South Korean semiconductor insiders and analysts believe that Samsung's Foundry business is catching up with a main rival. Earlier this month, TSMC leadership openly discussed an unprecedented demand for 2 nm wafer products. Industry moles believe that the Taiwan's top chipmaker is still ahead of contenders in nearby nations. As a result of an alleged leading and comfortable position, TSMC is reportedly upgrading its state-of-the-art facilities with brand-new equipment—indicating a push into 1.4 nm fields. According to a fresh Chosun Biz news article, Samsung engineers are in the process of narrowing the gap between their 2 nm Gate-All-Around (GAA)—also known as SF2—manufacturing node process and TSMC's equivalent technology.

Last month, leaks suggested SF2 trial yields passing the ~40% mark—in comparison, a ~60% figure was uttered by TSMC insiders. Chosun Biz's sources claim that the South Korean foundry team is close to getting their "2 nm process performance evaluation" into a crucial final stage. Yesterday's report posits that NVIDIA and Qualcomm are in the equation; these VIP clients are purportedly considering SF2 as a "second channel" option. Chosun Biz reckons that Team Green is sizing up Samsung Foundry flagship tech for next-gen commercial and enterprise GPUs. Meanwhile, the San Diego-based smartphone processor specialist could be eyeing up SF2 (for a future AP). The latest inside track info points to 2 nm GAA trial production runs breaking beyond aforementioned (approximate) 40% yield rates. TSMC 2 nm wafer charges are reportedly greater than expected, so big industry players are allegedly investigating "cheaper" non-Taiwanese production avenues.

Qualcomm Technologies, Inc., a technology leader enabling connected intelligent computing, and HUMAIN, a national artificial intelligence (AI) champion advancing Saudi Arabia's goals for AI, today signed a Memorandum of Understanding (MOU) with the intent to enter a strategic collaboration for the development of next-generation AI data centers, infrastructure and cloud-to-edge services to meet the rapidly growing demand for AI across the globe, including the Kingdom of Saudi Arabia. The MOU was signed during the Saudi-US Investment Forum in Riyadh, which was held during the official visit of President of the United States Donald Trump to the Kingdom of Saudi Arabia.

The data centers and supporting ecosystem are intended to provide both government and enterprise organizations with access to high-performance and power efficient CPU and AI cloud infrastructure and cloud-to-edge services. These offerings will enable the deployment of AI solutions that can make real-time predictions and decisions, in addition to significantly increasing the availability and value of advanced AI-enabled applications.

The Slormancer is now out of Early Access. Hello everyone! We're two indie devs who started this project over 8 years ago as a hobby. We've always been huge fans of hack'n'slash games so we decided to do one. We started from the bottom: watching YouTube tutorials on how to move sprites, draw pixel art, or generate dungeons procedurally. From there things escalated slowly. Four years later, The Slormancer launched in Early Access, and to our surprise, it found more success than we ever hoped for. That allowed us to go full-time and keep building the game we always wanted to play.

The Slormancer is now officially out of Early Access. Our game is the result of countless hours of work, but it was made possible thanks to the incredible support, feedback, and passion of our player community throughout its development. From the bottom of our hearts, thank you for being part of this adventure! Fight hordes of enemies led by The Slormancer, an evil warlock from the past, in an absurd yet epic Action RPG! Experience fast-paced gameplay, loads of shiny loot, tons of collectibles, countless hours of content, and carefully crafted pixel art.

Flow Computing, the pioneer in licensing on-die, ultra-high-performance parallel computing solutions to CPU vendors of all architectures - today announced it has successfully achieved a critical milestone in Flow's development roadmap towards commercializing its parallel processing ecosystem. Capable of increasing any CPU architecture by up to 100X, the company has been actively developing a compiler that enables source code to take advantage of its acclaimed Parallel Processing Unit (PPU) architecture - that compiler today entered Alpha testing.

Through the first target compilations, it has been determined that simple parallel workloads consist of a massive amount of loops in RISC-V CPU models without PPU assistance. Whereas in RISC-V CPU models incorporating the PPU, the amount of these loops is significantly reduced by recompiling the existing code, demonstrable proof that it is indeed possible to achieve a significant performance boost with a PPU-enhanced CPU design at up to 100X performance.

