Today's Hardware Pulse
Nvidia and SK Hynix ink multi-year AI memory chip pact
Nvidia and SK Hynix have signed a multi‑year agreement to co‑develop next‑generation memory chips optimized for artificial‑intelligence workloads. The collaboration will focus on memory for Nvidia’s Vera Rubin accelerator and broader AI infrastructure, pairing Nvidia’s AI expertise with SK Hynix’s high‑performance memory capabilities.
Also developing:
By the numbers: Rapidus raises $943M in funding
Thundercomm TurboX C7790 Android and Linux Development Kit Features Qualcomm Dragonwing Q-7790 Edge AI SoC
Thundercomm unveiled the TurboX C7790 development kit, an edge‑AI platform built around Qualcomm’s Dragonwing Q‑7790 SoC. The kit combines a 4 nm octa‑core processor, 12 GB LPDDR5X memory and 128 GB flash storage, delivering up to 24 TOPS of AI inference. It ships with a carrier board that supports Android and Linux, dual 1 GbE, HDMI, USB‑C, MIPI DSI, and a Sony IMX577 camera module. Priced at $1,499, the solution targets AI cameras, robotics, video‑conferencing and edge‑gateway deployments.
Google Home Speaker 2026: Gemini Finally Gets a New Home
Google unveiled the new Google Home Speaker, its first standalone smart speaker since the 2020 Nest Audio, priced at $135 and slated for a June 25, 2026 release. The device is built around Google’s Gemini large‑language‑model AI, delivering faster, conversational interactions via...
Analysts Lift Broadcom Targets to $640 as Stock Falls 14% on AI Outlook
Following Broadcom's Q2 earnings, a slate of Wall Street houses raised their price targets, with BNP Paribas moving its goal to $640 and Jefferies to $550. The upgrades come as the shares tumbled about 14% after the company fell short of...
SuperX Unveils 1.6 Tbps Optical Modules at Interop Tokyo, Targeting AI Data Center Supply Chain
SuperX AI Technology Limited announced the launch of its 1.6 Tbps optical module solution at Interop Tokyo 2026, alongside a full‑stack AI compute and power portfolio. The rollout aims to accelerate AI data‑center deployments in Japan and globally, tightening the telecom‑compute...
HP and Ferrari Unveil $5,599 AI‑Powered Limited‑Edition Laptop
HP and Ferrari announced a limited‑edition Scuderia Ferrari AI PC priced at $5,599, with only 4,999 numbered units. The laptop blends Ferrari’s design language with HP’s AI‑ready hardware, targeting collectors and enterprise users seeking premium performance.
Valve Confirms Summer 2026 Launch of Steam Machine and Steam Frame
Valve confirmed that its long‑delayed Steam Machine living‑room PC and Steam Frame VR headset will ship in summer 2026. The company rolled out a Verified program to assure developers of out‑of‑the‑box performance, even as global memory shortages keep pricing under...
SK Group Reconsiders SK Siltron Sale as AI Chip Supply Chain Value Rises
SK Group is reevaluating the planned sale of silicon‑wafer maker SK Siltron as the AI‑driven semiconductor upcycle lifts the strategic value of upstream wafer supply. Doosan Semiconductor was selected as the preferred bidder for a 70.6% stake, but negotiations have...
Check Out This Gachapon-Inspired Custom PC with a Floating Heart on Top, and Mini RAM Sticks in Its Capsules
A gachapon‑inspired custom PC debuted at Computex, built on a Thermaltake Tower 600 case and featuring a floating pink heart, spinning crank, and capsule‑styled RAM sticks. The machine houses an RTX 5060 GPU, a Ryzen 5 9600X CPU, and a custom water‑cooling loop, while...
University of Tokyo Creates 1‑nm Semiconducting Nanotubes, Paving Way for Ultra‑small Transistors
Scientists at the University of Tokyo have synthesized single‑walled molybdenum disulfide nanotubes only 1 nanometer in diameter, encased in boron‑nitride tubes. The achievement validates long‑standing theory on band‑gap scaling and opens a realistic path to gate‑all‑around transistors far smaller than today’s...
