Today's Hardware Pulse
Nvidia driver hints at next‑gen DLSS 5 work
Nvidia’s GeForce Game Ready driver 610.47 adds three new profile options—Enable DLSS NR override, Enable DLSS NR SL override, and Override DLSS NR presets—suggesting continued development of DLSS 5 neural‑rendering technology. The earlier DLSS 5 reveal had sparked debate over AI‑driven image alteration, and the new driver hints at a quiet but ongoing effort.
Also developing:
By the numbers: Cyient Semiconductors raises $30M Series A
Intel’s Crescent Island PCB Leak Shows Xe3P GPU with 160 GB LPDDR5X Memory
A PCB leak posted by leaker YuuKi_AnS confirms Intel's upcoming Crescent Island accelerator will use the Xe3P architecture and 160 GB of LPDDR5X memory. The design, aimed at air‑cooled data‑center inference, could give Intel a pricing edge over Nvidia and AMD's HBM‑based solutions.
Surface Pro 13‑inch: Fast, Long‑Lasting, Yet Fan‑Heavy
I've been working on Microsoft's 13-inch Surface Pro 12th Gen for Business running Intel Core Ultra Series 3, and am surprised by its speed and battery life, but oh that fan https://www.techradar.com/computing/windows-laptops/microsoft-13-inch-surface-pro-12th-gen-intel-core-ultra-series-3-hands-on
Super Semiconductor ETF Puts 40% Into Micron, AMD, Broadcom, Nvidia and Intel
The iShares Semiconductor ETF, newly highlighted for its concentrated exposure, assigns 40% of its portfolio to five AI‑focused chipmakers—Micron, AMD, Broadcom, Nvidia and Intel. The fund’s 30‑stock composition has delivered a 330% gain since early 2023, underscoring the rapid upside...
Imec Unveils First Quantum‑Dot Qubit Fabricated with High‑NA EUV Lithography
At ITF World, imec demonstrated a silicon quantum‑dot qubit device fabricated with high‑NA EUV lithography, achieving 6‑nm gate gaps and proving the method’s suitability for fab‑scale quantum hardware. The breakthrough links semiconductor manufacturing to quantum‑computing scaling.
Alibaba Launches Zhenwu M890 AI Accelerator, Tripling Performance of Its Own Chip
Alibaba Group unveiled the Zhenwu M890 AI accelerator at its Cloud Summit, promising three times the performance of the earlier Zhenwu 810E and 144 GB of GPU memory. The chip is aimed at agent‑centric workloads and marks a fresh escalation in...
Smart Glasses: Samsung and Google Take the Next Step in the AI Race
Samsung and Google have joined forces with fashion brands Gentle Monster and Warby Parker to launch a new generation of AI‑powered smart glasses, entering a market recently energized by EssilorLuxottica and Meta’s second‑generation release. The collaboration blends cutting‑edge vision technology...
Vishay’s New Automotive Optocouplers Combine 5300 VRMS Isolation and 11 Mm Creepage for EV Electronics Isolation Stages
Vishay Intertechnology has introduced two AEC‑Q102‑qualified optocouplers, the VOWA617A and VOWA618A, targeting high‑voltage isolation in electric‑vehicle power electronics. The devices deliver 5300 V RMS isolation, ≥11 mm creepage distance and a CTI of 600, exceeding typical automotive standards. They also offer a broader...
Smartphone Lidar Detects Hidden Objects Around Corners
Researchers just showed that cheap, smartphone-grade lidar can reveal hidden shapes around a corner. https://spectrum.ieee.org/smartphone-grade-lidar?share_id=9514826
Semidynamics Secures a Strategic Investment to Advance Memory-Centric AI Inference Chips
Semidynamics announced a strategic investment to accelerate its memory‑centric AI inference chips, aiming to overcome the industry’s “memory wall.” The funding will expand engineering teams, speed product development, and deepen ecosystem partnerships. By re‑architecting silicon around memory efficiency, the startup...
Amphenol Printed Circuits Shows PCBs of All Sizes at IMS2026
Amphenol Printed Circuits (APC) will exhibit a full spectrum of printed‑circuit‑board designs at the 2026 IEEE MTT‑S International Microwave Symposium in Boston, June 9‑11. The booth will feature high‑performance PCBs for rugged RF/microwave, high‑speed digital, and space‑qualified applications, ranging from tiny...
Alibaba Unveils New AI Chip as Nvidia Access Remains Stalled
Alibaba unveiled its Zhenwu M890 AI accelerator, claiming three‑times the performance of the earlier Zhenwu 810E and positioning it as a domestic alternative while Nvidia’s H200 chip remains stalled in China due to U.S. export restrictions. The new chip joins a Zhenwu...