Closer to the start of 2025, Kepler L2 shared "PlayStation 6" inside track info. The venerable leaker—of mostly AMD behind-the-scenes details—theorized that Sony's sixth-gen home console was based around an early fork that branched off from a mysterious "GFX13" GPU architecture. Earlier today, this target identifier has appeared again. Kepler L2's latest investigations have pulled the "GFX13+" target from kernel-level codebase updates. Data miners have often pulled compelling pre-release tidbits from official repositories; AMD's software engineering team open-source developments are quite visible. Over a year ago, similar detective work connected the "GFX12" IP to RDNA 4—nowadays, better known as the Radeon RX 9000 graphics card family. Despite Team Red's official insistence that RDNA 4 was a priority on desktop platforms, supposed discrete mobile variants were leaked last month. In addition, RDNA 3.5 seems to be a mainstay for at least another generation of integrated graphics solutions.

Looking beyond these current-gen graphics architectures, AMD leadership has already previewed a "unified" path forward. Last September's announcement revealed the convergence of RDNA (gaming) and CDNA (enterprise) GPU technologies: dubbed "UDNA." Present day leaks do mention a far out "RDNA 5" line (also "Navi 5x"), but in the same sentence as "UDNA." Kepler L2 commented on the (screen captured) presence of "ENABLE_WAVEFRONT" and "ENABLE_WAVEGROUP" kernel code properties: "I think it's related to SWC (Streaming Wave Coalescer/pseudo out-of-order execution). Each SIMD takes multiple wave32/wave64 (a wavegroup) as inputs and reorders the work items of each wave to reduce execution divergence." Fresh rumors suggest the distant RDNA 5/UDNA avenue being good enough to expand into higher-end gaming card territories. So far, RDNA 4 has hit a ceiling with Navi 48 GPU-powered Radeon 9070 XT offerings.

The status of Intel's Arc Xe2 "Battlemage" generation is still an unknown quantity—recent leaks have alluded to a repeat rollout of existing GPU designs, albeit in the field of productivity. Team Blue has hinted about upcoming second-gen professional-grade cards, but tech enthusiasts are still fantasizing about more potent gaming options. Certain industry watchdogs reckon that Intel's graphics division has moved onto getting Arc Xe3 "Celestial" discrete solutions out of the door. Sections of a speculated C-series could act as natural gap fillers; taking the place of "absent" B750 and B770 options. These theories were buoyed earlier on in May; with insiders linking "Celestial" to a pre‑silicon validation platform.

Since late last year, it has been very quiet on the Arc Xe4 "Druid" development front; older "Visual Compute" roadmaps (now completely irrelevant) alluded to a 2025+ launch of fourth-gen designs. x86deadandback has stumbled on new-ish evidence of Xe4 "behind-the-scenes" movement—yesterday's social media post included a (data mined) screenshot of Dawn LUCI CQ activities (dated "Jan 16"). At the time, company engineers seemed to be actively working on "Gen 15/Xe4" (aka "Druid"). x86deadandback's update aligns with official (December 2024) statements—back then, Tom "TAP" Petersen confirmed that his hardware colleagues had polished off "Xe3 IP," and were already shifted onto "the next thing."

Samsung Display announced on May 13 that it will take part in Display Week 2025, held from May 13 to 15 at the McEnery Convention Center in San Jose, California. The event, hosted by the Society for Information Display (SID), brings together display companies and experts from around the world to share cutting-edge technologies and R&D achievements.

Samsung Display is set to unveil a range of next-generation display technologies, including the industry's first non-cadmium 400-nit EL-QD and a 5,000 pixels-per-inch (PPI) RGB OLED on Silicon (OLEDoS) at Display Week 2025, reinforcing its leadership in cutting-edge panel technology. The company will also highlight its leadership in OLED innovation through a range of future-oriented technologies, including organic photodiodes (OPD), advanced sensors capable of measuring biometric data such as heart rate and blood pressure directly from light generated by a panel touched by a patient, and a high-resolution microdisplay that delivers 5,000 PPI in a compact 1.4-inch form factor.