Why Apple Might Put Cameras Into Its Next AirPods
Apple is reportedly testing AirPods that embed low‑resolution cameras to give Siri visual awareness. The hardware is said to be ready, but Apple executives fear Siri’s vision AI isn’t mature enough and that privacy risks could spark backlash. Use cases...
Samsung Could Return Snapdragon Chips to Galaxy Z Flip 8 in 'Some' Areas
Rumors on Naver suggest Samsung’s upcoming Galaxy Z Flip 8 could adopt a dual‑chip strategy, pairing Qualcomm’s Snapdragon 8 Elite Gen 5 with its in‑house Exynos 2600 in different regions. The tipster Lanzuk claims cost pressures on the 2 nm Exynos chip have led Samsung to accept...
NVIDIA's New Power Play: Why Fluence Is Surging
AI's rapid scaling is hitting power‑grid limits, creating demand for on‑site energy storage. Siemens unveiled a 136‑MW reference design for NVIDIA’s DSX Vera Rubin supercomputer that specifies Fluence Energy’s SmartStack battery system. Fluence posted Q2 FY2026 revenue of $464.9 M, below...
Nio Partners with Chinese Chipmaker GigaDevice to Co-Develop Automotive Chips
Nio has signed a strategic cooperation agreement with Chinese semiconductor supplier GigaDevice to jointly develop automotive‑grade chips for its next‑generation smart electric vehicles. GigaDevice will supply high‑performance NOR flash memory and microcontroller units, while Nio will guide chip architecture, requirement...
Why Waymo Settled for the Wrong Car
Waymo’s autonomous ride‑hailing fleet still relies on retrofitted Jaguar I‑PACE EVs, a platform that costs $120,000‑$200,000 per car, suffers reliability issues, and cannot be resold. The company is rolling out a purpose‑built Zeekr‑made Ojai minivan with a streamlined 23‑sensor suite,...
Dish TV Launches VZY Smart TVs in Kerala with DTH & OTT Integration
Dish TV India has introduced its VZY Smart TV lineup in Kerala, headlined by a 65‑inch model that merges direct‑to‑home (DTH) services with over‑the‑top (OTT) streaming. The integrated platform delivers live channels, on‑demand libraries, and sports content through a single...
Intel’s Crescent Island: A Smarter Strategy, But Proof Still Pending
Intel unveiled the Crescent Island GPU, a PCIe‑form‑factor accelerator that packs up to 480 GB of LPDDR5x memory and runs at 350 W using air cooling. The chip is aimed at the growing inference market, where capacity, cost per token and deployment...
The Idea of a Supercomputer on Your Desk
Nvidia and Microsoft unveiled RTX Spark‑powered Windows PCs that embed a petaflop‑class AI superchip, positioning the desktop as an active digital teammate. The machines can run large generative models locally, targeting developers, researchers, and enterprise teams seeking lower inference costs and...
Linux 7.1 + Mesa 26.1 Performance With The Radeon RX 9070 GRE, RX 9070 XT
AMD’s Radeon RX 9070 GRE and XT were benchmarked on Ubuntu 26.04 (Linux 7.0, Mesa 26.0) and on a development stack (Linux 7.1‑Git, Mesa 26.1.1). The author found virtually no performance difference between the stock and the newer kernel/driver combo across gaming and compute tests. A slight regression appeared...
KC-46’s New Remote Vision System Completes Initial ‘Non-Contact’ Testing
Boeing announced that its upgraded Remote Vision System (RVS 2.0) for the KC‑46 tanker has completed initial non‑contact flight testing, confirming optical performance and hardware maturity. The system, first flown in November 2025, replaces a problematic camera boom that caused...
