Micron is developing a memory architecture for humanoid robots, joining Samsung Electronics and SK Hynix in efforts to serve physical AI systems that must process sensor data while operating within battery power limits. The companies are pursuing both memory designs and storage products for robots and autonomous vehicles. The source provides no timeline, investment size or mass-production schedule; this development is an ongoing/planned R&D effort with no confirmed implementation or shipment status stated.
memory architecture for humanoid robotsmemory designsstorage productsMicron, Samsung Electronics and 3 totalDigiTimes Asia
Micron is developing edge inference memory (EIM) for physical AI applications, which stacks low-power DRAM atop a logic die to improve data processing efficiency, with mass production expected in the second half of 2028.
According to Commercial Times, citing institutional investors, Google's TPU v9 project, involving MediaTek, could see its initial tape-out delayed until January 2027 following design revisions, with shipments potentially pushed back to the first quarter of 2028; the November 2026 sampling target remains unchanged, and the production-version tape-out is expected in the second quarter. The report notes that revisions to Google's TPU v9 Core A design, codenamed Humifish... wait, Humufish, have delayed development by approximately three weeks, affecting the Core B and I/O designs handled by MediaTek. According to TrendForce, the chip is expected to adopt Intel's EMIB-T advanced packaging, with substrate yields remaining a key challenge to production scalability and cost competitiveness. Bloomberg, citing sources, reports that ByteDance is nearing an agreement to purchase MediaTek's data center chips, potentially making it MediaTek's second major ASIC customer after Alphabet's Google. Among Taiwanese ASIC design firms: GUC, supported by cloud customers' CPU and ASIC programs, reported record September revenue of NT$7.145 billion, up 97.8% year over year; Alchip posted September revenue of NT$9.979 billion, surging 496% year over year as shipments of 3nm AI accelerators increased. Faraday has obtained Samsung's 2nm process design kit (PDK) to pursue related design opportunities; per Commercial Times, Samsung's 4nm capacity is reportedly fully utilized, prompting some AI ASIC customers to evaluate its 2nm node. Faraday has also completed its design environment setup for Intel's 18A-P and begun offering design services, with plans to pursue more projects in 2027. Commercial Times, citing institutional investors, projects global AI ASIC accelerator shipments to reach 14.1 million units in 2027, up approximately 69% year over year.
Google TPU v9Core A (codenamed Humufish)Core BMediaTek, Google and 11 totalTrendForce
According to Seoul Economic Daily (with ZDNet having reported in September that Samsung is pursuing a dual-track strategy for custom HBM, tapping both its own foundry and rival TSMC), NVIDIA has asked Samsung, SK hynix and Micron to adopt TSMC's 3nm process for the base dies of its NVHBM, unveiled in August. Samsung's Device Solutions (DS) division reportedly agreed to the requirements following an internal review. The stated causes are NVIDIA's push to standardize production across suppliers to ensure consistent HBM performance and simplify product validation; additionally, industry estimates place TSMC's 3nm yields above 90%, versus around 60% for Samsung Foundry's 2nm (SF2) process. NVHBM's custom base die integrates the memory controller traditionally housed in the GPU; according to NVIDIA, the technology offers up to 30% more memory bandwidth, 15% lower HBM power consumption and 25% more XPU compute die area than standard HBM4E. The partnership is not entirely new: according to Seoul Economic Daily, Samsung and TSMC first collaborated four years ago to ensure HBM compatibility with TSMC's GPU packaging process, and this latest tie-up extends into HBM manufacturing for the first time. Samsung is keeping its own foundry in play: with TSMC's 3nm capacity facing competing demand from AI accelerators and smartphone chips, Samsung is validating NVHBM base die production on its 2nm process as a potential alternative should TSMC face capacity constraints. TSMC flagged tight 3nm capacity at its October 15 earnings call, with the node accounting for 33% of third-quarter wafer sales, its largest share among all process nodes; according to Economic Daily News, Chairman and CEO C.C. Wei announced plans for another 3nm fab in Tainan, on top of existing expansion projects in Arizona, Japan's Kumamoto and Taiwan's Southern Taiwan Science Park, with the additional fab slated to begin mass production in the first half of 2028. Samsung also continues in-house base die production for its standard HBM lineup: HBM4, which entered mass production earlier this year, and HBM4E, now undergoing production validation, both use Samsung Foundry's 4nm process for base dies, while their core dies are manufactured on 1c DRAM lines at its Hwaseong and Pyeongtaek campuses; Samsung plans to maintain its turnkey approach to base die production for custom HBM supplied to other customers, including AMD.