Developer and publisher Tripwire Interactive today announced that the highly anticipated next chapter in the studio's legendary Killing Floor franchise, Killing Floor 3, will launch worldwide on July 24 for PC (via Steam and Epic Games Store), the PlayStation 5 system, and the Xbox Series X|S console systems. Following a delay from the game's original March 25 release date, the development team has been focused on using the additional time to address feedback from the player community. Since the delay, the studio has shared updates directly with fans showing progress in everything from design, artwork, UI and animations. Additional development updates will be shared via the game's official forums and social channels, further demonstrating the team's commitment to evaluating player feedback and being transparent with their progress. Digital pre-orders for the Standard, Deluxe and Elite Nightfall Editions have been reinstated and are currently live on the respective platform store pages.

"Since delaying the launch of Killing Floor 3, we've dedicated the additional development time to responding to the community's feedback, ensuring the next chapter in the much-loved cult action/horror franchise lives up to our shared vision," said Bryan Wynia, Creative Director at Tripwire Interactive. "The support from our fans has been incredible, and with the new July 24th release date in sight, we're thrilled to bring that vision to players not just at launch, but through the continued support of new features and content that is a hallmark of our studio."

Late last month, La Quimera did not launch on its promised release date (April 26). At the time, the game's development team—Reburn—sent out an apology (via Discord), but their short statement did not delve into too much explanatory detail. Industry watchdogs proposed revised launch windows; potentially later on in 2025. Apparently, last month's "version 1.0" review code was not well received—content and performance-wise. Additionally, a mid-month closed beta program was reportedly nixed within 24 hours of kicking off—early community feedback suggests participants experiencing a myriad of problems. As promised during non-launch week, Reburn has provided a more comprehensive update about forward progress—in the renewed form of La Quimera; available now in Early Access.

Within the title's rebooted Steam page, Reburn outlines a revised release schedule (subject to change): "La Quimera, our narrative-driven sci-fi FPS, is now live on Steam in Early Access!..We plan to remain in Early Access for up to 18 months. This timeline may adjust based on feedback and the pace of development, but our goal is to deliver a fully polished experience within that timeframe...The Early Access version currently includes the first four levels of the campaign, featuring fully integrated narrative and gameplay elements. It offers approximately five hours of story-driven content, designed to demonstrate the core experience we are building toward. The current build is stable and fully playable, though bugs and balance adjustments are expected as we continue refinement." Interested parties can give this "still in progress" iteration a try; Steam early access pricing is $29.99.

In February, we announced Battlefield Labs as our most ambitious community collaboration in franchise history to validate the future of Battlefield. Today, we'll update you on our initial learnings and how we'll continue to scale testing into the future.

Across our four initial play sessions with a small group of core Battlefield players across Europe and North America, we've completed thousands of hours of gameplay, had hundreds of thousands of player spawns, and seen over a million environmental objects destroyed, including walls, windows, crates, and buildings your squad crashed the helicopter into.

Semidynamics, the only provider of fully customizable RISC-V processor IP, announces Cervell, a scalable and fully programmable Neural Processing Unit (NPU) built on RISC-V. Cervell combines CPU, vector, and tensor capabilities in a single, unified all-in-one architecture, unlocking zero-latency AI compute across applications from edge AI to datacenter-scale LLMs.

Delivering up to 256 TOPS (Tera Operations Per Second) at 2 GHz, Cervell scales from C8 to C64 configurations, allowing designers to tune performance to application needs—from 8 TOPS INT8 at 1 GHz in compact edge deployments to 256 TOPS INT4 in high-end AI inference.

Last September, AMD leadership publicly revealed UDNA—an "unforking" of previously separate enterprise and commercial GPU branches. Not long after this announcement, TechPowerUp's resident Serbian correspondent—AleksandarK—sat down with Team Red's Andrej Zdravkovic. The Chief Software Officer (and SVP) stated that a fair chunk of UDNA-related development work would be done by local engineers. Zdravkovic discussed this technology's eventual deployment in futuristic "AI PCs," but gamers have been salivating at the prospect of a proper successor to RDNA 4. A next-gen graphics architecture seeker—MrMPFR—has combed through official documents for any sign of UDNA preview material. The noted /Hardware subreddit member managed to distill their initial (very long) set of findings into an "easily digestible overview." They stated that this was just a small case of: "reporting and a little analysis on AMD's publicly available US patents filings," and other public-facing resources/archives.

