Semiconductors
Semiconductors news — Thursday 19 November 2026
150 stories from 10 sources, every one in English. Each links to its publisher.
-
White House Cyber Chief Sean Cairncross Leaving Post
Sean Cairncross, a top White House official who leads the Office of the National Cyber Director, is leaving his position. He was fired because he lacked the cyber credentials to perform the job, according to a person close to the White…
-
Nvidia builds physical AI ecosystem with Taiwan partners
The AI wave is moving beyond virtual cloud conversations into the era of physical AI, which interacts with the real world. Nvidia, Macnica Galaxy, and SignalPro have joined forces with Epson hardware to showcase an end-to-end ecosystem…
-
CHT chair urges spectrum reshuffle as 5G mmWave overlaps LEO satellite lane
Chunghwa Telecom (CHT) chairman Chih-cheng Chien on November 17 called for Taiwan to begin comprehensive planning and feasibility assessments for the 26GHz and 28GHz bands, saying 5G millimeter-wave spectrum currently overlaps with the…
-
Taiwan's economy minister opens Taiwan-EU industry policy dialogue in Belgium and Poland
Taiwan's Minister for Economic Affairs Ming-hsin Kung led a delegation to Belgium and Poland from November 14 to 19, meeting with the European Commission's Directorate-General for Internal Market, Industry, Entrepreneurship and SMEs (DG…
-
ANPEC confirms larger 1Q27 price hike to pass through foundry cost increases
Power management IC maker ANPEC Electronics has confirmed it will raise product prices in the first quarter of 2027, with a steeper increase than its previous adjustment, to offset rising foundry costs, General Manager Kang Sheng said at…
-
US AI chip curbs hit Southeast Asian data centers
Following controversy over Nvidia chip flows, the US Department of Commerce's Bureau of Industry and Security (BIS) has tightened export controls, requiring companies to obtain prior approval before exporting, reexporting, or transferring…
-
Chairman of CSIA Shaojun Wei projects China's 2026 chip sales to rop CNY1.5 trillion on AI boom
China's semiconductor sales are projected to surpass CNY1.5 trillion (approx. US$225 billion) in 2026, a sharp 89% jump from a year earlier, as surging artificial intelligence demand fuels rapid expansion across the domestic chip sector…
-
Samsung retools Austin chip plant to target SpaceX business
Samsung Electronics has sold KRW286.4 billion (approx. US$215 million) in chipmaking equipment to its US subsidiary as it retools its Austin, Texas, plant for specialty semiconductors, including chips it reportedly plans to supply to…
-
Samsung says neocloud debt poses limited risk to AI industry growth
Samsung Electronics rejected concerns that debt-financed expansion by neocloud providers could trigger a broader slowdown in artificial intelligence investment, saying these companies account for only about 8% of the overall AI market. The…
-
ByteDance's Azure deal exposes a new fault line in US-China AI controls
The recent upheaval surrounding Nvidia and former CEO Jensen Huang has sharpened an already intensifying debate in Washington over how far US restrictions on frontier AI should extend. Credit: Shihmin Fu
-
China audits 17 automakers over supplier payment delays
China's Ministry of Industry and Information Technology (MIIT) has commissioned the China Association of Automobile Manufacturers to launch a new audit of supplier payments at 17 major automakers, intensifying a nationwide crackdown on…
-
US tightens advanced chip foundry rules, restricting Chinese firms' use of third-country intermediaries
The United States is further tightening export controls on advanced semiconductors, extending restrictions beyond Chinese companies' access to AI computing power through overseas data centers to include advanced chip manufacturing and…
-
Physical AI robots open global path for Taiwan's precision suppliers
As AI applications move from cloud-based virtual conversations into the physical world, SignalPro Chairman Tien Sheng-yu has outlined the commercialization logic of Physical AI and the key advantages Taiwan's supply chain can leverage. He…
-
Democrats win as Trump backs geothermal for low-carbon baseload
