TrendForce forecasts that server CPU shipment growth will accelerate from 31% YoY in 2026 to 43% in 2027. The forecast was published at 12:00 PM on Oct 15, 2026. AMD's Lisa Su said at #OCP 2026, "This year is really the year of inference," adding that it's also a CPU story. The post also references what AMD, Arm, and NVIDIA showed at OCP (with a link). This is a forecast, not an achieved shipment result; no price movement or capacity event is stated, and the source does not state the cause.
The board of directors of Texas Instruments Incorporated (Nasdaq: TXN) declared a quarterly cash dividend of $1.52 per share of common stock on Oct. 15, 2026, payable on Nov. 10, 2026, to stockholders of record on Oct. 30, 2026. This is a company dividend distribution and does not involve changes to capacity, pricing, or supply and demand.
On October 15, 2026, SK hynix is disaggregating SSD controller cores, developing H‑TPUs (Hynix‑Tensor Processing Units) to offload low‑level read/write command processing, metadata management, and buffer control from the main controller CPU, and plans to deploy H‑TPUs in its next‑generation enterprise SSDs and Storage Next SSD controllers.
Elon Musk posted on X criticizing memory makers for what he calls 'extortionate' pricing amid an unprecedented shortage of both conventional DRAM and high-bandwidth memory (HBM), while praising Micron as 'reasonable' regarding future memory prices and saying it is 'making the smart move for long-term customer loyalty.' The article notes that Musk did not identify the companies he criticized and provided no specific details of the pricing or the terms offered by different suppliers (which tend to be confidential), so it is impossible to determine whether Samsung Electronics and SK hynix charge substantially more than Micron for comparable products; his remarks are nonetheless seen as directed at the South Korean memory manufacturers. The article adds that Micron's own profitability is exceptionally high: in its fourth quarter of fiscal 2026, Micron reported revenue of $54.23 billion and a GAAP gross margin of 81%, up from 40% a year earlier; its cloud memory unit earned $16.283 billion with an 83% non-GAAP gross margin; its core data center storage business sold $18.002 billion worth of goods with a 90% gross margin; automotive and embedded unit sales totaled $6.824 billion with gross margins reaching 84%; and mobile and client revenue was $13.114 billion with a 90% non-GAAP gross margin. Micron has signed 26 strategic customer agreements covering more than 35% of its expected revenue through 2030, with $32 billion in customer financial commitments, most consisting of cash deposits, and the company acknowledges its new agreements reflect higher prevailing memory prices and is by no means selling memory at a discount. Separately, Samsung has reportedly secured long-term contracts covering 75%–80% of its planned memory output in 2027, and SK hynix is reportedly pursuing similar arrangements. Market insiders and analysts expect memory supply shortages to persist through at least 2028. The article also states that manufacturers have raised memory prices severalfold this time without comparable increases in production capacity, and supply still lags behind demand.
DRAMHBMNANDMicron, Samsung Electronics and 3 totalTom's Hardware
Microchip Technology has introduced the POL-05-20S-T, a radiation-tolerant gallium nitride (GaN) point-of-load (POL) converter designed for spaceborne processing systems. The device delivers up to 20 A of continuous output current in a compact surface-mount package measuring 1.51 × 1.51 × 0.45 inches and weighing 25 g, with an input voltage range of 4.75 V to 13.2 V, an adjustable output from 0.7 V to 5 V, and an operating temperature range of −55°C to +100°C at full load. For radiation tolerance, Microchip specifies electrical parameters before and after exposure to total ionizing dose (TID) levels of up to 50 krad, and the converter is characterized for single-event effects (SEE) at linear energy transfer (LET) levels up to 43 MeV·cm²/mg. Targeting high-current, fast-transient loads such as the core voltage rails of FPGAs, microcontrollers, and ASICs, the converter is available now in production quantities.
radiation-tolerant GaN point-of-load (POL) converterMicrochip TechnologyPower Electronics News
Russian chipmakers demonstrated homegrown RISC-V and Arm chips with retro games at the Microelectronics 26 (ME26) forum in Sochi on October 11, 2026, as reported in an article published October 15, 2026. The demo featured a Baikal Electronics BE-U1000 microcontroller-based NES clone, an Elvis Hypercom-U SoC, and MCST's Elbrus-based PCs. The source does not state the cause. The event status is observed demonstration.