Gleaned information included: "finalized architectural characteristics in future RDNA generations, AMD DXR IHV stacks (driver agnostic), and AMD sponsored titles. But please take everything with a grain of salt given my lack of professional expertise and experience with Real-time ray tracing (RTRT)". MrMPFR believes that Team Red started picking up former NVIDIA and Intel engineering talent, back in 2022/2023. In addition, a lot of new hires were apparently sourced from academic institutions. In theory, these newer team members have not had the time to make major inroads—in terms of getting finalized products out into the wild. MrMPFR reckons that noticeable contributions will accelerate AMD's making of "RDNA 6+/UDNA 2+," and beyond. Early 2025 leaks have pointed to the company collaborating with Sony; their "PlayStation 6" console is tipped to be powered by some fork of Team Red's "UDNA" graphics technology.

Hollow Knight: Silksong's long gestation period has frustrated many Team Cherry fans. The Adelaide, Australia-based independent development and publishing house has only released one title: Hollow Knight (2017). The much-anticipated sequel was officially delayed two years ago; a "planned" H1 2023 launch window was scuppered. Since then, Metroidvania genre fanatics have pored over the slightest of new details or external insights. Team Cherry has maintained a very low profile the ensuing years, but hopes were elevated in March. At the time, Microsoft's ID@Xbox initiative boasted about their empowering (aka supplementary funding) of development on Clair Obscur: Expedition 33, Descenders Next, FBC: Firebreak, and "of course Hollow Knight: Silksong." Early last month, Team Cherry's PR and marketing chief—Matthew Griffin—confirmed that the team was working on versions of Silksong for Nintendo Switch and Switch 2 consoles. Aditionally, two new in-game screenshots emerged via Nintendo of Japan's official website.

Fresh signs of activity were welcomed and celebrated by the upcoming game's intended audience. More good news has turned up this week—eagle-eyed observers noticed that The Australian Centre for the Moving Image (ACMI) will be previewing a playable build later this year. As disclosed within a new exhibit description: "Game Worlds transports you inside the levels of 30 unforgettable video games, from global blockbusters like World of Warcraft (Blizzard Entertainment) to pioneer of the life simulation genre The Sims (Electronic Arts), Neopets (World of Neopia) and highly anticipated Australian title Hollow Knight: Silksong, plus more to be announced." The Melbourne-based museum could attract an unprecedented number of paying gamer tourists; playable titles will be showcased until February of next year.

We're excited to present IBM Granite 4.0 Tiny Preview, a preliminary version of the smallest model in the upcoming Granite 4.0 family of language models, to the open source community. Granite 4.0 Tiny Preview is extremely compact and compute efficient: at FP8 precision, several concurrent sessions performing long context (128K) tasks can be run on consumer grade hardware, including GPUs commonly available for under $350 USD. Though the model is only partially trained—it has only seen 2.5T of a planned 15T or more training tokens—it already offers performance rivaling that of IBM Granite 3.3 2B Instruct despite fewer active parameters and a roughly 72% reduction in memory requirements. We anticipate Granite 4.0 Tiny's performance to be on par with that of Granite 3.3 8B Instruct by the time it has completed training and post-training.

As its name suggests, Granite 4.0 Tiny will be among the smallest offerings in the Granite 4.0 model family. It will be officially released this summer as part of a model lineup that also includes Granite 4.0 Small and Granite 4.0 Medium. Granite 4.0 continues IBM's firm commitment to making efficiency and practicality the cornerstone of its enterprise LLM development. This preliminary version of Granite 4.0 Tiny is now available on Hugging Face—though we do not yet recommend the preview version for enterprise use—under a standard Apache 2.0 license. Our intent is to allow even GPU-poor developers to experiment and tinker with the model on consumer-grade GPUs. The model's novel architecture is pending support in Hugging Face transformers and vLLM, which we anticipate will be completed shortly for both projects. Official support to run this model locally through platform partners including Ollama and LMStudio is expected in time for the full model release later this summer.