H3C Launches World’s First Wi‑Fi 8 Enterprise Access Point with US‑made Broadcom Chipset
H3C announced the debut of the first enterprise‑grade Wi‑Fi 8 access point, powered by Broadcom’s US‑made BCM4918 chip. The device targets factories, campuses and dense offices, promising up to 30% better spectral efficiency and a 25% cut in latency through AI‑driven...
U.S. Trade Groups Warn Chip Shortage Threat as AI Demand Surges
The Semiconductor Industry Association and the National Association of Manufacturers sent a joint letter to the Biden administration, warning that explosive AI demand could outstrip U.S. chip production capacity. The groups call for accelerated policy measures to expand semiconductor manufacturing...
Lakshmikumaran & Sridharan Deploys Dell AI Factory with NVIDIA to Modernize Legal Services
Lakshmikumaran & Sridharan Attorneys (LKS) has rolled out an on‑premise AI platform built on Dell AI Factory and NVIDIA accelerated computing, giving its lawyers instant access to proprietary legal knowledge. The move aims to replace labor‑intensive research with an AI...
REDMAGIC Unveils 11S Pro Gaming Phone with $1 Early‑Bird Voucher
REDMAGIC launched the 11S Pro gaming smartphone, pairing a $1 early‑bird voucher that grants $30 off with a suite of performance upgrades, a 24,000 RPM turbo fan and a 7,500 mAh battery. The pre‑order window runs June 3‑9, giving buyers 24‑hour early access...
AMD Ships Second-Gen Versal Prime Accelerators
AMD announced the shipment of three new Versal Prime Gen2 adaptive SoCs, expanding its portfolio for space‑constrained applications. The devices integrate a quad‑core Cortex‑A78AE, six‑core Cortex‑R52, a Mali‑G78AE GPU, and high‑speed DDR5/LPDDR5X memory, delivering up to five times the scalar...
AMD Radeon RX 9070 GRE Shows Strong Value and Overclocking Potential at $549
AMD has launched the Radeon RX 9070 GRE worldwide at a $549 suggested price, offering a default boost of 2.79 GHz and notable overclocking headroom. Reviewers say the card punches above its weight against Nvidia’s RTX 5060 Ti and rivals the RTX 5070 in performance‑per‑dollar.
AMD Non-Committal on FSR 4.1 Support for RDNA 3.5 Integrated Graphics
AMD has indicated that support for its FidelityFX Super Resolution 4.1 (FSR 4.1) is not currently planned on RDNA 3.5 integrated graphics, citing a need to weigh performance benefits against implementation costs. While desktop Radeon RX 7000 and RX 6000 series cards are expected...
Nikkei Hits Fresh All‑Time High as AI‑Driven Buying Fuels 2.5% Surge
The Nikkei 225 closed at a fresh all‑time high of 68,402.13, up 2.5%, as investors poured into AI‑linked technology shares. The rally was powered by strong chip and semiconductor equipment earnings, a weak yen and a broader Asian equity bounce...
NVIDIA’s AI PC Push May Open Opportunities for India’s Semicon Ecosystem
NVIDIA unveiled the RTX Spark, an AI‑first chip built with MediaTek that reimagines Windows laptops as personal AI agents. Major OEMs—including Lenovo, HP, Dell, Microsoft, Asus and MSI—will launch RTX Spark‑powered models, giving consumers a third alternative to Intel and AMD. Analysts...
Nvidia Unveils RTX Spark PC Chip at Computex, Threatening Qualcomm's Snapdragon Lead
Nvidia revealed its RTX Spark AI processor at Taiwan's Computex, a joint effort with Microsoft that blends a Blackwell GPU, an Arm‑based CPU and 128 GB of unified memory. The surprise launch sent Qualcomm’s stock lower, underscoring a new clash over...
Maxsun Brings Intel Panther Lake and Wildcat Lake Mobile CPUs to Desktop Motherboards
At Computex 2026, Maxsun unveiled its first “Mobile on Desktop” (MoDT) motherboards built around Intel’s upcoming Panther Lake and Wildcat Lake mobile processors. The flagship ATX SK‑PTLNAS integrates a Core Ultra 7 356H SoC, offering up to 128 GB DDR5, five M.2 slots, dual...
