Memory substrate maker TLB said on Nov. 5 that third-quarter consolidated revenue rose 36.4% year-on-year to 93.965 billion won and operating profit surged 91.8% to 16.595 billion won, posting record quarterly revenue; revenue rose 6.5% and operating profit climbed 29.9% from the previous quarter, with operating margin at 17.7%, up 5.1 percentage points year-on-year and 3.2 percentage points quarter-on-quarter. The growth was driven by expanded shipments of SOCAMM2 substrates for AI servers; the company said the larger share of higher-value products improved its product mix and boosted profitability. SOCAMM2 is a low-power DRAM module used alongside the Vera CPU in Nvidia's next-generation Vera Rubin AI server platform, and its substrates command higher ASPs than conventional PCBs used in standard memory modules. TLB recently started developing SOCAMM3 substrates with two major domestic and overseas memory manufacturers, targeting mass production in 2028. The company also said some customers have agreed to reflect higher raw material costs in product prices in the fourth quarter, which is expected to contribute to earnings.
On a Nov. 5 third-quarter 2026 earnings conference call, Lotte Energy Materials said it expects its high-end circuit copper foil business for AI data centers to achieve a double-digit operating profit margin of at least 10% next year, driven by growing demand for high-value copper foil used in AI servers and high-speed networks, and the company also plans to expand production capacity. It plans to make circuit copper foil a core product category, raising its share of revenue from an estimated 14% this year to 40% by 2029. To support the expansion, Lotte Energy Materials will increase annual circuit copper foil production capacity at its Iksan plant in North Jeolla Province from 6,700 tons to 16,000 tons by 2027 by adding post-processing facilities within the existing plant, and is considering further expansion in 2028 plus dedicated production equipment for HVLP copper foil and ultra-thin copper foil (UTC). The company currently mass-produces Grade 3 and Grade 4 HVLP copper foil and is developing Grade 5; yield for Grade 1 HVLP is about 60%, below the 80% for conventional circuit copper foil, and the company plans to raise yields to levels comparable with global competitors by the end of this year. It forecasts global HVLP demand to grow at a 33% compound annual rate from 2026 through 2030 versus current demand of less than 30,000 tons, and expects circuit copper foil sales volume to more than double next year with high-end products accounting for over 50% of the total. Lotte Energy Materials reported third-quarter revenue of 289.7 billion won and an operating loss of 21.3 billion won; revenue rose 101.6% from a year earlier while the operating loss narrowed by about 13 billion won. The company expects its Malaysian subsidiary, which produces battery copper foil, to turn profitable in the first or second quarter of next year and to return to profitability on a consolidated basis in the second quarter of next year.
Adam White, president of Infineon's Power Systems division, held a press conference at Infineon Korea's headquarters in Samseong-dong, Seoul, on Nov. 5, outlining the company's “From Grid to Core” power semiconductor strategy and calling AI data center power solutions its most important growth driver. Revenue from AI data center power semiconductors exceeded 700 million euros in fiscal 2025, which ran from October 2024 through September 2025; in February the company set revenue targets of 1.5 billion euros for fiscal 2026 and 2.5 billion euros for fiscal 2027, and in August it raised its fiscal 2026 forecast to more than 1.6 billion euros. White said the company would announce a revised target at its headquarters' earnings presentation on Nov. 10, local time. White expects the power semiconductor market to expand significantly as AI data centers consume more electricity and transition to 800-volt direct current (800VDC) architectures: data centers' share of global electricity generation would rise from the current 3% to 7%–8% by 2030; power consumption per server rack is expected to increase from 100–150 kilowatts today to more than 1 megawatt after 2028, driven by GPUs whose individual power consumption will climb from 1,000 watts to more than 2,000W. The 800VDC architecture is being promoted by Nvidia, which plans to expand adoption alongside shipments of its Kyber racks in 2027. Customers are already shipping mass-produced solid-state circuit breakers, and preparations for mass production of solid-state transformers are accelerating in line with 800V DC adoption; Infineon expects the PSU market to shift from single-phase 8kW units currently in mass production to three-phase 30kW systems; vertical power delivery modules measuring less than 3 millimeters thick are scheduled for release in 2027. Infineon acquired Indian startup C2i Semiconductors in August, gaining digital controllers that support these architectures. Partnerships: with Hyosung Heavy Industries on high-voltage direct current (HVDC) transmission technology; an MOU with LS Electric signed in July to jointly develop SSTs and SSCBs; an MOU with SoluM signed on Nov. 5 to jointly develop server PSUs, BBUs and IBCs; plans announced in August 2025 to jointly develop vertical power delivery (VPD) modules for AI processors with Taiwan's Delta Electronics; a partnership with U.S.-based DG Matrix formed in March 2026 to supply SiC power semiconductors for SSTs; and an MOU with Estonia's Skeleton Technologies signed in September to jointly develop power solutions for AI data centers, including SSTs and high-power sidecar power systems.