While we initially set out to self-publish Battle Train, our goal is to make the best game possible. We've had many opportunities cross our paths, and one such opportunity felt too good to pass up. That said, we're proud to announce that we're now working with Bandai Namco Entertainment America Inc. to bring Battle Train to more folks in a couple of exciting ways. You've probably heard of Bandai Namco Entertainment. They've been around about as long as our entire industry and are known as the publishers behind everything from PAC-MAN to Elden Ring to Katamari Damacy (AKA, the silliest game with the best game soundtrack ever made).

Look, we know Battle Train is good. Our friends and family know Battle Train is good. But getting word to the general gaming community has been an uphill battle for us, so having Bandai Namco Entertainment on board to help increase our visibility is awesome… but while marketing is crucial, this news is about more than just marketing. With Bandai Namco Entertainment's support, we were able to extend development and add more to the game, polish the heck out of it, and really make this game sing. We're really pleased with the results. We've added extra cards, extra trains, extra trinkets, and extra time making sure Battle Train is an amazing gaming experience. ALSO, we're not just releasing on PC!!! We'll be bringing the ludicrous action of Battle Train to the Nintendo Switch as well! Surprise!!!

Hyper Light Breaker received some significant criticism immediately after launch, largely for clunky combat, insubstantial progression, and a lack of world-building. Some community feedback is to be expected, since Hyper Light Breaker is still in Steam Early Access, but the community backlash was unexpected, given the success that the game's predecessor, Hyper Light Drifter, garnered. The first major content update aimed at addressing much of the criticism has just dropped on Steam, and many of the changes seem to be welcome by the community.

Some of the biggest changes to Hyper Light Breaker introduced in the Buried Below update include changes to character progression, weapon synergies, and the overall gameplay loop itself. Heart Machine also introduced a new playable character, Rondo, two new melee weapons, a new mini boss—a Crown called Maw—with its own player perks, and new areas for players to explore. Heart Machine also says it has streamlined the game's onboarding process, and that the changes to the game should make the game easier to get into for new players but more difficult to master. In terms of the gameplay loop, the new update reduces player lives from four per run to one per run, making things much less forgiving, however, players will also now always spawn in one of two locations, making the start of a new run more predictable. There are also a number of improvements to the enemy AI that should make encounters and battles somewhat more challenging and interesting. Players are also locked to a single character and companion for the duration of each run, which is a substantial change from the previous system which allowed players to switch characters and Sycoms mid-run. There are a number of other balance patches and new additions, which you can read about in the full patch notes below. Overall, judging from the comments in the Steam Community discussion under the update patch notes, players seem to be tentatively excited about the majority of the changes to the gameplay loop, although some concerns remain.

AI-powered image generation has progressed at a remarkable pace—from early examples of models creating images of humans with too many fingers to now producing strikingly photorealistic visuals. Even with such leaps, one challenge remains: achieving creative control. Creating scenes using text has gotten easier, no longer requiring complex descriptions—and models have improved alignment to prompts. But describing finer details like composition, camera angles and object placement with text alone is hard, and making adjustments is even more complex.

Advanced workflows using ControlNets—tools that enhance image generation by providing greater control over the output—offer solutions, but their setup complexity limits broader accessibility. To help overcome these challenges and fast-track access to advanced AI capabilities, NVIDIA at the CES trade show earlier this year announced the NVIDIA AI Blueprint for 3D-guided generative AI for RTX PCs. This sample workflow includes everything needed to start generating images with full composition control. Users can download the new Blueprint today.

Siemens Digital Industries Software today announced that its continued collaboration with Intel Foundry has resulted in multiple product certifications, updated foundry reference flows, and additional technology enablement leveraging the foundry's leading-edge technologies for next-generation integrated circuits (IC) and advanced packaging. Siemens is a founding partner of the Intel Foundry Accelerator Chiplet Alliance - enabling a new and compelling solution for 3D IC and chiplet offerings to a breadth of semiconductor market verticals.

Intel 18A Certification Achievements
Siemens' industry-leading Calibre nmPlatform tool is now certified for the latest Intel 18A production Process Design Kit (PDK). Intel 18A represents a significant technological leap forward, featuring innovative RibbonFET Gate-all-around transistors and the industry's first PowerVia backside power delivery. This Calibre certification allows mutual customers to continue leveraging the Calibre nmPlatform tool as their industry-standard sign-off solution with Intel Foundry's most advanced manufacturing process, accelerating time-to-market for next-generation chip designs.