India Just Signed Its Most Consequential Chip Deal. The Hard Part Starts Now.
India and the Netherlands sealed a landmark MoU between Tata Electronics and ASML, committing the Dutch firm to supply EUV lithography tools and support for a new semiconductor fab in Dholera, Gujarat. The plant, built with a 50 % government stake...
Nvidia Ships First Vera CPUs to Anthropic, OpenAI, SpaceXAI and Oracle Cloud
Nvidia has delivered its first Vera CPU systems to Anthropic, OpenAI, SpaceXAI and Oracle Cloud Infrastructure, launching a custom processor designed for agentic AI. The rollout signals Nvidia’s push beyond GPUs into a new CPU class that could reshape AI...
Roundhill Memory ETF (DRAM) Nears 100% Gain in Six Weeks, Tops $9 B AUM
The Roundhill Memory ETF (ticker DRAM) has surged nearly 100% since its April 2 launch, climbing to more than $9 billion in assets under management. The fund’s rapid rise reflects soaring demand for DRAM and NAND chips that power AI data...
Hesai Ships 471,000 Lidar Units in Q1 2026, Revenue Jumps 30% to $99M
Hesai Group announced it shipped over 471,000 lidar units in the first quarter of 2026, driving a 30% year‑over‑year revenue increase to $99 million. The Chinese sensor maker also secured a strategic supply deal with Mercedes‑Benz for Level 3 autonomous models and...
Nvidia CEO Jensen Huang Signals Surprise China Market Opening Amid Rising Chip‑trade Tensions
Nvidia founder and CEO Jensen Huang made a surprise appearance in Beijing, indicating an unexpected opening to the Chinese market despite escalating chip‑trade disputes. The move, timed with a high‑profile U.S.-China summit, could reshape Nvidia's growth outlook and the broader...
UHF RFID Sessions: The Hidden Flag That Makes or Breaks Your Read Performance
UHF RFID tags contain four single‑bit session flags (S0‑S3) that dictate whether a tag responds during an inventory round. The flags toggle between states A and B, allowing a reader to mark tags as "already read" and avoid duplicate reads....
Building What Comes Next: Why Startups Need a Manufacturing-First Approach to Innovation
GlobalFoundries (GF) argues that semiconductor startups must embed manufacturability into early design decisions to avoid costly redesigns and capital inefficiency. The firm’s GF Labs and GF Accelerate programs provide startups with process design kits, multi‑project wafer runs, and venture funding...
Dell Tech World 2026: Eli Lilly and Company Scales AI-Driven Drug Discovery and Manufacturing with Dell
Dell Technologies deepens its 15‑year partnership with Eli Lilly, delivering compute and storage that power AI‑driven drug discovery and global manufacturing. Dell’s PowerEdge servers and PowerStore systems provide the backbone for LillyPod, a supercomputer feeding more than 1,000 GPUs with nearly...
Alibaba Bolsters AI Play with Fresh Chip, LLM
Alibaba unveiled its most powerful AI processor, the Zhenwu M890, featuring 144 GB of on‑chip memory and roughly three‑fold performance gains over its predecessor. The company also launched the Qwen3.7‑Max large language model, optimized for coding and autonomous, multi‑step tasks, and...
Hoth Plans Rocket One Rebrand and AI Chip Pivot
Hoth Therapeutics announced it will rename itself Rocket One, Inc. and restructure to focus on artificial‑intelligence semiconductor and spintronic computing technologies, while keeping its biotech programs in a separate subsidiary. The company secured exclusive licenses from Virginia Commonwealth University for...
Blackstone and Google Commit $5 B to Launch U.S. AI‑Chip Company Focused on TPUs
Blackstone announced a $5 billion equity commitment to a joint venture with Google that will create a U.S. company offering compute‑as‑a‑service using Google’s Tensor Processing Units. The venture aims to bring 500 MW of data‑center capacity online in 2027, positioning private‑equity at...
Flexible Organic-Inorganic Hybrid Synapse Advances Physical Reservoir Computing
Researchers unveiled a flexible organic‑inorganic hybrid charge‑trap synapse that delivers non‑volatile, analog weight storage with fast response and high endurance, enabling low‑power physical reservoir computing. The device combines organic semiconductor flexibility with inorganic charge‑trap layers to emulate synaptic dynamics while...
Dell Launches PowerStore Elite and 18th‑Gen PowerEdge, Triples Storage Performance for AI
Dell Technologies announced a sweeping refresh of its data‑center portfolio at Dell Technologies World, debuting PowerStore Elite storage and 18th‑generation PowerEdge servers. The new hardware promises up to three‑fold performance gains, a 6:1 data‑reduction guarantee and up to 70% higher...
