SpaceXAI and Tesla head Elon Musk believes that AI is not good enough for the most advanced computer programming tasks. Musk shared his thoughts in a tweet in response to his earlier remarks about his firm's Grok AI relying on reinforcement learning after finishing its pretraining. The latest Grok iteration, Grok 4.8, relies on 2.5 trillion parameters, which marks a significant bump over the previous Grok models. Elon Musk Says SpaceXAI's C/C++ Training Stack Is Written By Humans Since AI Is Not Good Enough September has been a good month for SpaceXAI after the firm released its Grok 4.6 model […]

Read full article at https://wccftech.com/elon-musk-admits-ai-isnt-good-enough-for-extremely-high-performance-software-says-grok-4-9-should-match-fable-class-models/

Intel and NVIDIA could cooperate on developing and building AI CPUs, says a fresh report from Korean financial firm GF Securities. The firm's analyst, Jeff Pu, made the remarks as part of his mid-quarter coverage for Intel and outlined that the firm could also benefit from headwinds stemming from agentic AI capabilities of GPT-6 AI models. Pu also discussed Intel's Foundry business and outlined that the yields of the 18A chip manufacturing process technology family were at 80%, with those of other larger products improving from June. Intel Might Gain Foothold In NVIDIA's Ecosystem Through x86 CPU, Says Analyst Intel's […]

Read full article at https://wccftech.com/intel-14a-production-capacity-tipped-to-sit-at-6000-wafers-per-month-by-2027-end-with-18a-yields-at-80-says-analyst/

The Taiwan Semiconductor Manufacturing Company (TSMC) is planning to increase its 3-nanometer and 2-nanometer manufacturing process technology capacities in 2027, suggests a supply chain report from Taiwan. TSMC, the world's leading contract chip manufacturer, is facing high demand from the AI buildout and high-performance computing (HPC) chips. Its 2 and 3-nanometer technologies are the most advanced in its portfolio, with earlier reports having suggested that the next-generation 1.4-nanometer technology could enter production in 2027. TSMC To Bump 2-nanometer & 3-nanometer Production Capacity By 22% and 16% By Mid-2027, Says Report Over the past couple of weeks, several reports have discussed […]

Read full article at https://wccftech.com/report-tsmcs-2-nanometer-push-outpaces-3-nanometer-growth-with-capacity-aimed-at-110000-wafers-per-month-by-h2-2027/

Chinese electronic design automation (EDA) firm Empyrean Technology is using agentic artificial intelligence to accelerate its chip design process. The firm's chairman, Liu Weiping, shared the details at a conference in Shenzen, China, last week. His remarks come as Chinese firms continue to manage around the constraints levied on them through chip equipment sanctions. According to Weiping, his firm has managed to cut down design times for some tasks by as much as 75% through using agentic AI, and his comments came days before China's Ministry of Industry and Information Technology issued a new development plan for the information and […]

Read full article at https://wccftech.com/chinese-chip-firm-empyrean-slashes-circuit-design-time-by-75-using-agentic-ai-amidst-race-to-agentic-ai-semiconductor-design/

The Taiwan Semiconductor Manufacturing Company (TSMC) might double its chip-on-wafer-on-substrate (CoWoS) capacity from 2026 levels in 2028, suggests a report from the Taiwanese media. TSMC's CoWoS is the leading packaging technology for AI GPUs and is used by major firms such as NVIDIA. The details suggest that the capacity expansion comes at a time when existing constraints have led to a spillover to other firms, such as Taiwan's UMC and Amkor. Intel Expected To Reach Up To 45,000 Wafers Per Month Of Equivalent CoWoS Capacity In 2028, Says Report The booming demand for TSMC's CoWoS technology and the resulting shortage […]

Read full article at https://wccftech.com/report-tsmc-to-double-cowos-capacity-by-2028-as-ai-chip-shortage-spills-over-to-rivals/

The Taiwan Semiconductor Manufacturing Company's (TSMC) next-generation 1.4-nanometer production could kick off in April, suggests a new report in the Taiwanese press. TSMC's 1.4-nanometer process technology, also called the 14A manufacturing node, will succeed the firm's 2-nanometer technology. According to the details, TSMC's 1.4-nanometer facility in Taichung could enter trial production in April 2027, with mass production expected in the second half of the year. TSMC Submits Application To Build Two Temporary Offices At Taichung 1.4-nanometer Facility, Says Report According to the details, the technology park where the plant is being built has revealed that construction of the plant started […]

Read full article at https://wccftech.com/tsmc-reportedly-accelerates-1-4nm-production-to-april-2027-pulling-its-most-advanced-chips-a-full-year-ahead-of-schedule/