power semiconductorssolid-state transformers (SSTs)solid-state circuit breakers (SSCBs)Infineon, Nvidia and 9 totalThe ElecPartnership/M&A
Related manufacturers
Infineon, Nvidia, Hyosung Heavy Industries, LS Electric, SoluM, Delta Electronics, DG Matrix, Skeleton Technologies, C2i Semiconductors
The Korea Display Industry Association (KDIA) issued a statement on Nov. 6 welcoming the approval of the Advanced Display National Research Platform Project. The project passed the preliminary feasibility review at a meeting of the National Research and Development Project Evaluation Steering Committee on Nov. 5. The project has a total budget of 368.87 billion won, including 278.9 billion won in government funding, and will invest from 2027 to 2033 to establish research infrastructure capable of validating the entire display development process, from panel design and manufacturing to module integration, supporting open R&D infrastructure for resolutions of up to 10,000 pixels per inch (ppi) along with core technologies for next-generation displays. The government plans to establish an Advanced Display Research Institute through the project, where companies, universities and research organizations can jointly use equipment and research infrastructure. The KDIA said South Korea maintained its global leadership in the OLED market with a 68.7% share in 2025, but Chinese manufacturers are rapidly catching up in OLED and next-generation technologies. The project has only passed the preliminary feasibility review and remains in the planning stage; implementation has not begun.
Nvidia disclosed power efficiency benchmarks and gaming support plans for RTX Spark, which combines its own CPU and GPU, at the RTX Spark Press Event held at the JW Marriott Hotel in Seoul on Nov. 6, presented by Jeff Yen, Nvidia's Asia-Pacific (APAC) director of technical marketing. According to Nvidia's benchmark data, RTX Spark delivered approximately twice the performance of a laptop equipped with an Intel Core Ultra 9 CPU and Nvidia RTX 5070 discrete GPU at 60 watts; the comparison was based on performance across 20 major games and prompt processing in large language models (LLMs). When running on battery power, RTX Spark maintained approximately 98% of its plugged-in performance. Gaming figures disclosed include 142 FPS in Doom: The Dark Ages, 137 FPS in 007 First Light and 135 FPS in Cyberpunk 2077. RTX Spark uses an Arm-based CPU architecture, so x86 games cannot run directly; Microsoft's Prism emulator for Windows on Arm exists, but emulation can cause performance degradation or compatibility issues. Nvidia is collaborating with major developers to expand the number of games running natively rather than via emulation; partners include EA, Activision, Riot Games and Krafton, with supported titles including Apex Legends, the Call of Duty series and PUBG: Battlegrounds. Riot Games plans to support League of Legends and Valorant on RTX Spark using its own anti-cheat solution. Nvidia has also incorporated DLSS 5, a second-generation transformer-based graphics enhancement technology, into RTX Spark, using neural rendering to reproduce lighting, materials and textures in games.
Maxsun has shared full specifications for the SK-PTLNAS and SK-WCLAIO, two NAS-focused MoDT (Mobile on Desktop) motherboards built around Intel's Panther Lake and Wildcat Lake chips. Both were first shown at Computex 2026 and are expected to launch in November, though pricing is still unknown. The SK-PTLNAS now tops out at the Core Ultra 9 386H (16 cores, four Xe3 GPU cores, 25 W base power, up to 80 W turbo power), supports up to 128 GB of DDR5-7200 across two SO-DIMM slots, five M.2 slots, three Mini SAS ports, a PCIe 5.0 x8 slot, 10GbE, 2.5GbE and two 1GbE ports, two Thunderbolt 4 ports, and lists Windows 11 and Ubuntu support; the Core Ultra X9 388H option listed at Computex no longer seems to be on the table. The SK-WCLAIO measures just 90 x 165 mm, a big change from the mATX design shown in June, and runs the six-core Core 7 350 (2P + 4LPE) with two Xe3 cores, a single DDR5-6400 SO-DIMM slot for up to 48 GB, 1 TB of onboard UFS 3.1 storage and four M.2 slots; since Wildcat Lake has very few PCIe lanes, those drives probably won't each get full x4 bandwidth. Maxsun rates the system at 45 W when active and 16 W with drives on standby, and the power input is unclear for now, with the listing mentioning both a 12 V DC jack and USB-C PD.