Dutch company Prodrive Technologies said at SEMICON West 2026, held at the Moscone Center in San Francisco on Oct. 13 (the event runs Oct. 13 to 15), that it plans to open an office in South Korea between late 2026 and early 2027 to strengthen support for local customers, including Hanwha Semitech, and expand its presence through sales activities. The company said it is recruiting local employees in preparation for the office opening and expects to open "around the end of this year or early next year," considering Dongtan and Pangyo as potential locations. The Korean office will handle sales and customer support, with a small team being recruited; a representative said there are no plans to establish manufacturing facilities in Korea because Korean customers can be supported from Malaysia, China and Europe. Prodrive is also building a new manufacturing facility in Malaysia, operates production sites in the Netherlands, Boston in the US, and China, has sales offices in Japan and Israel, and employs approximately 2,300 people worldwide. It develops modules and provides system integration services for semiconductor equipment manufacturers; in South Korea, Hanwha Semitech collaborated with Prodrive to develop a hybrid bonder, and earlier this year the second-generation hybrid bonder, SHB2 Nano, was manufactured and assembled at Prodrive's Netherlands facility before being brought to South Korea. Prodrive also supplies linear motors and servo drives for magnetic levitation systems used in ASML's extreme ultraviolet (EUV) lithography equipment.
An analytical article states that agentic workloads increase CPU demand, while reasoning models, MoE architectures, and reservoir computing fuel AI acceleration, altering the balance between CPUs and GPUs. The source provides no specific manufacturers, part numbers, figures, or prices; this is background commentary.
South Korea's LX Semicon is seeking buyers for more than 60 silicon carbide (SiC) power semiconductor patents as it winds down a business it entered five years ago to reduce its reliance on display driver ICs (DDIs). According to the source, the retreat comes as Taiwan's Novatek gains ground in LX Semicon's core display chip market, increasing pressure on the company to develop new revenue sources. The patent sale is at the stage of seeking buyers and has not been confirmed as completed.
According to a report published by The Register on Wednesday, 14 October 2026 at 16:00 UTC, Gigabyte recently announced new motherboards that use DDR4 RAM, a standard that launched in 2014. For AM4 processors, Gigabyte offers the new B450M H ARGB with two DDR4 slots and the B550 X-Eagle with four slots, while Intel-platform buyers can get the B760M Aorus Elite. The article states this is partly due to the current price rises and component shortages induced by large-scale, cloud-based LLMs. In the People's Republic of China, several vendors sell new DDR3 motherboards for circa 15-year-old server CPUs, such as Huananzhi's X79-4D or X79 Supreme, or the Machinist X79 V2.82H, all using Intel's LGA-2011 socket, which launched in 2011. The DDR5 standard was agreed in 2020, with mainstream PC adoption following in 2021. The article also argues that Linux lets older hardware remain in service, making second-hand kit more usable than ever.
SK hynix has detailed plans for an in-house core named a Hynix Transaction Processing Unit (H-TPU), intended to reduce reliance on power-hungry CPU cores. The plan was presented by Koo In-jae, the company's VP of SoC design, at ISOCC 2026 (the International SoC Design Conference), and remains a plan rather than a shipping product: H-TPU has yet to be proven in an actual enterprise SSD controller product, and no date has been given. The H-TPU uses a four-stage, in-order pipeline, is placed in clusters to handle repetitive tasks such as read/write command processing, metadata management and buffer control, with more than 100 H-TPUs in total; management remains with the traditional Arm CPU, making it an offload rather than a replacement. The cores are claimed to draw only about 2.5% of a general-purpose CPU core's power, with around 4% of its area (the CPU core used for comparison is not named). The reported cause is that SSD controller CPUs are at their limit on power and area, and PCIe bandwidth doubles every generation while controller CPU clocks and control capability reportedly cannot keep up. SK hynix reportedly plans to apply H-TPU across its NAND products, including next-gen enterprise SSD controllers, and to high-bandwidth flash (HBF), which the outlet said is due next year; the H-TPU is reportedly already being used in 'Storage Next', Nvidia's name for the joint AIN P project with SK hynix.
TDK introduces the Micronas HAR 3025 dual-die stray-field robust motor position sensor with redundant design, ASIL D ready (SEooC) per ISO 26262:2018 and stray-field compensation per ISO 11452-8:2015; samples are available now and start of production is planned for the first half of 2027. The source does not state the cause.
On Oct. 12, Nvidia unveiled its Vera CPU Rack specification, introducing a server rack design centered on central processing units (CPUs) for the first time, optimized for agentic artificial intelligence (AI) workloads. According to Nvidia, agentic AI relies heavily on CPU computing; under its architecture, the GPU-based Vera Rubin NVL72 rack handles traditional large language model (LLM) training and inference, while the Vera CPU Rack is designed to process agentic AI workloads. The Vera CPU Rack is based on Nvidia's third-generation MGX architecture; MGX is a modular reference specification for data center servers that Nvidia provides to server manufacturers, with partners including Dell, Supermicro and Quanta using it to design and manufacture their own server products. The rack integrates computing, networking, power and cooling systems into a common modular platform. Nvidia also said it submitted designs for the Vera Rubin rack and drawer-style compute trays based on MGX 1.2, an updated version of its third-generation MGX specification, to the Open Compute Project's (OCP) Open DC Initiative, intended to help server manufacturers develop next-generation AI factory infrastructure based on Vera Rubin and bring it to market more quickly. MGX 1.2 is compatible with 19-inch server rack equipment, supports up to 250 kilowatts (kW) per rack, and introduces rack-level liquid cooling; Nvidia said standardized interconnects, reconfigurable input/output (I/O) slots and fanless compute trays reduce assembly and maintenance time.