In the latest blow to the gaming industry, Respawn Entertainment, a subsidiary of EA, has officially announced the cancellation of two new games it had in the early stages of development. The official announcement does not specify what the cancelled games were, but previous reports from early this month indicated that Respawn had quietly canned a game in the Titanfall IP. It seems somewhat like this announcement from Respawn is simply confirmation of the cancellation that was previously already leaked. Along with the cancellation of the game, and likely more importantly, Bloomberg claims that the "targeted team adjustments" Respawn mentioned in the release would mean that as many as 100 employees on the Respawn team will be laid off.

As for the rest of the restructuring, Respawn says that it will be focusing its efforts on Apex Legends and Star Wars Jedi. Previous rumors claimed that Respawn was developing a mainline Titanfall 3 game, which will supposedly launch at the 2025 The Game Awards show, which is scheduled for December 11, 2025. It seems as though the Titanfall IP game that was cancelled as of this announcement was something of a spin-off extraction shooter set in the Titanfall universe. This news comes at a time when gaming industry lay-offs seem to be at a peek and shortly after the creation of the United Videogame Workers Union—a union aimed at protecting North American video game workers from exactly these sorts of unexpected layoffs that are seemingly aimed at reducing the workforce in the name of short-term profits.

Ansys today announced thermal and multiphysics signoff tool certifications for designs manufactured with Intel 18A process technology. These certifications help ensure functionality and reliability of advanced semiconductor systems for the most demanding applications—including AI chips, graphic processing units (GPUs), and high-performance computing (HPC) products. Intel Foundry and Ansys have also enabled a comprehensive multiphysics signoff analysis flow for Intel Foundry's EMIB technology used for creating multi-die 3D integrated circuit (3D-IC) systems.

Recognized as industry-leading solutions, RedHawk-SC and Totem deliver speed, accuracy, and capacity to analyze the power integrity and reliability of Intel 18A RibbonFET Gate-all-around (GAA) transistors with PowerVia backside power delivery. For scalable electromagnetic analysis, Ansys is introducing HFSS-IC Pro, a new addition to the HFSS-IC product family. HFSS-IC Pro is certified for modeling on-chip electromagnetic integrity in radio frequency chips, Wi-Fi, 5G/6G, and other telecommunication applications made with the Intel 18A process node.

Today at Intel Foundry Direct Connect, the company will share progress on multiple generations of its core process and advanced packaging technologies. The company will also announce new ecosystem programs and partnerships, and welcome industry leaders to discuss how a systems foundry approach enables collaboration with partners and unlocks innovation for customers.

Intel CEO Lip-Bu Tan will open the event by discussing Intel Foundry's progress and priorities as the company drives the next phase of its foundry strategy. Naga Chandrasekaran, Intel Foundry chief technology and operations officer, and Kevin O'Buckley, general manager of Foundry Services, will also deliver keynotes during the morning session, sharing process and advanced packaging news while highlighting Intel Foundry's globally diverse manufacturing and supply chain.

At today's Intel Foundry Direct Connect 2025 event, Synopsys, Inc. announced broad EDA and IP collaborations with Intel Foundry, including availability of its certified AI-driven digital and analog design flows for the Intel 18A process node and production-ready EDA flows for the Intel 18A-P process node with RibbonFET Gate-all-around transistor architecture and the industry's first commercial foundry implementation of PowerVia backside power delivery. To drive multi-die design innovation forward, Synopsys and Intel Foundry are collaborating to enable Intel's new Embedded Multi-die Interconnect Bridge-T (EMIB-T) advanced packaging technology with an EDA reference flow powered by Synopsys 3DIC Compiler. With its EDA flows, multi-die solution, and broad portfolio of Synopsys' foundation and interface IP on Intel 18A and Intel 18A-P, Synopsys is helping designers accelerate the development of highly optimized AI and HPC chip designs from silicon to systems.