Tom's Hardware reported on 8 November 2026 that the latest CPU-Z hardware survey shows the first-ever year-over-year decline in PC RAM capacity: 64GB configurations plummeted 18% and 32GB configurations fell 2.5% year on year (data released by the CPU-Z developer on 6 November 2026). Meanwhile, 16GB configurations rebounded by 12.9%, 8GB configurations grew 5.5%, and 24GB configurations increased 19.7% as a compromise for consumers needing more than 16GB without paying for premium 32GB kits. On memory types, DDR4 remains dominant at a 54.5% share with 6% growth, while DDR5 holds 32.9% with a 2.9% decline. The source attributes this to the AI rush gobbling up the global DRAM supply, which has driven DDR4 and DDR5 prices through the roof, forcing consumers to reconsider the cost-benefit of larger capacities and opt for more affordable kits. CPU-Z processor charts show significant Zen 3 increases, with Ryzen 7 5700X, Ryzen 5 5600 and Ryzen 5 5500 up 42.8%, 36.1% and 67.3% respectively; supporting developments include AMD rereleasing the Ryzen 7 5800X3D, vendors restarting DDR4 and DDR4 motherboard production, and Intel preparing rewarmed Raptor Lake Next chips for 2027. Note that CPU-Z only tracks systems where it is installed, so the data represents its active user base and is not retail sales figures in any sense.
According to the latest CPU-Z Validator statistics cited by Doc TB, one of the developers behind CPU-Z, this is the first time RAM capacity on PCs has declined year over year; the figures only cover systems users actually submitted to the validator over the last three months, so they are not a full picture of the PC market. 32 GB remains the most popular configuration at 37.2% of validations, though its share is down 2.5% compared to a year ago; 64 GB took a much bigger hit, falling 18% to 7.3%; 16 GB rose 12.9% to 33.0% and 8 GB rose 5.5% to 11.0%, while 96 GB and 128 GB setups dropped by more than 26% each. Memory type tells a similar story: DDR4 now makes up 54.5% of validations, up 6.0%, while DDR5 slipped 2.9% to 32.9%. The article notes Steam's September survey showed 32 GB overtaking 16 GB as the most common configuration among gamers, but that flipped with the October survey. The source ties this to memory prices having climbed sharply this year, saying plenty of people are sticking with older DDR4 systems or settling for 16 GB instead of paying current prices; it also mentions Team Group expects the worst of the shortage to land between Q4 2026 and Q2 2027, and DDR4 motherboards for unannounced LGA1700 processors expected in early 2027.
According to Intel leaker Jaykihn on X, Intel's Razor Lake desktop processors, the generation following Nova Lake-S, are currently slated for 2028; asked whether Razor Lake-S was planned for 2027, he replied "not currently," then said Razor Lake is a 2028 product for now. A leaked partner slide placed the first Nova Lake-S SKUs in Q1 2027 and indicated Intel plans to keep LGA-1954 for multiple generations, so Razor Lake-S should drop into the same boards roughly a year later. Razor Lake will reportedly use new Griffin Cove P-cores, called "an entire tock" rather than a refinement; the same core will reportedly also power Intel's next-generation Coral Rapids Xeon processors, though SMT stays exclusive to Xeon. Earlier leaks pointed to TSMC's N2X node for some Razor Lake production, and the 12 Xe3P core iGPU SKU originally meant for Nova Lake-S was reportedly pushed to Razor Lake. Intel has not confirmed any of this, so 2028 should be treated as a target and timelines can move.