NVHBMNVHBM base dieHBM4NVIDIA, TSMC and 5 totalTrendForcePartnership/M&A
Kulicke & Soffa reported fiscal Q3 2026 net revenue of $330.4 million, up 123% year-over-year, with gross margin of 47.8% and non-GAAP EPS of $1.20, beating expectations. The company guided to roughly $375 million for Q4 fiscal 2026 ending October 3, 2026, citing strengthening demand across semiconductor, memory, automotive, and industrial markets.
Lee Cheol, vice president at Micron, said at the "Semiconductors Powering Humanoids" conference held as part of SEDEX 2026 at COEX in Seoul on Oct. 16 that Micron is preparing edge inference memory (EIM) for physical AI applications and is discussing with major customers the development of products suited to emerging markets including autonomous driving, humanoid robots, AR and VR. EIM is positioned between AI chips and memory to reduce data processing latency and increase bandwidth; it stacks LPDDR on top of a logic die and, unlike PIM, its computing architecture has not been disclosed, and Micron has not disclosed specific performance targets. The company is developing a method of applying EIM to LPDDR6, with industry observers expecting mass production in the second half of 2028. Micron noted that current AI models for robots use around 1 billion to 2 billion parameters, so memory bandwidth requirements are not particularly high, but as more advanced vision-language models (VLMs) emerge, LPDDR alone will become insufficient, requiring a transition to HBM, which Micron expects to eventually be adopted in humanoid robots. Micron expects future high-performance humanoid robots to require approximately 200GB of DRAM and 4TB of storage, compared with 12GB of DRAM and 256GB of storage used in current service robots. The company said it is collaborating with robot manufacturers, chipset companies and hyperscalers from the early design stages.
Glen Wilk, vice president of technology at ASM, said during a CEO Summit panel session at SEMICON West 2026, held at the Moscone Center in San Francisco on Oct. 14 local time, that traditional scaling has effectively reached the end of its useful life and that materials will no longer merely enable scaling but will drive semiconductor innovation. He explained that process miniaturization alone can no longer meet the computing performance and power efficiency requirements of AI, citing transistor gates as an example: silicon dioxide, used as a gate dielectric for more than 50 years, was replaced by high-k materials such as hafnium oxide to reduce gate leakage current and improve transistor performance. He also emphasized the need for materials innovation in thermal management. ASM's main revenue sources are ALD and epitaxy equipment; new interconnect metals such as molybdenum can be deposited using ALD, and silicon and silicon-germanium thin films for gate-all-around (GAA) transistors are grown using epitaxy equipment, so the introduction of new materials creates revenue-growth opportunities for ASM. The remarks contain no specific price, output or order data; the source does not state the cause beyond the stated rationale.
According to Wccftech, a South Korean semiconductor engineer using the X handle @TTHUEONE accused Tesla of demanding access to Samsung's internal fab information and audit rights for its AI5 and AI6 chip contracts while aggressively pushing down foundry prices; the engineer also alleged that Terafab had poached numerous employees from Samsung's Taylor fab in Texas and that Tesla turned to Samsung Foundry for its AI5 and AI6 chips because TSMC would not meet its demands. Earlier this week, Musk praised Micron, saying the company 'has been reasonable regarding future memory prices, while others have been extortionate,' remarks widely interpreted as criticism of Samsung. As background: Business Post warned in March that Terafab (planned site about a 30-minute drive from Samsung's Taylor foundry) could intensify competition for talent and equipment and flagged risks to Samsung's intellectual property and manufacturing know-how; Financial News reported on October 12 that Samsung has begun deploying advanced-node yield engineers from South Korea to its Taylor facility while expanding local recruitment ahead of planned 2nm mass production in 2027, aiming to begin operating the fab by the end of 2026, with Tesla among key customers; Seoul Economic Daily reported in September that Samsung had begun trial production of Tesla's AI5 chips at the facility with yield verification underway ahead of full-scale production in 2027. AsiaToday earlier reported that Tesla's AI5 and AI6 orders are expected to support initial capacity utilization at the Taylor fab, and that TSMC's potential involvement in Terafab could complicate Samsung's efforts to secure future chip orders. TSMC Chairman C.C. Wei confirmed the company is in discussions over Terafab, though the outcome remains uncertain, and emphasized that any fab involving TSMC would have to be operated by the company itself; per Reuters, Musk said on October 7 that his companies would build and operate Terafab, with TSMC potentially subleasing part of the facility rather than taking an operational role.