AMD Takes a Third of Server CPU Market as Shipments Grow
Processor shipments fell over 6% in Q1 2026, but server CPUs bucked the trend, growing more than 10% year‑over‑year. AMD captured a third of the server CPU market at 33.2%, up six points from a year earlier, while Intel’s server...
Reliance Global Group Positions EnQuantum to Secure the Post-Quantum Economy Before Q-Day
Classiq and Pontificia Universidad Católica de Chile have launched Latin America’s first Quantum Machine Learning Consortium for computational pathology, a 12‑month effort focused on renal disease analysis. The Entropy 2026 conference will convene researchers in Barcelona from July 1‑3 to explore quantum...
CoolIT Systems Demonstrates 15kW Coldplate, Extending Single-Phase DLC Roadmap Far Beyond 2030
CoolIT Systems unveiled a 15kW coldplate, delivering nearly four times the performance of its previous 4kW single‑phase direct liquid‑cooling (DLC) design. The new split‑flow microchannel architecture was validated with standard water‑glycol coolant at 1.2 L/min/kW, suitable for 45 °C warm‑water environments. This...
Silicon Control
Kerman Kohli’s June 4, 2026 essay “Silicon Control” argues that AI is not in a bubble and that the industry faces a decade‑long supply constraint rather than an overbuilt infrastructure. He outlines why compute capacity is effectively finite, how demand will continue...
Infineon’s New H-DPAK Puts a 750 V CoolSiC Half-Bridge in a Top-Side-Cooled Package for EV OBCs and DC-DC Converters
Infineon unveiled the H‑DPAK, a top‑side‑cooled package that embeds a 750 V CoolSiC G2 half‑bridge for automotive power‑conversion. The split‑lead‑frame design spreads heat efficiently and lowers parasitic loop inductance, curbing switching noise. At 2.3 mm tall, the package matches Infineon’s existing Q‑DPAK and...
100W Ultra-Wideband GaN Power Amplifier
Micable Inc. announced a new 100 W ultra‑wideband GaN power amplifier covering 0.5‑6 GHz continuously. The device delivers over 100 W saturated output, up to 64 dB gain, and a rugged 5:1 VSWR design with built‑in over‑temperature and over‑current protection. Priced from $16,800 and...
Primax Advances AI Conference Solutions with Qualcomm-Powered Edge AI Platforms
Primax Electronics announced an AI‑powered conference platform built around Qualcomm’s Dragonwing Q‑7790 Edge AI processor. The solution merges vision, audio, and interaction modules into a single AISF architecture, targeting hybrid‑work meeting rooms. By leveraging its ODM+ model and Tymphany’s professional audio...
SolidRun and Peridio Accelerate Development of Physical AI Deployments by Adding Avocado OS to RZ/V2N Vision AI Platforms
SolidRun and Peridio announced that Avocado OS, a production‑grade Linux for Physical AI, is now available on SolidRun’s RZ/V2N‑based HummingBoard RZ‑V2N‑AIoT and SolidSense AIoT V2N platforms. The RZ/V2N SoC combines a quad‑core Cortex‑A55, a 15 TOPS DRP‑AI3 neural accelerator, and rugged...
Pinterest Works with AWS to Power Next Chapter of AI-Driven Visual Search Discovery
Pinterest announced a $4 billion multi‑year agreement with Amazon Web Services, the largest cloud spend in its history, to power the next phase of AI‑driven visual search. The deal runs through 2031 and expands use of AWS custom silicon, including Trainium...
Novacium, LN Innov’ and Groupe Zekat to Showcase a Fully Integrated European Drone Propulsion Solution at Eurosatory 2026
HPQ Silicon’s French partner Novacium will unveil an Integrated Drone Propulsion System (IDPS) at Eurosatory 2026, together with LN Innov’ and Groupe Zekat. The IDPS combines Novacium’s silicon‑enhanced lithium‑ion batteries, LN Innov’ high‑performance electric motors, and Zekat’s intelligent electronic speed controllers, all manufactured in...
