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
Samsung Display reported its Q3 2026 results: according to Samsung Electronics' third-quarter earnings conference call on Oct. 29, Samsung Display recorded revenue of 8 trillion won and operating profit of 1.1 trillion won. Revenue declined approximately 1.2% and operating profit fell 8.3% year-on-year, while revenue increased 6.7% and operating profit rose 57.1% from the previous quarter. Shipments of small and medium-sized OLED panels increased following new product launches by major customers; small and medium-sized business revenue and profit both increased from the previous quarter, and the large OLED business improved earnings from the previous quarter through increased sales of premium products and cost reductions. Samsung Display expects panel demand to increase in the fourth quarter, supported by the year-end peak season and new product launches, though external factors such as exchange rate fluctuations could weigh on earnings improvements. The outlook for next year remains uncertain: rising demand for memory used in AI servers is pushing up mobile memory prices, increasing cost pressures on smartphone manufacturers. Samsung Electronics' Mobile eXperience (MX) and Networks businesses posted an operating loss of 600 billion won in the third quarter, their second consecutive quarterly loss; although foldable smartphone sales volumes and revenue increased year-on-year following launches such as the Galaxy Z Fold8, higher memory prices weighed on profitability. Samsung Display plans to increase the proportion of premium products including foldable displays, improve crease visibility and durability in foldable panels, and ramp up operations at its new eighth-generation 8.6 IT OLED production line; in large OLED it will focus on ultra-large TVs and gaming monitors while expanding its customer base into gaming consoles and automotive applications. Capital expenditure this year is expected to reach 3.9 trillion won, up 1.1 trillion won from last year, with cumulative investment through the third quarter totaling 2.4 trillion won. Heo Cheol said market uncertainty is expected to persist in 2027 with demand remaining weak, and the company will pursue productivity improvements and cost reductions to maintain profitability.
small and medium-sized OLED panelslarge OLED panelsfoldable displaysSamsung DisplayThe Elec
Shin Seung-cheol, vice president of Samsung Electronics' System LSI Business under its Device Solutions (DS) division, said during the company's third-quarter earnings conference call on Oct. 29 that the Exynos 2700 has improved AI and gaming performance and that development and mass production are proceeding without disruption, with Samsung expecting to further expand market share by continuously strengthening SoC competitiveness. According to Counterpoint Research, Samsung Electronics held a 9% share of the global smartphone AP market by shipment volume in the second quarter of this year, up 2 percentage points from 7% in the previous quarter and up 3 percentage points from a year earlier; the Exynos 2600, used in the Galaxy S26 series, drove the increase in shipments. The Exynos 2700, which recently entered mass production, has increased production volume 10% compared with the previous generation and has been confirmed for use in the new Pro model of the Galaxy S27 series, scheduled for release next year, in addition to the standard and Plus models for South Korea and Europe. Samsung's System LSI Business has also secured a new custom SoC project by establishing a business partnership with a new customer. The source does not state the cause.
Delta Electronics approved plans on October 28 to establish a large manufacturing base in Penang, Malaysia, as the power and thermal management company expands overseas production to support customer demand. The site is expected to ramp up to capacity comparable to that of Delta's Thailand operations over five years, with space reserved for process research and development and a potential future R&D center. The stated purpose is to support customer demand. The event is an approved plan and has not yet been implemented.
The article examines whether Texas Instruments' (TXN) share price, now at US$309.83, can be justified by expected future cash flows: over the past 3 years the stock has gained 133.7%, latest twelve-month free cash flow is about $5.2b, and the DCF model assumes growing cash generation with undiscounted forecasts rising into the mid-teens of billions by 2030, yet the DCF output comes in substantially below the current share price of $309.83, implying the market is baking in richer cash flow or lower risk. The article cites Texas Instruments' strong Q3 2026 earnings and upbeat Q4 outlook, along with community narratives: a bull case (29% undervalued, citing a multiyear capacity-expansion cycle that is temporarily suppressing free cash flow while enhancing long-term competitive position) and a bear case (9% overvalued, questioning whether the 14.2% growth input may be reading an analog cycle recovery as secular growth), and notes the planned Silicon Labs acquisition and expanding manufacturing capacity could affect how quickly future cash flows arrive and how much reinvestment they require. The source does not state the cause.