In a keynote presentation at today's event, John Koeter, Senior Vice President, for the Synopsys IP Group, emphasized: "The successful collaboration between Synopsys and Intel Foundry is advancing the semiconductor industry with silicon to system design solutions to meet the evolving needs for AI and high-performance computing applications. Our production-ready EDA flows, IP, and multi-die solution, provides our mutual customers with comprehensive technologies to accelerate the development of chip designs that meet or exceed their requirements."

MiTAC Computing Technology, a global leader in high-performance and energy-efficient server solutions, is proud to announce its participation at the OCP EMEA Summit 2025, taking place April 29-30 at the Convention Centre Dublin. At Booth No. B13, MiTAC will showcase its latest innovations in server design, sustainable cooling, and open-source firmware development - empowering the future of data center infrastructure.

C2810Z5 & C2820Z5: Advancing Sustainable Thermal Design
MiTAC will debut two new OCP server platforms - C2810Z5 (air-cooled) and C2820Z5 (liquid-cooled), built to meet the demands of high-performance computing (HPC) and AI workloads. Designed around the latest AMD EPYC 9005 series processors, these multi-node servers support the latest EPYC 9005 processors and are engineered to deliver optimal compute density and power efficiency.

OKI Circuit Technology, the OKI Group printed circuit board (PCB) company, has successfully developed 124-layer PCB technology for wafer inspection equipment designed for next-generation high bandwidth memory, such as HBM mounted on AI semiconductors. This is a roughly 15% increase in the number of layers over conventional 108-layer designs. OTC is seeking to establish mass production technology by October 2025 at its Joetsu Plant in Joetsu City, Niigata Prefecture, which has a proven track record and advanced development and production capabilities in the field of high multilayer, high-precision, large-format PCBs for semiconductor inspection equipment.

AI processing requires the transmission of vast data volumes between graphics processing unit (GPU) semiconductors and memory. As semiconductor performance increases, the memory installed is also required to have high-speed, high-frequency, and high-density data transfer capabilities. HBM features a stacked DRAM structure, requiring technology capable of fabricating wafers even more thinly and precisely. This configuration also requires that the PCBs used in inspection equipment meet even higher levels of performance and quality.

M31 Technology Corporation (M31), a global provider of silicon intellectual property (IP), today announced that its eUSB2 PHY IP has achieved silicon-proven status on TSMC's 3 nm process and has successfully completed tape-out on TSMC's 2 nm process. As a member of TSMC Open Innovation Platform (OIP) IP Alliance since 2012, M31 has been honored with TSMC OIP Partner of the Year Award for seven consecutive years. In 2020, M31 pioneered its eUSB2 IP solution on the TSMC 7 nm process node, further solidifying its leadership in advanced interface IP development. Since then, M31 has steadily expanded its eUSB2 IP portfolio across TSMC's 5 nm, 3 nm, and most recently, 2 nm process technologies—closely aligning with TSMC's cutting-edge roadmap to accelerate the adoption of AI-enabled smart devices. Looking ahead, M31 is actively developing the next-generation eUSB2 Version 2.0 (eUSB2V2) PHY IP, with ongoing efforts focused on both TSMC's 3 nm and 2 nm process nodes.

Reinforcing its robustness and reliability on advanced nodes, M31's eUSB2 IP solutions have been widely adopted by leading global companies in high-end smartphone chipsets and AI-driven image processing applications. Building on this success, M31 is currently advancing the development of eUSB2V2 IP on TSMC's N3 and N2 process technologies - expanding its comprehensive eUSB2 portfolio to include eUSB2 V1, V2 PHY, and eUSB2 Repeater solutions. Leveraging the eUSB2 standard, eUSB2V2 enhances data transfer rates while maintaining a low-voltage interface and leveraging asymmetric bandwidth technology—allowing TX and RX to operate at different data rates. This significantly improves transmission efficiency, making it ideal for embedded applications such as AI edge computing, smart surveillance, and image processing chips. To accommodate diverse design needs, eUSB2V2 leverages and enhances the I/O architecture based on the eUSB2 standard, supporting data transfer speeds from 480 Mbps up to 4.8 Gbps. The solution delivers a comprehensive eUSB2 platform for high-end SoCs—optimizing power efficiency, performance and design flexibility, while maintaining full compatibility with legacy USB 2.0 devices.