According to a report from The Information, two people familiar with the initiative said TSMC has begun preparing plans for 3D DRAM manufacturing following a request from its largest customer, Nvidia; the initiative is reportedly overseen by Hsu Kuo-chin, TSMC's vice president responsible for advanced packaging and former head of Micron's Taiwanese operations. The report attributes the backdrop to surging demand for AI hardware straining global DRAM supplies and driving up prices, turning memory from a commodity into a strategic asset. TSMC denied the report: a TSMC spokesperson told Tom's Hardware, "Please note that TSMC focuses on our core business at all times, which is dedicated IC foundry," with a mission "to be the trusted technology and capacity provider of the global logic IC industry for years to come," and TSMC states it is a pure-play logic foundry and plans to remain one. The article notes that 3D DRAM is years away from market viability even for major memory makers, so a 3D DRAM strategy could not solve memory shortages for at least five years; all three leading DRAM makers — Micron, Samsung, and SK hynix — are working on their own 3D DRAM projects, and Micron's DRAM strategy includes at least four more planar DRAM process technologies before 3D DRAM becomes a reality. TSMC lacks the processes and production capacity to make DRAM, would need to invest billions of dollars and years of R&D if it entered the DRAM business, and would face cyclical oversupply risk, potential complications with large DRAM makers — in particular SK Hynix, which partnered with the foundry in 2024 to manufacture logic base dies for next-generation HBM products — and possible antitrust scrutiny; the article also notes TSMC could benefit from advanced memory demand without making DRAM, via producing logic base dies for HBM, integrating memory with processors using its SoIC hybrid bonding technology, and developing customized memory solutions with partners such as GUC and established DRAM manufacturers. The 3D DRAM exploration is an unconfirmed media report that TSMC has denied.
TCL CSOT announced it has started moving production equipment into its t8 factory in Guangzhou, China. Feng Zhang, Senior Vice President of TCL CSOT, said the move-in of the first piece of core production equipment marks the t8 project's transition to the equipment installation and commissioning phase. The fab is an 8.6-generation inkjet-printed OLED production line, with mass production scheduled for Q4 2027. Construction started on September 16, 2025, and the main factory structure was completed in less than a year (226 days). The project has a total investment of approximately RMB 29.5 billion (around $4.42 billion). When completed, the t8 fab will process around 22,500 glass substrates monthly, each measuring 2,290 x 2,620 mm, the kind used for manufacturing medium to large-sized OLED panels. Initial production will focus on displays for tablets, notebooks, and monitors, and TCL CSOT is also exploring inkjet-printed OLED technology potential for flexible and foldable panels.
DIGITIMES observes that as large language model (LLM) and agentic AI development become increasingly widespread, this is not only driving significant growth in data center CPU demand but also a sharp surge in key-value cache (KV cache) storage requirements. As DRAM and HBM capacity face severe supply shortages, demand for data center SSDs is rising rapidly, which establishes three key directions for SSD development. The source provides no specific figures, dates, or manufacturer information.
Silanna has launched the Plural Nano family of high-performance analog-to-digital converters (ADCs) for Edge AI and other applications with limited PCB space. The devices offer resolutions from 12 to 16 bits and sampling rates from 5 to 250 Msps in compact 64-ball BGA packages. The family includes two package options: a 3.4 × 3.4 mm version supporting single- and dual-channel 12-bit configurations as well as single-channel 14- and 16-bit devices, and a larger 3.7 × 3.7 mm package accommodating dual-channel 14- and 16-bit variants. According to Silanna, competing ADCs operating above 125 Msps are almost universally housed in packages of at least 9 × 9 mm; against these, the two Plural Nano package variants reduce footprint by 83% and 86% respectively, and against competing 125 Msps converters typically measuring at least 5 × 5 mm, the reductions reach 45% and 54%. Plural Nano shares the proprietary factory-configurable architecture of Silanna's original Plural ADC family, letting teams establish specific product configurations (SKUs) after fabrication. At launch, 90 ADC configurations covering 12-, 14-, and 16-bit resolutions will be available, and Silanna plans to announce additional resolutions and features later in 2026. The devices integrate reference and clock-management circuitry, support LVDS and CMOS outputs, and operate from -40°C to +85°C; ADCs and evaluation kits are available through Silanna's authorized distributor network.