China's Semiconductor Manufacturing International Corporation (SMIC) was one of the fastest-growing foundries in the second quarter, according to fresh data from market research firm TrendForce. SMIC grew its revenue by 20% during the second quarter and was the third biggest chip manufacturer in the world in terms of revenue, as it comfortably sat above GlobalFoundries. Taiwan's TSMC, courtesy of its advanced manufacturing process technologies and stable revenue, remained the biggest foundry and commanded 72.5% of the overall market. Consumer Electronics Helped SMIC To Grow Revenue 20% In Q2, Shows Data The booming demand for artificial intelligence memory chips has reshaped […]

Read full article at https://wccftech.com/smic-outgrows-tsmc-and-samsung-with-20-surge-overtaking-globalfoundries-as-worlds-third-largest-foundry/

With the high demand for high bandwidth memory (HBM) outpacing supply, Samsung Electronics has reportedly allocated more than 50% of its 4-nanometer capacity to making the memory chips. The report comes courtesy of ZDNet Korea, and it concerns the base die for the memory chips. As opposed to the HBM DRAM modules, which are fabricated with custom manufacturing nodes, the base die is fabricated using technologies used to build logic chips. Samsung Aiming To Expand HBM4 Memory Chip Production In Q3 2026, Says Report As part of its coverage, ZDNet quotes sources from both within and outside Samsung to comment […]

Read full article at https://wccftech.com/report-samsungs-4nm-lines-run-at-full-tilt-as-hbm4-chips-eat-up-to-60-of-foundry-output/

China's Naura Technology Group, a key domestic semiconductor manufacturing equipment provider, claims to have developed a two-step etch process for manufacturing 3D DRAM chips. 3D DRAM chips are the next generation of DRAM technology as they seek to overcome the constraints of current chips when it comes to cell density. A paper published by Naura in the IEEE journal explains progress made in etching uniformly across the 64 layers of 3D DRAM to shift the focus to vertical manufacturing instead of the horizontal constraints encountered by non-EUV lithography. Chinese Chip Equipment Manufacturer Naura Details Breakthrough In Semiconductor Etch Engineering That […]

Read full article at https://wccftech.com/chinese-chip-making-equipment-firm-naura-claims-big-breakthrough-that-could-enable-cxmt-to-overcome-euv-sanctions-make-3d-dram-chips/

The CEO of German lens manufacturer Zeiss's semiconductor business, Frank Rohmund, believes that while China might be 14 years behind current Western chip manufacturing technology, restrictions on the country will serve to accelerate its pace of innovation. The executive made the remarks to Bloomberg during the SEMICON conference in Taiwan, and he added that China's advances in deep ultraviolet (DUV) lithography have to be evaluated on several parameters before comparing them to non-Chinese technology. Zeiss Executive Says China's Progress In DUV Lithography Technology Is Unsurprising Owing to Western restrictions, news about Chinese lithography developments are almost always intensely scrutinized. For […]

China is unlikely to match Dutch chip manufacturing ASML's EUV machines for ten years, suggests a new report from investment bank UBS. EUV lithography scanners are a key link in the global advanced chip manufacturing industry, and without them, making the latest chips is complicated and costly. US restrictions on China prevent the country's major chip manufacturers from acquiring the EUV machines, and UBS outlines that China can overcome the gap for older immersion-based DUV lithography much sooner than it can for EUV technology. Chinese EUV Chip Manufacturing Technology Development Is Similar To ASML's In 2004, Says UBS Back in […]

Read full article at https://wccftech.com/china-trails-asmls-advanced-euv-chip-making-technology-by-10-years-says-investment-bank-ubs-as-progress-is-stuck-at-asmls-2004-level/

Intel's chief financial officer, David Zinsner, believes that improvement on the firm's 14A manufacturing process technology is the fastest in its history since the 22-nanometer node was launched in 2012. The executive made the remarks at Deutsche Bank’s 2026 Technology Conference and discussed the process technology's defect density. A technology's defect density measures flaws or contamination per unit area of a wafer and is typically measured in the number of defects per centimeter squared. Intel CFO David Zinsner Says Company Is Designing Products For 14A Chip Manufacturing Technology Node Intel's 14A manufacturing process technology is quite important in the firm's […]

Read full article at https://wccftech.com/intels-david-zinsner-says-14a-improving-fastest-since-2012-node-betting-foundrys-future-on-tsmc-fight/

With workers at Samsung's memory chip production facility having successfully negotiated bonuses for themselves, those at Micron's factories in Taiwan appear to be angling for similar benefits. After the US firm's Taiwan union held a meeting in July to study Samsung's strike, a report from Reuters quoting the union members suggests that workers are moving towards a strike unless management meets their demands. Micron Union Workers Say More Than 80% Of Workers Back Strike The July Micron meeting saw union management respond to worker queries about a strike. It also sought opinions about Micron's policies regarding a 200% bonus cap […]