The energy-structure revolution driven by the enormous power demand of AI data centers (AIDCs) is accelerating. Alphabet's Google recently signed the world's largest enhanced geothermal system (EGS) procurement deal to date with geothermal…
-
Walsin Lihwa targets AI, nuclear power and seven other growth industries
Walsin Lihwa Corporation said it would accelerate a strategic transformation as it approaches its 60th anniversary in 2026, shifting its focus toward nine emerging industries. The Taiwanese industrial group will target AI data centers…
-
Samsung memory procurement costs surge 263% as phone prices rise 13%
Samsung Electronics' smartphone business is facing mounting cost pressure as AI server memory demand squeezes supply for mobile memory. The average purchase price of memory used in Samsung phones rose 263% in the first three quarters of…
-
Taiwan robot suppliers target 2027 validation, 2028 volume production
Tesla's Optimus has entered early production at a California factory, raising expectations that the humanoid robot industry is moving from prototype development toward commercialization. Taiwan suppliers, while staying out of complete…
-
Humanoid robots need two breakthroughs before revenue takes off
The humanoid robot market remains in the process of finding a viable business model, while supply chain companies say product designs, component costs, end-user applications, and customer requirements all remain unsettled. Specialized…
- Global goodwill assets hit $10.5tn on M&A boom but interest rate risk looms
-
Score TP-Link Archer with full Wi-Fi 7 support and 10 Gbps wired networking for incredible 55% off its regular price
Internet connection speeds are consistently rising most everywhere in the world, and more than a few spots already have multi-gigabit service available. Making those kinds of speeds available throughout your house, however, requires a…
-
Alphabet Energy Deals Delayed Amid Weak Muni Market
Two prepaid energy deals that would provide $3 billion to Google parent company Alphabet have been paused due to muted investor demand and market shakiness, Bloomberg reported. The deals are still expected to go ahead, but the delay…
-
Apple Developing AI Agent to Rival Meta and OpenAI
Apple is developing an AI agent that can browse the web, use apps, and write code, according to recent job postings. The iPhone maker is building an AI assistant that can use a computer much like a person does, the postings say. Such a…
-
Anthropic Closes First Trading Day at $97 a Share, 17% Above IPO Price
Anthropic’s shares closed up 17% at $97 on the company’s first day as a public company, well down on its opening trade. Even so, the company finished the day with a market capitalization of around $1.75 trillion on a fully diluted basis…
-
Nvidia Nears Acquisition of Diffusion AI Developer for More Than $500 Million
Nvidia is in advanced talks to acquire Inception, a two-year-old startup that develops a type of AI that aims to perform faster than the one that powers products such as ChatGPT and Claude, according to three people with knowledge of the…
-
Memory R&D Is a High-Stakes Game
The need for significant investment in R&D is nothing new for memory makers, but Micron Technology’s creation of a dedicated facility that looks beyond the next node illustrates why the industry cannot
- International students in Japan face language hurdles when seeking jobs
- Japanese automakers brace for 3rd year of global sales decline
- Anthropic closes 17% above IPO price after 34% opening surge
- Rural Japan fears new visas will divert foreign workers to Tokyo
-
New Windows 11 lock screen zero-day unlocks BitLocker-protected PCs with just a few keystrokes
Considering the current status of the cybersecurity landscape, some may well say that "0-day" has lost its meaning, The latest high-stakes vulnerability pertains to a bug in Windows 11 (and possibly 10), wherein an attacker can partially…
-
AI data center power demand will double by 2030 thanks to $4 trillion in spending, adding an entire Great Britain to energy strain
The AI data center build-out is pushing electricity grids across the globe, with many regions facing demand from projects that would exceed the current peak load that each system could support. According to the International Energy…
- Anthropic IPO draws rush of Japan retail investors
- U.S. Steel agrees with union to raise wages 19.25% over 4 years
-