Samsung and Google Launch AI-Powered Smart Glasses with Gentle Monster and Warby Parker
Samsung Electronics and Google unveiled a new line of AI‑enabled smart glasses at Google I/O 2026, built on Google’s Gemini AI and Samsung hardware. The devices, co‑designed with fashion houses Gentle Monster and Warby Parker, aim to create a hands‑free...
Microsoft Set To Bring Biggest India Data Centre Online
Microsoft will launch its biggest data centre in Hyderabad by mid‑2024, part of a $20.5 billion investment in India. The facility aims to meet soaring demand for Azure cloud and Copilot AI services, which Indian firms such as Infosys, Cognizant and...
Alibaba Is Designing AI Chips Around Agents, and that Changes What the Race Is Actually About
Alibaba’s semiconductor arm T‑Head unveiled the Zhenwu M890, an AI accelerator built specifically for agent‑style workloads. The chip delivers roughly three times the performance of its 810E predecessor and is the first in a three‑year roadmap that includes the V900...
Schunk Supplies 123 Depot Charging Pantographs to King County Metro in Seattle
Schunk Transit Solutions is delivering 123 inverted SLS 201 pantographs to King County Metro’s Tukwila Base depot for electric‑bus charging. The equipment integrates into the depot structure rather than mounting on vehicle roofs, and full commissioning is planned for spring 2026. Metro...
MTN to Turn Its African Towers Into an AI Inference Grid
MTN Group announced it will retrofit its African cellular towers with open‑GPU hardware, creating a distributed AI inference grid that runs both radio functions and edge AI workloads. The move complements two new AI‑enabled data centres slated for South Africa...
China Reportedly Blocks NVIDIA GeForce RTX 5090 D V2 Imports
Chinese customs have reportedly denied import permits for NVIDIA's GeForce RTX 5090 D v2, a GPU specifically engineered for the Chinese market. The block appears to originate from Chinese regulators rather than new U.S. export controls, catching NVIDIA off guard. Because the card...
KIOXIA EXCERIA PRO G2 (PCIe Gen 5.0 NVMe SSD @ 14,9 GB/S) Review
Japanese memory maker KIOXIA has launched the EXCERIA PRO G2, a consumer‑grade PCIe 5.0 ×4 NVMe SSD delivering up to 14.9 GB/s sequential reads and 13.7 GB/s writes. The drive pairs its high bandwidth with integrated DRAM cache, a 110 % improvement in power efficiency, and a...
How Data Movement Defines Performance for AI Silicon
AI silicon performance now hinges on efficient data movement rather than raw compute power. In data‑center GPUs, over 80% of dynamic energy is spent shuttling data to DRAM, while edge AI devices can waste up to 90% of inference time...
Intel Nova Lake ES Chips Reportedly Double Multi-Core Performance
Intel has begun shipping engineering samples of its next‑generation Core Ultra 400 (code‑named Nova Lake) desktop processors to hardware partners for validation. The chips introduce a redesigned hybrid architecture with Coyote Cove performance cores and Arctic Wolf efficiency cores, and feature...
Swissbit Launches N7000 PCIe 4.0 SSD Series With Up to 10 DWPD
Swissbit unveiled the N7000 series, a new line of PCIe 4.0 NVMe SSDs built on its eighth‑generation BiCS8 NAND. The portfolio includes DRAM‑less models that rely on Host Memory Buffer technology and offers two endurance tiers: a TLC‑based 0.3 DWPD variant and...
Acer Launches Aspire 14 AI Ryzen AI 400 Copilot+ Notebooks
Acer unveiled the Aspire 14 AI notebook series, featuring two 14‑inch models powered by AMD’s Ryzen AI 400 processors. The laptops combine 16 GB LPDDR5X memory, Radeon 840M graphics and a 50‑TOPS NPU for on‑device AI tasks such as text generation and image editing. A dedicated...
COLORFUL Launches iGame B850M ULTRA Series AM5 Motherboards
COLORFUL has unveiled the iGame B850M ULTRA Series, two micro‑ATX motherboards for AMD Ryzen 9000 AM5 platforms. The ULTRA‑S model features a 14+2+1‑phase VRM and supports DDR5‑8400 memory, while the ULTRA‑OC pushes DDR5‑8800 and adds 5 Gb Ethernet for extreme overclocking. Both boards include...
Intel Crescent Island Xe3P AI Accelerator PCB Leaks Online
Intel has leaked PCB images of its upcoming Crescent Island AI accelerator, built on the Xe3P "Celestial" architecture. The design departs from previous GPU layouts, featuring a central square processor and 20 LPDDR5x memory modules that can deliver up to...
