 (227,500 monthly), and climbing another 27% to 3.48 million wafers in 2028 (nearly 290,000 per month). At its latest earnings call, Chairman and CEO C.C. Wei highlighted strong yields for its 5x reticle-size packaging, already in mass production, while confirming development of 9x and 14x solutions; the company is also advancing CoPoS to support larger packages at lower costs, with Commercial Times pointing to a 2028 timeline for 14x packaging. The price increases remain reported or planned and the article does not state they have been formally implemented.

    CoWoS2nm3nmTSMCTrendForcePrice Movement