CEO Interview with Khaled Alashmouny of AIDAChip
Khaled Alashmouny is the founder and CEO of AIDAChip. He spent 13 years at Apple leading analog and mixed-signal design teams and has about two decades of experience building silicon. He founded AIDAChip to solve the problem he saw on…
- Japan ethylene output depressed despite normal naphtha supplies
- Itochu, Sony partner to market artist-branded merchandise in IP push
-
Why Bending Spoons Could Tap Convertible Bonds Next
Since going public in July, Bending Spoons has tapped the debt and equity markets to raise money mostly to buy more tech firms, and to run its growing portfolio. One financing option the company hasn’t yet used is convertible bonds, but…
-
Meta Bets on $300 Muse Charm to Boost AI Hardware Business
Meta Platforms plans to charge $300 for its forthcoming Muse Charm device and offer user subsidies, according to a person familiar with the matter, betting that the pocket-sized AI gadget will help energize its hardware business and draw…
-
Community-Focused Academic Brings AI Tools to Colorado Farmers
Many farmers and ranchers in eastern Colorado are curious about artificial intelligence, but they don’t have a local tech-savvy person to show them how to use it. Xue “Alice” Dong has made that her job. Dong, an IEEE senior member, is an…
-
Wall Street Short-Seller Alleges AMD Has Business Ties to China
Shares of Advanced Micro Devices dropped 9% on Thursday after Wall Street short-selling research firm Hunterbrook Capital published a report detailing AMD’s business relationships in China, including ties to companies blacklisted by the…
-
ServiceNow Builds an AI Weapon to Target Salesforce Customers
After a late summer rally, the Saaspocalypse slump in software stocks such as ServiceNow and Salesforce has returned. One reason may be signs that Anthropic’s expansion into tools outside of coding—such as tools to automate legal, finance…
- India to let Japan use firing ranges for surface-to-ship missile tests
- AI volatility 'expected' as winners, losers emerge: JPMorgan co-president
-
November 2026 Top500 New Systems Are Few and Far Between
We check out the new systems in the November 2026 Top500 list and see some of the major trends shaking out The post November 2026 Top500 New Systems Are Few and Far Between appeared first on ServeTheHome .
-
Google Selects Bezos’ Blue Origin for Data Center Satellite Test
Google will partner with Jeff Bezos’s Blue Origin rocket company for Project Suncatcher, Google’s experimental bid to test solar-powered data centers in space, the companies announced Thursday morning. The partnership comes as companies…
- Myanmar committed genocide against Rohingya, ICJ rules in landmark verdict
- Japan spacecraft supported by NEC, others set to reach Mercury
- Japan weighs raising new-drug prices closer to overseas levels
- Anthropic opens up 34% in trading debut after record IPO
-
Anthropic Opens Trading at $111.5, Passes $2 Trillion in Market Cap
Anthropic shares began trading at about $111.5 per share, 34% above its IPO price of $83. This gives Anthropic a market capitalization of $2 trillion on a fully diluted basis, making it one of the world’s largest companies. The…
-
Gartner Warns of Risks Behind Sovereign Tech Claims
BARCELONA, Spain — Addressing enterprise and government executives at the recent Gartner IT Symposium/Xpo , Gartner analysts René Büst and Bettina Tratz-Ryan delivered a critical assessment of
-
The AI Chip Boom’s Uncomfortable Arithmetic
The semiconductor industry’s biggest risk may be misunderstanding its own success. When revenues rise dramatically, the natural response is to expand capacity, secure supplies and prepare for sustained growth. But those decisions become…
- APEC live: China releases leaders' statement after its disappearance from APEC website
- Japan's Logisteed to buy US peer Neovia, expanding 3rd-party logistics
- Steelworkers union USW announces tentative labor pact with U.S. Steel
-
Alibaba CEO Warns of Impact From New U.S. Chip Rule
Alibaba Group CEO Eddie Wu said Thursday that Washington’s new rule barring Chinese companies from accessing Nvidia’s advanced AI chips through overseas data centers “will affect next year’s products and services.” Alibaba is stepping up…
-
Will Anthropic’s IPO Lead To An Exodus of Anthropic Employees?
