HBM4 deal; war’s impact on supply chain; U.S. AI framework; GPU smuggling; restart of H200 processors for China; SK hynix’s memory prediction; TFLN photonics deal; test and metrology platforms; NoC verification automation; Tesla probe; AI distillation attacks.
The Iran War’s toll on the chip industry is widening. Over 95% of Taiwan’s energy is imported, causing the country to secure alternative sources. Korea is also heavily dependent on energy imports from the Middle East.
Shortages of key materials are cropping up everywhere. Helium from Qatar, the second largest producer behind the U.S., is constrained by hostilities in the Persian Gulf. And tungsten, a metal used in semiconductor production, is in short supply due to Chinese export restrictions and competing military demands. The price has more than doubled this year, reports Bloomberg.
AI chip exports
Nvidia is restarting production of its H200 AI chips in anticipation of being able to send them to Chinese customers.
The U.S. Dept. of Justice charged three individuals with conspiring to divert billions of dollars of servers with Nvidia GPUs to China, in violation of export rules. One of those charged is a co-founder of Super Micro Computer, according to CNBC.
Memory
Samsung will supply HBM4 memory for AMD‘s next-gen AI accelerator and DDR5 for CPUs, an expansion of their existing collaboration.
Micron completed its $1.8B acquisition of PSMC’s P5 site in Tongluo, Taiwan, to expand production of leading-edge DRAM products, including HBM. The company also started high-volume production of HBM4 designed for Nvidia’s Vera Rubin.
At the GPU Technology Conference (GTC), Nvidia and partner highlights include:
Vera Rubin platform for AI computing, with 7 new chips for AI factories;
The semiconductor market surpassed $830B in 2025 and could reach $1T in 2026 at the current growth rate, according to Omdia.
Charles H. Bennett and Gilles Brassard are the recipients of the 2025 ACM A.M. Turing Award for their contributions to the field of quantum information science and quantum cryptography.
The EU granted Open EU Foundry status to Silicon Box‘s packaging and test project in Italy, a designation that will speed up approvals and priority access to pilot lines.
The UK plans to spend up to ~$2.7B on quantum computing, sensing and navigation, and networking over the next four years.
The Netherlands‘ Future Network Services consortium will receive an additional $246M to advance 6G network chips and technology.
Siemens invested $165M to expand its production facilities for low- and medium-voltage electrical equipment for data centers in North and South Carolina, USA.
UMC, Wavetek, HyperLight, and Jabil are teaming up to deploy thin-film lithium niobate (TFLN) photonics in hyperscale AI data center interconnects.
Imec will transfer its LPCVD silicon nitride photonic IC platform to New Origin, a Dutch photonics foundry now building its first fab, as part of an ongoing collaboration on PIC technology.
Lightmatter and the Open Compute Project launched an initiative to develop open specifications for a shared reference architecture that enables interoperable co-packaged optics in AI systems.
Taiwan Explained: Why China Claims It, and Why the US Is Involved (CFR)
Taiwan’s Semis: Backbone Of The Global Tech Economy (Eurasia Review)
New Technology
Manufacturing, Test
Onto Innovation launched an extensible inspection and metrology platform for front-end and advanced packaging applications, with high throughput and the ability to detect defects as small as 150nm by using multiple optical techniques, multi‑mode illumination, and new ML algorithms.
Teradyne debuted a PCBA and sub-assembly test platform that integrates structural, parametric, high-speed interconnect, and functional tests. Teradyne also launched an opto-electric automated test platform for high-volume silicon photonics and co-packaged optics manufacturing.
EDA
Axiomise unveiled an app that automates the verification of custom NoC implementations with exhaustive guarantees generated by formal proofs. The app can be configured to handle different bus protocols, different channel types, as well as complex routing rules.
Keysight released an AI inference emulation platform, simulating real-world workloads and workload patterns.
Siemens EDA introduced a domain-scoped autonomous AI agent that plans and orchestrates multi-tool and multi-agent complex semiconductor, 3D-IC, and PCB system workflows spanning design, verification, and manufacturing sign-off.
Infineon introduced two new high-voltage Intermediate Bus Converter (HV IBC) reference designs, ultra-low noise XENSIV TLE4978 hybrid Hall and coil current sensor, and the XDM700-1 system for monitoring and reporting IC for high-side or low-side current and voltage sensing. The company also expanded its digital power controller portfolio and AI data center voltage regulation portfolio.
A UCLA research team developed a technique to improve electrical current in perovskite semiconductors with a “contact-induced charge-transfer doping method using silver oxide nanoclusters formed at the interface.”
Volkswagen opted to use chips from China’s Xpeng instead of Nvidia‘s.
Rivian and Uber will jointly deploy 50,000 autonomous robotaxis in San Francisco and Miami before moving to more cities by 2031.
Government actions
NHTSA announced an expansion of its probe into Tesla’s full self-driving technology in the wake of crashes, with a focus on the system’s abilities to handle reduced roadway visibility conditions.
Think tank IAPS highlighted the fast-growing threat of AI-related attacks, with these two reports:
AI distillation attacks, defined as “systematic efforts to extract capabilities from frontier AI systems to train competitor models.” The firm recommends adding offenders to the US Bureau of Industry and Security Entity List, assessing sanctions, and utilizing IP laws and NIST to establish minimum standards.
Highly autonomous cyber-capable agents, including systems that autonomously conduct multi-stage cyber attacks comparable to today’s top illegal hacking or nation-affiliated threat groups. The 159-page investigation dives into the implications and policy recommendations of the emerging threat.
Alerts
Nine critical flaws in low-cost KVMs (keyboard, video, mouse) were identified across four vendors, giving hackers physical access to every connected machine, reports Eclypsium.
Keysight established a SW bill of materials to assist companies in complying with new EU CRA regulations.
Infineon and Nvidia are collaborating on safe and secure humanoid robots, utilizing digital twins of Infineon’s actuators and sensors and other virtual model technology.
Sandia National Laboratories is now offering a cybersecurity residency, a fast-tracked 6-month hands-on training program that includes memory forensics.
In an alliance with Micron and NSF, Syracuse University is paying each student $2,400 in order to generate interest in working with semiconductor, optics, and quantum technology.
Arm teamed up with Anglia Ruskin University to open the ARU Arm AI Lab in Cambridge, England.
SIA sees military veterans as one potential pipeline for addressing the chip industry’s growing workforce shortage in the U.S. Meanwhile, the UK is also ringing a similar warning bell, projecting 79,000 extra AI, cybersecurity and semiconductor workers will be needed by 2035, in this report by Frontier Economics.
Trending Video
Improving Yield Through Shared Data: Jayant D’Souza, technical product director at Siemens EDA, talks about the underlying drivers for sharing data, how it can be done securely, and why this is suddenly getting so much attention.
InP and SiPho join CMOS as critical technologies. Lasers, CPO and OCS will be everywhere (indium phosphide, silicon photonics, co-packaged optics, optical circuit switch).
Reducing variation in manufacturing, monitoring behavior over time, and targeting specific workloads can have a big impact on power, performance, and area/cost.
This website uses cookies to improve your experience while you navigate through the website. The cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. We do not sell any personal information.
By continuing to use our website, you consent to our Privacy Policy. If you access other websites using the links provided, please be aware they may have their own privacy policies, and we do not accept any responsibility or liability for these policies or for any personal data which may be collected through these sites. Please check these policies before you submit any personal information to these sites.
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.
Leave a Reply