Vera CPU RackVera Rubin NVL72 rackMGX 1.2Nvidia, Dell and 4 totalThe Elec
Asia Vital Components (AVC) unveiled new liquid-cooling solutions, including direct liquid cooling (DLC) and an In-Row CDU, at the Open Compute Project (OCP) Global Summit held October 12–15 in San Jose, US. The company is expanding liquid cooling beyond AI GPUs and CPUs to support rising AI computing demand and increasingly heat-intensive data-center racks.
direct liquid cooling (DLC)In-Row CDUAsia Vital Components (AVC)DigiTimes Asia
The article discusses how to measure the performance demands of agentic AI, assessing CPU bottlenecks, software stack (SW stack) efficiency, and system-level tradeoffs. It is a technical background discussion with no price movement, capacity change, policy, halt, or corporate event, and provides no specific figures, dates, or manufacturer information.
On October 14, 2026, TrendForce reported from the OCP Global Summit 2026 that Arm said the CPU market is expected to surge from $52 billion in 2026 to $300 billion by 2030 due to agentic AI redefining data center architectures; Arm’s AGI CPU debuted in March 2026.
In an SK hynix Newsroom article series [Inside the Research] Part 2 dated October 14, 2026, the company detailed its research paper "Co-Optimizing Cell, Non-Cell, and Page Schemes for Energy Efficient Analog Computing in 3D FeNAND," published in the 2025 IEEE International Electron Devices Meeting (IEDM), authored by Won-Tae Koo and Jihun Kim of SK hynix's Global RTC. The research proposes a 3D FeNAND-based analog compute-in-memory (A-CiM) architecture that merges storage and computation, and by co-optimizing cell, non-cell (interconnect/word line), and page schemes, achieves approximately 10 to 20 times higher throughput and around 4,000 to 5,000 times greater energy efficiency than conventional 3D arrays under the same chip-area constraints; the article notes these figures reflect 3D FeNAND's high-density advantage over conventional 2D nonvolatile memory arrays widely studied as A-CiM candidates. Reported optimization results include: using a low-resistance word line metal improved throughput by up to 16% and energy efficiency by up to 3%; air gaps reduced word-line parasitic capacitance by approximately 60%, improving throughput by 1.2 times and energy efficiency by two times; jointly optimizing word line resistance and capacitance improved throughput by up to 1.4 times and energy efficiency by up to 2.2 times; in-page pipelining (IPP) doubled energy efficiency and improved throughput by nine times versus the baseline; and splitting a conventional 16KB page into four 4KB subpages improved energy efficiency by 1.2 times and throughput by 1.7 times. The team also demonstrated four-level-per-cell operation in the low-current region with stable data retention at the 300mm wafer scale. The article states the research shows NAND's potential to expand beyond a storage medium into a computing medium; the source contains no price, shipment, capacity, or market-condition information.
Fabless semiconductor startup Bolt Graphics announced the bring-up of its Zeus test chip and the opening of a Taipei office to manage manufacturing partnerships as it prepares for production of its first fully commercial GPU, which is planned for general availability in 2028 (a plan, not an achieved result). Built on Bolt's rendering technology architecture, the 12 nm TSMC test chip accelerates electromagnetic simulation and path tracing in software and hardware, enabling full-product simulation without physical prototypes. Bolt stated its GPU exceeds performance expectations, validating the chip architecture, and that it has proven complex physics algorithms can be executed on fixed-function hardware, accelerating training of advanced robotics and world models. The source does not state the cause (this is a company progress announcement, not a price or supply change).
Intel proposes an open software blueprint that positions open software as connective tissue linking CPUs, GPUs, data, and business applications, emphasizing upstream contributions to frameworks like PyTorch, SGLang, and vLLM, aiming for Day 0 readiness, describing Xeon as the operational foundation for inference, extending to disaggregated inference, and presenting microservices and enterprise agent toolkits as building blocks; the article notes it is a blueprint, not proof of execution.
During her keynote at the OCP Global Summit 2026 on October 12, AMD CEO Lisa Su expressed strong confidence in the CPU market's growth prospects and called 2026 the "Year of Inference." She also raised AMD's forecast for the 2030 CPU TAM to $200B+. This is a forward-looking forecast target, not an achieved result; the source does not state the cause, only the confidence expression.