Today is a major milestone for Anthropic . The five-year-old maker of Claude beat its archrival OpenAI to the public markets, and many of its employees are going to become millionaires, if not billionaires, when the company starts to trade…
-
Intel continues layoffs with cuts to its IT department, hundreds of employees reportedly affected
Intel has reduced its workforce by approximately 27,800 employees under Lip-Bu Tan, based on the company's reported headcount figures, but this does not seem to be enough. The company has reportedly initiated another round of layoffs, this…
- APEC live: Joint statement disappears from APEC website
- Alibaba trains next-gen AI model on own chips amid tighter US curbs
-
Shaping the Form, Function, and Future of Humanoid Robotics
Humanoid robotics is moving rapidly from concept toward real-world applications, creating new engineering challenges across sensing, movement, assembly, reliability and long-term serviceability. As robots
-
Anthropic Is Pushing New Products, and Big Discounts, to Justify Its Huge Valuation
To justify its enormous new valuation as a public company, Anthropic needs to show that it can expand quickly in other areas of business the way it did with coding. Its early efforts are showing signs of promise—and highlighting the steep…
-
PSS 3.1 Opens New Possibilities for Hardware and Software Verification
Accellera’s approval of the Portable Test and Stimulus Standard (PSS) 3.1 marks an encouraging step toward a more connected, reusable approach to verification. Announced on November 11, the release is the first update since PSS 3.0 in…
-
Supermicro sued by shareholders over claims Nvidia executive’s wife knew about illicit China sales
More information about the alleged knowledge of chip diversion by a Supermicro executive is being revealed as shareholders’ civil lawsuit against the company progresses. According to The Telegraph , the investors alleged that there was “no…
- Xi, Trump, Putin promise to raise energy supplies and lower nuclear risk
-
Developer vibe codes Windows 11 alternative operating system for amd64, arm64, and riscv64 hardware
There’s a new vibe-coded OS in town going by the name of LiberNT . Its name provides a big clue to its nature: Liber (free), NT (New Technology architecture, as in Windows NT ). As it has been released under GNU’s GPL3.0 license, LiberNT…
-
Intel 18A yields hit 70% as company signs Microsoft as EMIB customer
Lip-Bu Tan, chief executive of Intel, said on Thursday that yields of chips made using the company's latest 18A (1.8nm-class) fabrication process have achieved 70%, which is more or less in line with what one would expect from an all-new…
-
Pakistan undersea cable fault causes internet failure near Strait of Hormuz
Pakistan's telecommunications authority, the PTA, has confirmed that internet connectivity issues affecting the country's residents are related to the failure of an international submarine cable. SEA-ME-WE 5 is located almost seven miles…
- Alibaba quarterly profit more than triples as cloud services boost sales
- APEC live: Joint statement disappears from website
-
Why a Computer Scientist Thinks Superintelligence Won’t Happen
Many people were caught off guard this past summer when employees of one of the world’s leading AI companies began casually debating whether their products would lead to human extinction. It started when Anthropic researcher Jacob Coxon…
-
Silicon Motion demos SM8466 PCIe 6.0 SSD controller, expected to ship by 2028 — controller boasts improved quality of service, samples imminent
Silicon Motion is showing off its PCIe 6.0 SSD controller for data center applications for the first time at the SC26 conference in Chicago . The final iteration of Silicon Motion's SM8466 controller was taped out recently, and drives…
-
Intel Diamond Rapids shows signs of life in SC26 demo running live server
Intel’s next-gen Xeon Diamond Rapids CPUs appeared at Supercomputing 2026 ( SC26 ) in two separate booths, but neither was from Intel directly. Most interesting was a running server at the CXL booth, showcasing a demo with a Diamond Rapids…
-
Ukraine deploys homemade 'flying modems' in daring drone attack to destroy Crimean Bridge, ditching Elon Musk's Starlink
Ukrainian forces ditched Elon Musk's Starlink wireless system and used their own relay network of "flying modems" to maintain a live feed of an attack on the Crimean Bridge, even amid Russian attempts to block the signal. The attack, which…
- Xi, Trump, Putin put nuclear arms, Ukraine war on table at historic summit
- Xi tells Trump to 'do right thing' on Taiwan as trade truce remains in flux
- APEC live: Xi-Trump-Putin trilateral ends
-
Three 'e-waste' Nvidia V100s hit 100 tokens per second on Qwen 3.6 35B in a $1,042 homelab rig
A homelab builder benchmarked three used Nvidia V100s on Qwen 3.6 35B, achieving about 100 tokens per second in a custom setup costing $1,042 with just “e-waste” GPUs. The smaller, dense Qwen 3.6 27B manages about 24 tokens per second on…
-
Get 32GB of DDR5 for $240 with this blistering Intel gaming PC bundle
Buying a bundle in the Black Friday sales could save you a ton of money on expensive RAM and SSDs, and this latest one from Newegg shaves $240 off the price of 32GB of DDR5. The price on this bundle is so good Newegg won't even show it to…
-
Save 36% on this 4TB Samsung 990 Pro SSD upgrade for your gaming PC, now just 17.5 cents per GB
SSDs are expensive right now, thanks to AI, so picking one up on sale is the best way to do it. Luckily, thanks to the start of Amazon's Black Friday sale, you can pick up a 4TB Samsung 990 Pro for $699.99 right now . That's a $100 price…
-
Critical Java vulnerability puts millions of sites, services, and devices at risk of remote code execution
We've just had a few high-stakes security vulnerabilities making systems administrators run around with their heads on fire, and the deluge doesn't look like it's stopping any time soon. Java, the programming language that Oracle has…
-
Nvidia Reports $115.8B in Quarterly Revenue Under Interim CEO
Nvidia reported third-quarter revenue of $115.8 billion on Tuesday, more than twice its revenue in the same period last year. Cloud providers and tech giants kept buying its AI processors at a strong pace.