AI5 chipAI6 chip2nm processTesla, Samsung and 6 totalTrendForcePartnership/M&A
According to Reuters, SK hynix is reportedly asking the U.S. government to authorize additional semiconductor manufacturing equipment shipments to its Dalian facility in China in exchange for a commitment to build a new NAND fab in the U.S. The Dalian fab is the world's only facility to produce 192-layer floating gate 3D NAND for Solidigm's data center-grade SSDs for AI, enterprise, and hyperscale data centers; its manufacturing operations rely on equipment from American suppliers including Applied Materials, KLA, and Lam Research, whose exports to China require a U.S. government license. Solidigm has requested U.S. export licenses for additional tools, but the U.S. government has yet to authorize the shipments; last December the U.S. Commerce Department granted SK hynix a limited authorization for equipment deliveries to Dalian for 2026, but has not issued a similar equipment export authorization for 2027. Solidigm wants to install additional manufacturing equipment in the unused portions of its Dalian Mod B facility, completed in 2023, and is also considering an initial public offering as early as 2027 to raise additional capital. The proposed U.S. fab and the equipment export arrangement remain reported proposals, not approved or implemented actions.
3D NAND192-layer floating gate 3D NANDdata center-grade SSDsSK hynix, Solidigm and 5 totalTom's HardwarePolicy/Regulation
Related manufacturers
SK hynix, Solidigm, Applied Materials, KLA, Lam Research
Techwing said in a regulatory filing on Oct. 14 that preliminary third-quarter revenue reached approximately 87.8 billion won and operating profit about 18 billion won, year-on-year increases of 110.18% and 153.04% respectively, benefiting from a boom in the memory semiconductor market. Cumulative revenue for the first nine months of this year was approximately 197 billion won with operating profit of about 40.2 billion won, both already exceeding last year's full-year revenue of about 159.1 billion won and operating profit of approximately 15.8 billion won. The company attributed the growth to increased capital spending by Samsung Electronics, SK hynix and Micron, the three major memory chipmakers; as the three memory manufacturers expand their fabs and increase wafer production, demand for testing has grown and demand for memory test handlers has surged. Techwing supplies HBM testing equipment to all three companies, Micron is a major customer, and memory test handlers such as the TW350HT account for the largest share of the company's business. Revenue from a new product supplied to Samsung Electronics was also recognized in the third quarter; Techwing disclosed in August that it had been supplying HBM testing equipment worth 9.72 billion won to Samsung since June, confirmed to be its Cube Prober equipment. The Cube Prober is an HBM die testing system that performs comprehensive testing of logic and core dies and can test 256 dies simultaneously at temperatures ranging from minus 40 to 150 degrees Celsius; Techwing has been supplying Cube Prober equipment to Samsung Electronics since last year and also supplies it to SK hynix.
memory test handlersHBM testing equipmentCube ProberTechwing, Samsung Electronics and 4 totalThe Elec
ASML published its Q3 2026 earnings on October 14, 2026, saying the unprecedented spending and investment brought about by the AI boom is not slowing down, with customers already lining up to place orders for its advanced chipmaking tools for 2029 and 2030 output; CEO Christophe Fouquet said the company would struggle to meet demand despite plans announced in July to raise capacity by 30% in 2027 and another 30% in 2028. The company also confirmed it will raise prices starting in 2027, with the full effect showing in 2028. The Q3 2026 report detailed a net profit of €4.96 billion on net sales of €11.7 billion; the company projects Q4 sales of €14.2 billion to €15.2 billion ($15.9 billion to $17 billion) and boosted its full-year sales projections from at least €43 billion to €44-45 billion. ASML's shares rose 2% on Wednesday, up almost 70% so far this year, with its €586 billion market valuation making it Europe's largest firm by market cap. Additionally, Terafab will reportedly receive its first EUV chipmaking tool from ASML in mid-2027; CFO Roger Dassen said it is "still kicking the tires on the exact size of that (Terafab) opportunity."
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
On October 15, 2026, Dutch specialty foundry XIVER announced it has acquired an ASML PAS 5500 DUV lithography tool for its Eindhoven foundry, with the tool expected to be operational in 2027. The DUV stepper will complement XIVER’s existing i-line lithography equipment, enabling finer and more precisely aligned structures, including double-sided wafer patterning and sub‑micron 3D micromachining. XIVER says it will be the only mid‑size pure‑play foundry offering in‑house DUV lithography. the source does not state the cause. XIVER was carved out of Philips at the High Tech Campus in Eindhoven and acquired by Dutch investors in early 2025. Since then, it has been expanding its manufacturing capabilities in line with its customers’ roadmaps and expected capacity needs.
ASML PAS 5500 DUV lithography toolXIVER, ASMLEvertiqCapacity Change
Micron Technology on October 15 opened Micron Advanced Assembly and Test Taiwan 2 (AATT2) in Taichung, its second advanced assembly and test facility in Taiwan. The plant expands the company's packaging and testing capacity for high-bandwidth memory (HBM) as AI demand grows and HBM supply tightens. The source does not state the specific amount of capacity added.