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
A research commentary article discusses the 2023 paper 'Machine Learning-Based Microarchitecture-Level Power Modeling of CPUs', published in the IEEE Transactions on Computers with authors from UT Austin plus Hyperstone and U Stuttgart in Germany; the paper has 18 citations and extends a 2021 learning-based power modeling paper (which covered fixed function accelerators) to programmable in-order and out-of-order CPUs, with the learning method based on DNNs. The authors train decision-tree models to predict power for a single clock cycle on low level CPU building blocks such as fetch, decode, and load-store, trained using golden power data from a post-P&R gate-level netlist and simulation traces, then assemble the models hierarchically to construct a cycle-accurate power profile from a cycle-accurate micro-architectural simulation trace of the complete design. Results on two open source designs (a simple in-order CPU and the complex out-of-order CPU BOOM), implemented on an open source 45nm process and library, show that with only 20k cycles of training data using 8 very simple test programs (quick sort, Towers of Hanoi, Dhrystone, ...), power estimation for BOOM on the CoreMark benchmark has an average error in any clock cycle of less than 4%; at the resolution of individual building blocks, 80% of the blocks are within a 15% error in any clock cycle; and the runtime overhead of processing the decision trees is negligible. The commentary also cites that gate-level reference power estimation of the BOOM processor (approx. 460K standard cells) took a total of 40-65 hours running at 2-3 cycles/s versus 7-9 hours using the learned models and Verilator-based simulation, a speedup of roughly 5-9x; that the 3.6% MAE on CoreMark is normalized to the average power of the entire CPU and does not mean power is predicted within 3.6% in every cycle, with accuracy varying strongly across blocks (ALU 15% MAE, issue unit 18%, memory issue unit 23%); and that the decision trees are calibrated against a particular synthesized and placed-and-routed design, with the authors providing no evidence that characterized blocks preserve accuracy across different configurations, so applicability to design-space exploration remains unclear.
A new report from Frost & Sullivan and LeadLeo Research indicates that SenseTime Group secured nearly half of China's neocloud market in the first half of 2026, with the market share captured primarily through its AI infrastructure platform, SenseCore. As background, the report was published amid Elon Musk's prediction that more than 200 new cloud providers (neoclouds) face impending bankruptcy or acquisition in a global market consolidation. The source does not state the cause beyond the stated capture primarily through SenseCore.
On October 27, 2026, Samsung Electronics Chairman Lee Jae-yong marked his fourth anniversary as chairman by visiting a robotics research center, where he reviewed humanoid robot technology and preparations for commercialization, instead of holding a commemorative event. As Samsung integrates its research organizations and plans manufacturing-floor applications, robotics is emerging as a key pillar of its physical AI strategy. The source does not state any direct impact on component supply or prices.
Next-generation semiconductor process data of Samsung Electronics was allegedly removed without authorization by a university-industry research partner in South Korea, raising concerns over security vulnerabilities in industry-academia collaborations. The researcher, who reportedly retained Samsung employee access while working on the project, was indicted after allegedly transferring extreme ultraviolet (EUV) lithography information to a personal cloud account for use in the collaboration.
EUV lithography informationnext-generation semiconductor process dataSamsung ElectronicsDigiTimes Asia
According to new research from the Global Electronics Association and Accuris, an engineering intelligence and electronic component data provider, applications for silicon carbide (SiC) and gallium nitride (GaN) power devices are expanding across the electronics industry, affecting not only component selection but also product design, manufacturing capabilities and supply chain planning. The source provides no specific figures, companies or event status changes beyond this stated demand-driven trend.
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.
According to a product news release from Murata Manufacturing Co., Ltd. dated 10/29/2026, the company has begun mass production of the SWL series MM8330-2600RJ5 and MM8330-2600RK20, the world's smallest RF/microwave coaxial connectors with a switch (based on Murata research as of October 28, 2026). Measuring 1.0 mm × 1.0 mm × 0.55 mm, the connectors reduce PCB mounting area by approximately 40% compared with Murata's previous SWJ series (1.4 mm × 1.2 mm × 0.65 mm); when used with dedicated probes, the PCB footprint including the keep-out area is reduced by approximately 60% versus the SWJ series. The SWL series supports DC to 15 GHz with a nominal impedance of 50 Ω; the MM8330-2600RJ5 is supplied on reels of 5,000 pieces and the MM8330-2600RK20 on reels of 20,000 pieces. The dedicated probes are scheduled to enter mass production in December 2026.