-
China’s Moonshot AI Scraps $150 Million Deal to Access Nvidia Chips in Indonesia
Moonshot AI, the Beijing-based developer of the Kimi models, has scrapped a $150 million deal to access Nvidia Blackwell chips through a data center in Indonesia, after the U.S. this week announced a rule to restrict Chinese companies from…
-
DDR6 RAM shrinks by 44% while delivering 4X DDR5's baseline bandwidth — JEDEC previews 12.8 GT/s modules with up to 1 terabyte of capacity per slot
Today's best RAM will look huge compared to upcoming DDR6 memory. JEDEC has just released a preview of the DDR6 specification, confirming that next-generation memory modules will be 44% shorter than DDR5. That is only the physical aspect…
- Japan's Takaichi criticizes 'arbitrary' export curbs in brief China visit
- APEC live: Xi says China, Russia, US must cooperate on global challenges
-
China expands 6G pilot program to eight provinces and cities
China is accelerating its 6G industrial rollout, moving from communications equipment prototypes and experimental spectrum access to field trials, as the domestic supply chain advances into system-level validation. On November 17, China's…
-
Exclusive: TikTok to Appear at House Meeting Over National Security Concerns
TikTok will attend a public roundtable with the U.S. House Select Committee on China on Dec. 15 to answer questions relating to China’s influence on its algorithm and how the company protects Americans’ data, a congressional aide said on…
-
Nvidia doubles down on Taiwan supply chain as AI squeezes substrate and OSAT capacity
Nvidia is deepening its ties with Taiwan's artificial intelligence supply chain by issuing capacity prepayments to fund expansion across IC substrates, advanced packaging, and testing—locking in production capacity for the next three to…
- BlackRock CEO warns tighter AI regulations could lead to US recession
- Virtual 'Jensanity' rocks Taipei in support for Nvidia chief
- Japan Tobacco cuts profit forecast by $8.7bn due to asset seizure in Russia
- Trump, Xi and Putin hold historic trilateral meeting
- South Korea rides AI boom to new height as exports hit $1 trillion
-
Nvidia maintains product roadmap as AI chip supply chain impact stays limited
Following Nvidia's latest earnings conference, the company's operating performance, outlook, and product development timeline showed no apparent impact from the departure of former CEO Jensen Huang. AI chip demand remains strong, although…
-
Tighter US chip rules set to favor Qualcomm and MediaTek
The US Department of Commerce has introduced stricter chip technology controls affecting China and several Southeast Asian and Latin American countries, requiring case-by-case approval for shipments of products built on processes below…
-
ASE invests NT$248.9B in Renwu K52 plant for AI testing
Taiwan Fore Electronics, a subsidiary of ASE Technology Holding, will invest NT$248.9 billion (US$7.54 billion) in a new K52 plant at the J block of Kaohsiung's Renwu industrial park, with construction scheduled for phased completion in…
-
Beyond Pass/Fail: Characterizing Silicon Margin With PVTW
Production test is necessarily built around decisions. PASS. FAIL. Bin 1. Bin 2. Ship. Do not ship. Characterization has a different objective. It asks not only whether the device meets specification, but how far it is from the boundary…
-
AI Accelerators Drive Changes In Power Delivery Networks
Key Takeaways: AI accelerators and GPUs consume thousands of amperes at increasingly low voltages, causing a shift to vertical power delivery that brings regulation closer to loads. Advances in integrated voltage regulators and passive…
-
How Virtual Twins And Machine Learning Optimize DRAM Yield
As memory technologies continue to scale, manufacturing processes are becoming increasingly complex. Additional process steps, tighter process windows, and new sources of variation are making it progressively more challenging to reach…
-
From Metrology To Autonomy: Scaling Yield, Quality, And Innovation
With the growth of increasingly complex devices, improving yield, quality, and manufacturing speed requires a new approach. Across 14 presentations and posters presented at SEMICON West 2026, Intel Foundry leaders and engineers shared…
-
What Is Driving Silicon Wafer Demand Recovery?