Topco Scientific (TSC) is evaluating a new base in Boise, Idaho, and plans to strengthen its local technology, engineering, and supply chain capabilities to pursue memory and advanced-process opportunities. The company already operates service bases in Phoenix, Arizona, and Dallas, Texas. The source links this plan to the US entering a fresh wave of chip manufacturing expansion and Micron ramping up advanced memory investment. The event remains at the evaluation/planning stage and is not a confirmed implementation outcome.
Intel and Micron Technology urged Taiwanese semiconductor materials, components and equipment suppliers to establish operations in the United States, speaking at the SEMICON West 2026 Taiwan Supply Chain Forum in San Francisco on Oct. 12. Ben Fulmer, vice president of construction and sourcing at Intel Foundry, encouraged Taiwanese companies to invest locally so they are ready when selected as a supplier. Shane Smith, senior director of front-end materials strategic sourcing and supply assurance at Micron, said cooperation with Taiwanese suppliers had enabled the company to relocate a significant portion of its DRAM business to the US, noted that transporting all the equipment and materials is a challenge, and emphasized that investment in the US was necessary to develop the ecosystem. Both companies are expanding US manufacturing: Intel is investing more than $100 billion across Arizona, New Mexico, Oregon and Ohio, while Micron plans to invest more than $250 billion by 2035. The forum was organized to connect Taiwanese suppliers with US manufacturers, academic institutions and government agencies; Intel and Micron delivered keynote addresses. Participating Taiwanese companies included Topco Scientific, Challentech International, Gallant Micro Machining, United Integrated Services and Yesiang Enterprise. The urging is an appeal, not a confirmed investment commitment.
Applied Materials opened the EPIC Center in California with up to $5 billion investment, partnering with TSMC, Intel, Samsung, Micron, and SK hynix to accelerate the commercialization of advanced semiconductor technologies, particularly for AI leadership.
Applied Materials, TSMC and 6 totalThe ElecPartnership/M&A
Related manufacturers
Applied Materials, TSMC, Intel, Samsung Electronics, Micron, SK hynix
On its earnings call, ASML warned that even with a planned 30% capacity increase in 2027 followed by another targeted 30% expansion in 2028, its planned capacity would fall short of customer demand in both 2027 and 2028; Reuters, citing CEO Christophe Fouquet, notes that orders for 2028 are already pouring in, with some customers looking to secure equipment as far ahead as 2029 or even 2030. ASML expects to ship around 65 systems, up more than 50% from last year, and the outlook for DUV immersion tools has improved significantly: after initially anticipating a decline in shipments, the company now expects volumes to remain roughly flat year over year. ASML confirmed it will introduce the NXE:3800F EUV scanner in 2027, boosting wafer throughput from today's 230 wafers per hour to 260 wph, a breakthrough the company said is expected to meaningfully scale customer production capacities starting in 2028 and beyond; it is also upgrading key DUV immersion tools scheduled for delivery in 2027 to a record 330 wph. On pricing, ASML said pricing discussions with major customers have concluded, with the impact expected to be minimal in 2026, beginning in 2027 and fully visible from 2028 onward; according to The Elec, ASML is reportedly seeking a 10% price hike on replacement parts supplied to South Korea, with Samsung Electronics and SK hynix already notified. Financially, ASML posted record third-quarter net sales of €11.66 billion, up from €9.33 billion in the previous quarter; net profit surged to €4.96 billion from €2.92 billion and gross margin expanded to 57.5% from 54%. It forecasts Q4 revenue of €14.2 billion–€15.2 billion ($15.9 billion–$17.0 billion) with gross margin around 57%, and raised the lower end of full-year revenue guidance by €1 billion to €44 billion–€45 billion while maintaining the upper end, with gross margin of approximately 56%. The price impact has not yet taken effect and the capacity expansions remain planned.
SEMI expects Americas 300mm fab equipment spending to more than double from US$18.1 billion in 2026 to US$40.6 billion in 2029, driven by U.S. investment and expansions by TSMC, Intel, Micron, and Samsung.
300mm fab equipmentTSMC, Intel and 5 totalTrendForceCapacity Change
CXMT detailed its 4F² DRAM architecture roadmap, targeting DDR5 RDIMMs by end-2026, while its global DRAM market share rose to 9.5% in 2Q26. The 4F² tech reduces cell area by 33%, placing CXMT roughly two years ahead of SK hynix and Samsung. Bearish for global DRAM prices due to intensified competition and potential supply expansion.
DRAMDDR5LPDDR5CXMT (ChangXin Memory Technologies), SK hynix and 3 totalTrendForceCapacity Change
Related manufacturers
CXMT (ChangXin Memory Technologies), SK hynix, Samsung