Read full article at https://wccftech.com/micron-workers-in-taiwan-move-toward-strike-as-80-surveyed-back-walkout-over-bonus-caps/

The growing demand for high bandwidth memory (HBM) from the AI buildout has led to Samsung having booked 70% of its memory production capacity for long-term agreements (LTAs), according to a new report from the Korean media. These agreements are providing Samsung's customers with significant price reductions compared to the HBM spot price, which is as much as five times higher than that of the LTAs. Samsung's Primary LTA Contract Partners Are Big Tech Players Such As Microsoft, Google & NVIDIA, Says Report According to a report in the SEDaily, the high demand for HBM has led to tightness in […]

Read full article at https://wccftech.com/samsung-locks-70-of-memory-capacity-into-2031-deals-long-term-hbm-deals-run-5x-cheaper-than-spot-as-ai-demand-chokes-the-dram-market/

The Taiwan Semiconductor Manufacturing Company (TSMC)'s bonuses for the second quarter outpaced the firm's revenue growth, as shown in the firm's filings with the Securities and Exchange Commission (SEC). During the second quarter, TSMC paid out NT$36 billion in bonuses, which marked a 50.6% growth over the year ago figure of NT$35 billion. At the same time, the firm's second quarter revenue of NT$1.2 trillion marked a 36% annual growth. TSMC Paid 4% Of Its Operating Profit As Bonuses To Employees In Q2 2026 Bonuses have become a contentious issue in the semiconductor industry following the AI boom. The ball […]

Read full article at https://wccftech.com/tsmc-bonuses-surge-50-and-outpace-revenue-growth-as-sec-filings-expose-the-ai-talent-wars-real-cost/

Owing to the surge in DRAM prices, one kilogram of memory chips manufactured in Korea is now worth as much as 620 grams of gold. The global memory shortage, ushered in by the ongoing AI infrastructure buildout, has led to triple-digit percentage increases in memory chip prices. Owing to the price rises, memory companies SK hynix and Samsung have also paid generous bonuses to their employees, with Micron's workers in Taiwan also purportedly interested in similar bonuses. Data From Korean Trade Statistics Agency Shows One Kg Of DRAM Costs As Much As 620 Grams Of Gold Data from Korea's Trade […]

Read full article at https://wccftech.com/dram-prices-explode-401-as-one-kilogram-of-korean-memory-now-costs-as-much-as-620-grams-of-24k-gold/

According to a fresh report from The Wall Street Journal, chip designer NVIDIA Corporation has stopped its revenue sharing and credit support initiative to support AI growth. The Journal's report comes a day after NVIDIA's latest earnings report, which saw the firm report $108 billion in commitments to OpenAI and other firms as part of its initiative to support data center infrastructure development. Concerns About Antitrust Scrutiny Lead NVIDIA To Pause Credit Support & Revenue Sharing Agreements With Cloud Providers Back in July, NVIDIA had announced in a blog post that it was partnering with cloud providers to set up […]

Read full article at https://wccftech.com/nvidia-quietly-freezes-its-cloud-credit-support-deals-after-internal-talks-warn-of-antitrust-exposure/

With NVIDIA's fiscal second quarter earnings official, Amazon and the firm announced earlier today that they will expand their collaboration for using NVIDIA's AI chips in Amazon's AWS cloud computing platform. Through the deal, Amazon will now deploy up to three million NVIDIA AI GPUs in its global infrastructure to power workloads for agentic AI, automation, physical AI and other applications. Amazon Signs Up For NVIDIA's Vera CPUs As Part Of Its Efforts To Provide "Broadest Possible" Set of Compute Options In AWS As part of its release, Amazon outlined that the partnership expansion with NVIDIA will complement its custom […]

Read full article at https://wccftech.com/amazon-commits-to-3-million-nvidia-gpus-in-aws-tripling-a-deal-that-started-at-just-one-million/

As part of its earnings release for the second quarter of fiscal 2027, NVIDIA shared that it has entered land, power and shell guarantees for its AI cloud partners that are worth $108.5 billion. NVIDIA's guarantees and agreements with its customers have caused considerable debate, with the firm's detractors accusing it of relying on circular financing to create a safety net for its customers. NVIDIA Assures That Its Project Guarantees Will Drop As The Year's Pass & Eventually Turn Into Revenue One interesting outcome of the AI infrastructure buildout has been NVIDIA providing commitments to its customers and helping them […]

Read full article at https://wccftech.com/nvidia-ties-103-billion-in-commitments-to-openai-site-that-could-return-200-billion-in-revenue/