Global silicon wafer shipments are forecast to grow 11.8% in 2026, according to the Q3 2026 SEMI Silicon Wafer Market Monitor . This increase is primarily due to strong demand for AI, which is projected to drive shipments through 2029…
-
Simulation of Plasma Chambers for Predictive Simulation of Semiconductor Manufacturing
Plasma etching and deposition processes increasingly determine the manufacturability of advanced semiconductor devices, where smaller features, more complex geometries, stricter wafer-uniformity requirements, and new chemistries demand…
-
Detachable Fiber Connectors Pose New Challenges For Photonics
Key T ake a ways: CPO and NPO require new approaches for connecting optical fibers to chip packages. Detachable connectors can simplify assembly, test, and replacement of optical components. Numerous bespoke designs exist today, but…
-
Can AI Make Its Own Manufacturing Decisions?
Key Takeaways: AI can identify promising process conditions and recognize manufacturing problems, but those results still require physical validation before they can be trusted in production. Incomplete manufacturing data and uncertainty…
-
US AI curbs disrupt Southeast Asia shipments; MSI sees worst-case impact of one month
The recent US expansion of export controls on artificial intelligence (AI) products is rippling through the Southeast Asian operations of Taiwanese system integrators, as well as motherboard and graphics card makers. Micro-Star…
-
Microsoft reaffirms Nvidia partnership after CEO departure
Microsoft's chief executive has expressed regret over the departure of Nvidia's chief executive while reaffirming confidence in the chipmaker and its partnership with Microsoft. The comments, reported by TechCrunch , followed the Nvidia…
-
Chinese battery-materials firms restructure overseas plants as US rules tighten
Chinese battery-materials suppliers are selling stakes and transferring operations at overseas plants to preserve access to the US market under tighter American tax-credit rules, even as they retain financial and technology ties to the…
-
Intel lands first customer for 18A-P foundry process
Intel has signed its first external customer for its 18A-P chipmaking process, advancing its efforts to expand its contract manufacturing business and compete with TSMC. Credit: AFP
- BlackRock CEO warns tighter US AI regulations risk recession
- APEC live: Takaichi says Japan's position on Taiwan remains the same
- Trump touts 'balanced' China ties as Xi highlights stability
- Anthropic hikes IPO to $83 on strong demand; $1.5tn valuation possible
-
Ford ships first Michigan LFP cells, but the batteries are still licensed from China
Ford has begun shipping the first lithium-iron-phosphate (LFP) battery cells from its Michigan plant, marking a significant step in the automaker's attempt to localize lower-cost EV battery production in the US. Credit: Ford
-
MSI targets over 10% non-PC revenue with on-premises AI
Micro-Star International (MSI) is reshaping its product and operating structure as AI computing demand expands, export controls tighten, and rising memory prices pressure the PC market. President Jeans Huang said PC shipments are expected…
- APEC live: Trump says US-China trade is balanced
- Trump and Xi hold 3rd summit of year as friendly vibes mask deep divisions
-
Xi outlines four APEC cooperation pillars as trilateral talks with Trump, Putin draw focus
The 33rd Asia-Pacific Economic Cooperation (APEC) Economic Leaders' Meeting opened in Shenzhen, China, on November 19, with Chinese President Xi Jinping proposing four pillars—trade, connectivity, innovation, and development—to deepen…
-
Analysis: CXMT sees edge AI driving memory growth as enterprise and device markets emerge as twin engines
ChangXin Memory Technologies (CXMT) held an earnings briefing on November 17, during which company director and president Cao Kanyu outlined the outlook for enterprise, edge-device, consumer, and automotive memory markets. Credit: AFP
-
White House reportedly mulls public disclosure rules for US firms using Chinese AI
The White House is reportedly mulling new rules that would require US companies to publicly disclose their use of Chinese AI models, a move that would seek to discourage them from using the tools while avoiding hard restrictions. The…
- APEC live: Trump tells Putin 'We've had a great relationship for many years'