RF / microwave coaxial connectors with a switchMurata Manufacturing Co., Ltd.Murata News
Ting-Hao Lin, R&D vice president at Kinsus Interconnect Technology, said at the 2026 TPCA Show that surging demand for AI computing is driving advanced packaging toward larger form factors. To support this trend, substrate materials must deliver low warpage and exceptional dimensional stability while overcoming challenges in thermal expansion mismatches and structural stress. The source states no specific capacity, pricing or expansion plans.
advanced packaging substratesKinsus Interconnect TechnologyDigiTimes Asia
According to Kinsus Interconnect Technology R&D vice president Ting-hao Lin at the 2026 TPCA Show, Kinsus is exploring aluminum nitride (AlN) ceramics as a core material for substrates used in large-scale advanced packaging, seeking an alternative to glass. The exploration of a glass alternative is set against ongoing brittleness and reliability challenges of glass. The effort remains at the exploration stage, i.e., planned/research status, not an achieved mass-production or commercialized result.
aluminum nitride (AlN) ceramicsglass core materialsubstratesKinsus Interconnect TechnologyDigiTimes Asia
According to a consensus among economic research institutions, 2026 will be a breakout year for Taiwan's economic growth, with GDP forecast to expand at a double-digit rate—more than twice China's growth rate. The forecast also states that over the next six months, the electronics and machinery sector is expected to remain resilient, while securities, textiles, chemicals, and petrochemical raw materials could weaken. The source does not state the cause behind these projections.
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.
Finetek will supply 10 billion won worth of OLED manufacturing equipment to Samsung Display Vietnam, consisting of bonding systems used in OLED panel manufacturing. The order follows a 19.5 billion won contract signed in July. Finetek stated on Oct. 28 that cumulative orders this year, including the latest contract, totaled approximately 37 billion won, equivalent to 80% of last year's revenue, and will be reflected in revenue sequentially from the second half of this year. Finetek also plans to expand its existing OLED equipment business and new businesses, including robotics.
QRT received a Ministerial Commendation from South Korea's Ministry of Science and ICT at the International Research Industry Convention 2026, held at the Suwon Convention Center on Oct. 28. The company was recognized for developing specialized semiconductor reliability testing equipment, which has reduced South Korea's dependence on foreign suppliers by 70%. QRT has commercialized single-event effect (SEE) testing equipment for radiation-induced soft errors in semiconductors used in space, aviation, defense, autonomous driving and AI applications, as well as a radio-frequency high-temperature operating life (RF-HTOL) testing system for communications applications; its SEE equipment for neutron and heavy-ion testing has established a foundation for reliability verification, and the RF-HTOL system has enabled evaluation of the operating life of high-power and high-frequency semiconductors. QRT said it delivered RF-HTOL equipment to a European defense customer in the first half of this year and has continued to secure equipment orders from global customers in the second half.
Microsoft and Nvidia publicly released the Surface Laptop Ultra on October 7, built around the application processor originally referred to as the N1X and now renamed the GB10. The chip was co-developed by NVIDIA (which contributed the Grace Blackwell GPU subsystem) and MediaTek (which supplied the multi-core CPU cluster and reportedly also handled full SoC integration duties); it was originally intended for—and may eventually still show up in—Arm-based Windows PCs, but repeated development hurdles reportedly delayed both SoC and system shipment plans, and lingering functional bugs precluded Windows compatibility (explaining the DGX Spark's Linux OS foundation). The machine comes in two versions sharing the same NPU core: 5120-core GPU with 18-core CPU, and 6144-core GPU with 20-core CPU; system memory ranges from 24 GB to 128 GB of LPDDR5x unified memory; SSD options start at 512 GB (PCIe Gen 4) with 1 TB and 2 TB (PCIe Gen 5) configurations; pricing is $2,599 baseline up to $5,899.99 top-end, with delivery promised 'later this year'; an RTX Spark-based developer platform at $5,999.99 and up is also available for pre-order. In the same period, Amazon launched its latest-generation Alexa-branded (no-longer-Fire) tablets in 8-inch, 11-inch and 12-inch 'Pro' flavors, running 'vanilla' Android with a return to the Google Play Store, priced from $299 up to $679.99 (not including cases or the high-end model's optional pen and keyboard accessories). Apple's week began Monday with a new-generation iPad mini (now with the A20 Pro, a slightly larger 8.5-inch display switched from LCD to OLED, priced from $699 and up; the 2024 entry-level 128 GB Wi-Fi-only model was originally $499 but was bumped to $599 in June of this year), Tuesday brought a 14-inch entry-level MacBook Pro now running the M6 SoC (versus last year's M5) plus a 24-inch iMac making a two-generation jump from the M4 to the M6, and today Apple unveiled its first touchscreen-augmented MacBook Pro laptops in both 14-inch and 16-inch displays with ProLED XDR panels and both M5 Pro and Max SoC options, priced from $2,899 to $14,549.
Nvidia RTX Spark N1X (GB10) application processorApple A20 Pro SoCApple M6 SoCApple, Microsoft and 6 totalEDN