This is a paid press release. Contact the press release distributor directly with any inquiries. General Purpose Microcontroller Unit (MCU) Global Market Report 2026 | Capture $20.29 Billion by 2030 as STMicroelectronics, TI, Infineon and NXP Accelerate Ultra-Low-Power IoT Research and Markets Fri, November 6, 2026 at 3:01 PM GMT+1 5 min read STM IFX.DE 005930.KS TXN Company Logo Players: Samsung, Texas Instruments, STMicroelectronics, Infineon and NXP. The data profiles their participation and product innovation in the general-purpose MCU market. Dublin, Nov. 06, 2026 (GLOBE NEWSWIRE) -- "General Purpose Microcontroller Unit (MCU) Market Global Report 2026" has been added to ResearchAndMarkets.com's offering. The global general-purpose microcontroller unit (MCU) market is experiencing strong growth as manufacturers expand the use of programmable, energy-efficient processing solutions across connected devices and electronic control systems. The market is projected to increase from $13.23 billion in 2025 to $14.46 billion in 2026, representing a compound annual growth rate (CAGR) of 9.3%. Historic market growth has been supported by the increasing adoption of embedded electronic…
GlobalFoundries revealed new details about its FDX Fusion technology, which promises 7nm-class performance and 8nm-class logic density using planar FD-SOI transistors without EUV lithography tools. The company targets 2030 for high-volume manufacturing of the node and shared what is next for the FDX Fusion node and the FDX program in general. This is a technology roadmap disclosure: high-volume manufacturing remains a plan/target for 2030 and has not been achieved; the source does not state the cause.
ASUS announced the XG Core Radeon RX 9060 XT LP 16 GB GDDR6, a compact external graphics solution that adds discrete graphics to compatible laptops, gaming handhelds and mini PCs through USB4 connectivity. Powered by AMD RDNA 4 architecture, it pairs 16 GB of GDDR6 memory and supports gaming, creative applications and compatible AI workloads. Users do not need to make any internal hardware modifications to access discrete graphics; the external design adds a dedicated GPU when needed without replacing the entire system.
ASUS today announced the XG Core Radeon RX 9060 XT LP 16GB GDDR6, a compact external graphics solution that adds discrete graphics to compatible laptops, gaming handhelds and mini PCs through USB4 connectivity. The product is powered by AMD RDNA 4 architecture and pairs 16 GB of GDDR6 memory with support for gaming, creative applications and compatible AI workloads. According to the source, users do not need to make any internal hardware modifications to their device to access discrete graphics. The source does not state the cause; this is a product announcement, with no pricing, availability timing, or any capacity or price change mentioned.
AMD announced that it is sampling the Versal AI Core XQRVC1902 in an enhanced space-grade lidless organic package, built for 15-year missions and tested to MIL-PRF-38535 Class Y. The device is pin-compatible with the VC1902 and brings adaptive AI compute to GEO, cislunar, and deep-space payloads without sacrificing reliability. The source does not state the cause, nor does it provide volume, pricing, or supply-timing information.
AI infrastructure company Xiilab plans to issue 10 billion won in convertible bonds (CBs) to investors including the National Growth Fund. On Nov. 6, Xiilab said its board approved the issuance of 10 billion won in its first series of unregistered, unsecured, privately placed convertible bonds, with a conversion price of 8,106 won per share and a five-year maturity; the subscription agreement is scheduled for Nov. 10, with payment due Nov. 16. The bonds carry a 0% annual coupon rate and a 0% yield to maturity, allowing the company to raise funds without interest expenses through maturity; they contain no downward conversion price adjustment (refixing) provision. Xiilab said it also secured a call option covering up to 50% of the subscription amount to mitigate dilution for existing shareholders if its share price rises, and investors may exercise their put options starting 30 months after issuance. The bonds will be subscribed by National Growth Fund sub-funds managed by Korea Investment Value Asset Management, DB Asset Management, Suseong Asset Management and Orion Capital Management, totaling 8 billion won, with the remaining 2 billion won from Orion Capital Management's KOSDAQ Venture Fund and other investors. The proceeds will be used to expand the AIDC business from infrastructure deployment into operations, including preparations for growing demand for next-generation GPUs such as Vera Rubin, broadening GPU management capabilities through the proprietary AstraGo platform and generating recurring revenue from operations and management services. The issuance is board-approved but not yet subscribed or paid, so it remains a planned event.
GPUs (Vera Rubin)AI data center (AIDC)XiilabThe Elec
Q3 DRAM revenue grew 369% YoY to nearly $200 billion, driven by ASPs reaching $1.9 per gigabit (Gb), while bit shipments rose only about 4%. Samsung led with a 38% market share, SK hynix 29%, Micron 21%, and CXMT 9%. This is a pricing super-cycle driven by AI turning memory into a scarce resource, with prices still rising and suppliers on track for new earnings records next quarter.
DRAMDDRLPDDRSamsung, SK hynix and 4 totalElectronics WeeklyPrice Movement