- Vietnam's new ASEAN internet cable part of $1bn Japan partnership
-
Towards Fair Resource Sharing in Agentic LLM Serving
arXiv:2611.23558v1 Announce Type: cross Abstract: As LLM agents increasingly execute complex, multi-step tasks on shared inference infrastructure, resource contention across the end-to-end LLM serving stack causes severe and uneven…
-
Run, Read Back, Refine: Observation Fidelity in Execution-Grounded EDA Agents
arXiv:2611.19056v2 Announce Type: replace Abstract: Design engineers lose the most time orchestrating flows and patching the gaps where closed EDA tools refuse to interoperate. We present an execution-grounded LLM agent for PyAether…
-
RFAmpDesigner: A Two-Level LLM-guided Optimization Framework for Automated Differential CMOS RF Amplifier Sizing
arXiv:2605.10093v2 Announce Type: replace Abstract: Automated RF amplifier sizing is challenging because the design process is inherently hierarchical, the parameter space is high-dimensional and strongly coupled, and high-fidelity circuit…
-
A Runtime System for Mitigating Crossbar Port Contention on SM-Partitioned GPUs
arXiv:2611.24458v1 Announce Type: cross Abstract: To avoid scheduling delays when running concurrent workloads on a GPU, Streaming Multiprocessor (SM) partitioning mechanisms such as CUDA Green Contexts restrict each workload to separate…
-
MicroAssert: Microarchitecture-Aware Structured Assertion Generation for CPU Verification
arXiv:2611.24872v1 Announce Type: new Abstract: Automatic assertion generation reduces manual verification effort, yet existing approaches struggle to capture microarchitectural behaviors and cross-region interactions in CPUs. We present…
-
Mechanistic Critics for Sample-Efficient NPU Design-Space Exploration
arXiv:2611.24733v1 Announce Type: new Abstract: NPU design-space exploration for LLM inference mainly relies on an expensive simulator that often returns only scalar feedback such as latency and power. Such feedback can rank evaluated…
-
TechReport: Implementing CHERI in gem5
arXiv:2611.24460v1 Announce Type: new Abstract: Despite countless mitigation efforts, memory safety issues still prevail. The recently discovered security issues in PHP underline this fact. One of the most promising approaches to solving…
-
Hardware Acceleration of Vision-Language-Action Models with Weight Precision Rising over the Flow Steps
arXiv:2611.24234v1 Announce Type: new Abstract: Vision-language-action (VLA) models such as GR00T control robots through an action head: a diffusion transformer that turns noise into a chunk of actions in a few flow-matching steps. Every…
-
ReCIM-HD: A 40 nm On-Device Learning Accelerator with Hyperdimensional Computing via Importance-Based Hybrid Digital-ReRAM CIM
arXiv:2611.23973v1 Announce Type: new Abstract: ReCIM-HD is an on-device learning (ODL) accelerator that integrates gradient-free hyperdimensional computing (HDC) with resistive RAM (ReRAM)-based analog compute-in-memory (CIM). ReCIM-HD…
-
Inference-Time Scaling in Agentic Hardware Verification Workflows: A Multi-Strategy Characterization Across Protocol Complexity
arXiv:2611.23905v1 Announce Type: new Abstract: Inference-time scaling, allocating more compute during generation rather than pre-training, has demonstrated dramatic quality improvements for LLM reasoning tasks. Its behaviour in structured…
-
Programmable Probabilistic Computer with 1,000,000 p-bits
arXiv:2606.25313v2 Announce Type: replace-cross Abstract: Many problems in physics and machine learning require sampling the collective states of large numbers of interacting variables. On modern graphics processors, repeated movement of…
-
PULSAR-V: I/O-Aware Post-Update Attestation for RISC-V
arXiv:2611.24135v1 Announce Type: cross Abstract: Post-update attestation proves that a device installed the intended firmware, not that the image stayed protected until the report was signed. In a multi-master SoC, a master outside the…
-
Compute Before You Digitize: ADC-less Mixed-Signal Processing for Scalable Cryogenic Qubit Readout
arXiv:2611.23448v1 Announce Type: cross Abstract: We introduce DIRAC, an ADC-less in-cryostat qubit readout architecture that uses analog computing to redraw the analog-digital boundary. Directly operating on the front-end's analog output…
-
Reliable arithmetic circuits using less reliable (sub-digital) logic