Valeo’s New Ineez Bidirectional EV Charging Station Implements ISO 15118-20, Enables V2G
Valeo unveiled the Ineez AC charging station at the Drive to Zero show in Paris, featuring bidirectional charging and the ISO 15118‑20 protocol. The new charger creates a secure, universal link between electric vehicles and the grid, enabling real‑time energy exchange...
TSMC CEO Warns AI‑Driven Chip Shortage Could Last Years, Threatening Manufacturers
TSMC chief executive C.C. Wei told shareholders that AI‑driven demand for advanced semiconductors will outpace the foundry’s supply for several years, even as the company expands capacity in the United States. The warning signals prolonged pressure on manufacturers that rely...
MSI Unveils First Intel Arc G3 Extreme Handheld, Challenging Steam Deck and ROG Ally
MSI announced the Claw 8 EX AI+ at Computex 2026, the world’s first handheld gaming PC built on Intel’s Arc G3 Extreme platform. Priced near $1,500, the device promises superior performance and AI‑driven upscaling against rivals like Valve’s Steam Deck...
Silicon Motion Increases Sales of SSD Controllers Amid NAND Shortage, but Expects NAND Shortages to Get Worse in 2027 —...
Silicon Motion reported Q1 2026 revenue of $342.1 million, up 23% quarter‑over‑quarter and 105% year‑over‑year, driven by a 40‑45% YoY rise in SSD controller sales. High‑end PCIe 4.0/5.0, UFS 3.1 and eMMC 4.1 controllers offset weakness in low‑end consumer chips. The company warns that NAND‑memory...
Intel: Pushing EMIB Forward: Design Methodology Insights with Synopsys Tools
Intel is extending its Embedded Multi‑die Interconnect Bridge (EMIB) platform by partnering with Synopsys to embed advanced EDA capabilities into the design flow. The collaboration delivers a co‑design methodology that lets engineers simulate signal, power and thermal behavior across multiple...
GCC Git Enables Additional Tuning For AMD Zen 6
The GNU Compiler Collection (GCC) has introduced additional tuning flags for AMD's upcoming Zen 6 (znver6) architecture in its development branch. The new optimizations fuse ALU operations with subsequent conditional jumps when the ALU uses memory operands or immediate‑plus‑displacement operands, and...
How to Turn the Odin 3 Into an ARM-Based Steam Deck Alternative
The Odin 3 handheld, powered by the Snapdragon 8 Elite and an Adreno 830 GPU, can run Rocknix Linux to provide native ARM64 Steam and Proton game support. Users can install the OS on internal storage or a microSD card, leveraging 16 GB of RAM...
Samsung Galaxy Ring 2: The Shocking Release Date Update
Samsung plans to unveil the Galaxy Ring 2, its next‑generation wearable, in the second half of the Galaxy Unpacked 2027 event. The ring promises a 9‑10‑day battery life, a thinner lighter profile, and upgraded health sensors for temperature, sleep and cardiovascular...
DIY Handheld LattePanda Linux Cyberdeck Build Guide From Scratch
Ben Makes Everything’s DIY guide walks readers through building a handheld cyberdeck centered on a LattePanda μ x86 single‑board computer. The device packs a 7‑inch 1080p 120 Hz touchscreen, a thumb‑typing keyboard, 1 TB SSD, and a custom 4S lithium‑ion battery with USB‑C...
Computex 2026: Adata Trusta AI Scaler Extended Memory Solution Breaks GPU Limits
Adata’s enterprise storage brand Trusta unveiled the AI Scaler Extended Memory Solution, featuring the AI Scaler Toolkit that lets on‑premises AI workloads spill over from GPU memory to system DRAM and high‑speed SSDs. The integrated hardware‑software approach claims more than...