Computex 2026: What to Wait For Before You Buy a Laptop
Computex 2026, themed “AI Together,” kicks off June 2‑5 in Taipei, showcasing a clash of ARM laptops, NVIDIA’s AI‑centric roadmap, and heavyweight keynotes from Jensen Huang, Qualcomm’s Cristiano Amon and Intel’s Lip‑Bu Tan. The show promises to reveal NVIDIA’s long‑rumored N1X Arm platform, Intel’s...
Sony Xperia 1 VIII: Europe Leak Confirms Qualcomm’s New Platform in Sony’s Flagship
Sony has officially confirmed the Xperia 1 VIII flagship for Europe, equipping it with the Snapdragon 8 Elite Gen 5 SoC, 12 GB of RAM in the base model and up to 16 GB in a higher‑end version, and storage options ranging from 256 GB to 1 TB. The device...
NVIDIA Vera: First In-House CPU Effort Lands with Leading AI Labs
NVIDIA has begun shipping its in‑house Vera CPU to four marquee AI customers—Anthropic, OpenAI, Oracle Cloud Infrastructure, and SpaceXAI. The processor packs 88 custom Olympus cores, up to 1.2 TB/s memory bandwidth and 3.4 TB/s fabric bandwidth, targeting the latency‑critical orchestration layer...
Global OLED Monitor Shipments Soar 78% YoY for 1Q 2026
Global OLED monitor shipments surged 78% year‑on‑year in Q1 2026 despite an 11% quarter‑on‑quarter dip, driven by abundant QD‑OLED panel supply. ASUS retained the top market share at 24% thanks to a broad OLED lineup, including a new 34‑inch 360 Hz...
ST Image Sensors Bring Always-On Vision to Next-Gen Personal Electronics
STMicroelectronics has introduced the VD55G4 monochrome and VD65G4 RGB image sensors, the latest members of its BrightSense family. These ultralow‑power global‑shutter chips deliver always‑on vision while consuming up to ten times less power than conventional sensors. Designed for wearables, AR/VR...
Rethinking System Design Amid the DRAM Crunch
DRAM scarcity is hitting AI hardware, with high‑capacity modules now costing three to four times previous levels and facing long lead times. This pressure is prompting a redesign of AI systems toward lower memory footprints, leveraging edge accelerators that run...
GatesAir Supports Successful ATSC 3.0 Direct-to-Mobile Trial in India
GatesAir announced the successful completion of a three‑year ATSC 3.0 Direct‑to‑Mobile (D2M) field trial in Delhi, India, partnered with public broadcaster Prasar Bharati and Tejas Networks. The trial used GatesAir’s Maxiva XTE exciter and a hybrid high‑power/high‑tower and low‑power/low‑tower transmission architecture operating in...
TSMC Lays Out ‘Three-Layer Cake’ Vision for AI Chips
Taiwan Semiconductor Manufacturing Company unveiled a "three‑layer cake" roadmap that fuses its 2nm N2 compute node, advanced 3D CoWoS packaging, and the Compact Universal Photonic Engine (COUPE) for high‑speed optical links. The strategy was announced at TSMC’s 2026 Technology Symposium...
Cusp-Singularity-Enhanced Coriolis Effect for Sensitive Chip-Scale Gyroscopes
Researchers have introduced third‑order cusp singularities into a chip‑scale Coriolis vibratory gyroscope, creating a singularity‑enhanced Coriolis effect. The approach yields a cubic‑root response that lifts the effective Coriolis factor by up to 1,010 times, improves signal‑to‑noise ratio 253‑fold, and boosts precision...
Hybrid AI Chip Achieves 100× Compute‑per‑Watt via Analog Flash
On the silicon side, it's the era of analog inference. This company combo will deliver the first hybrid AI platform with 100x more compute per watt than the best GPUs. How? Mythic performs the core AI operation (8-bit multiply & add)...
Amazon’s Trainium AI Chips Gain Traction with Developers as Software Matures and Nvidia GPUs Face Capacity Constraints
Amazon’s custom Trainium AI chips are gaining momentum as Anthropic and OpenAI secure large capacity commitments and developers migrate workloads thanks to maturing software. The AWS custom silicon portfolio now exceeds a $20 B annualized run rate, equivalent to roughly $50 B...
How Micron (MU) Is Targeting AI Data Centers With Higher-Capacity Server Memory
Micron Technology announced on May 12 that it has sampled a 256GB DDR5 registered DIMM built on its 1‑gamma DRAM process, delivering speeds up to 9,200 MT/s—about 40% faster than existing volume‑produced modules. The memory uses advanced 3D stacking and...
How AMD (AMD) Is Bringing AI Acceleration Into Existing Data Center Racks
Advanced Micro Devices unveiled its Instinct MI350P PCIe GPU on May 7, a dual‑slot accelerator designed to run AI inference in existing air‑cooled data‑center racks without major power or cooling upgrades. The cards target generative and agentic AI workloads, allowing enterprises...