arXiv:2611.24591v1 Announce Type: new Abstract: The frequent static failure of CMOS logic gates at ultra-low voltages has historically been perceived as a strict barrier to extreme voltage scaling: this paper introduces a statistical…
-
Dissecting an Unexplored SASS Control-Code Field: Barrier Exemption and Programmable Warp Priority
arXiv:2611.24437v1 Announce Type: new Abstract: Every instruction in NVIDIA's native GPU ISA (SASS) carries a control code that guides hardware scheduling, and prior work has decoded its known fields to optimize GPU kernels. However, one…
-
HETEROVLM: Edge Multi-Request Vision-Language Model Serving on a Heterogeneous Versal SoC with AI Engines and Programmable Logic
arXiv:2611.23878v1 Announce Type: new Abstract: Edge devices in vehicles and robots serve concurrent requests with different deadlines to a vision-language model (VLM), each mixing a compute-bound encoder and prefill with a memory-bound…
-
Everlight ramps up AI optical communications push, targets double-digit growth in 2027 on robust coupler orders
Taiwan-based Everlight Electronics is accelerating its entry into the AI optical communications market through overseas technology partnerships, automated packaging investments and strategic upstream supply agreements. With customer…
-
Qualipoly accelerates shift to electronic chemicals, targeting over 50% revenue contribution
Growing demand for AI servers and high-speed switches is driving upgrades in high-frequency, high-speed copper-clad laminate (CCL) materials, prompting Qualipoly Chemical to accelerate its transition toward electronic chemicals. Credit…
-
UBright eyes semiconductor, passive components and smart acoustics growth
As competition intensifies in the optical film market, UBright Optronics is accelerating its product and technology transformation. The company is expanding into semiconductor packaging and optical waveguides while investing in passive…
-
Samsung's silence on 2027 8-inch MPW plan worries South Korean IC design firms
Samsung Electronics' foundry division is gradually winding down its 8-inch process operations but has yet to announce its 2027 8-inch multi-project wafer (MPW) plan, raising concerns among South Korean IC design firms about losing access…
-
NCT builds capacity to commercialize gallium oxide with Mitsubishi Electric
Novel Crystal Technology (NCT), a Japan-based developer of gallium oxide (Ga₂O₃) wafers, is accelerating commercialization by building its own production capacity rather than adopting a fabless model. The company is also working with…
-
Nvidia's role in connecting Alibaba with Alice Huang ahead of US chip curbs deepens Megaspeed scrutiny
It has been reported that Nvidia's sales organization is the party that introduced Chinese technology giant Alibaba Group to Alice Huang, the executive at the center of an escalating US-Singapore semiconductor diversion investigation…
-
DIGITIMES Insight: Supply snags, not election or Nvidia CEO exit, behind 4Q26 server shipment cut
Global server shipments for the fourth quarter of 2026 have been slightly revised down. Still, the cause lies in component mismatches and weak production yields on the supply side, not order cuts by cloud service providers (CSPs)…
-
Google looks to bypass Broadcom with direct HBM purchases
Google is looking to buy high-bandwidth memory (HBM) directly for AI chips developed with MediaTek, potentially cutting billions of dollars in procurement costs and changing how Samsung Electronics, SK Hynix and Micron Technology negotiate…
-
SemiFive, PFN stack DRAM directly on logic as HBM alternative
South Korea's SemiFive and Japan's Preferred Networks (PFN) are developing an AI inference chip that stacks DRAM directly above a logic die, offering an alternative to the side-by-side processor and HBM layout used in conventional 2.5D…
-
Commentary: What it would take for Jensen Huang to return to Nvidia
Jensen Huang says he is ready to come back. Nvidia's board says it is not. DIGITIMES analyst Luke Lin argues that the deciding voices sit outside the company, and that timing may matter more than what either side wants. Credit: AFP
-
GlobalFoundries invests US$10 million in India-based Cyient Semiconductors
Cyient Semiconductors has approved a US$10 million strategic equity investment from GlobalFoundries, bringing the Indian company closer to a major semiconductor manufacturer. The deal is intended to expand its custom ASIC capabilities…
- APEC live: Trump says US has been 'Pacific power for hundreds of years'