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
TECNO CAMON 50 Unites Technology, Fashion and Productivity in One Next-Generation Device
TECNO unveiled its CAMON 50 series at MWC 2026, positioning the lineup as the centerpiece of an AI‑powered ecosystem that spans smartphones, AIoT devices, and productivity tools. The devices combine fashion‑forward design, 5G connectivity, and a new LightMaster imaging engine that uses AI to balance shadows, reduce glare, and deliver super‑zoom clarity. Integrated AI productivity features enable instant idea capture, organization, and content creation, turning the phone into a daily work companion. By blending style, intelligence, and performance, TECNO aims to redefine the mid‑range smartphone market and attract creators seeking both aesthetics and advanced imaging.
UK's New Supercomputer Is Coming 13 Months Too Late, NAO Report
The National Audit Office reports that the United Kingdom’s next‑generation supercomputer will miss its original 2026 launch by roughly 13 months, now expected in early 2027. The delay stems primarily from prolonged government funding approvals and procurement bottlenecks. The project,...
Nvidia GTC 2026: HPE Alletra Storage MP X10000 Becomes First Nvidia-Certified Storage Object-Based Platform for Enterprise AI
HPE announced that its Alletra Storage MP X10000 has earned Nvidia‑Certified Storage Foundation validation, becoming the first object‑based platform to achieve this milestone. The certification confirms the system can sustain AI workloads scaling to 128 GPUs, delivering the throughput and...
Nvidia GTC 2026: KIOXIA Announces New SSD Model Optimized for AI GPU-Initiated Workloads
At Nvidia’s GTC 2026, KIOXIA unveiled its new Super High IOPS SSD, the GP Series, designed for direct GPU access in AI workloads. The drive leverages KIOXIA’s XL‑FLASH storage class memory to deliver higher IOPS, 512‑byte granularity, and lower power...
Nvidia GTC 2026: Vdura Unveils RDMA Support and Context-Aware Tiering for GPU-Native AI Infrastructure
Vdura announced GPU‑native RDMA support for its Data Platform, allowing direct memory access between GPUs and storage without CPU involvement. The company also previewed Phase 1 of Context‑Aware Tiering, which will automatically move data across local NVMe SSD, DRAM and persistent...
Breker Hosts an Energetic Panel on Spec-Driven Verification
At DVCon 2026, Breker hosted a panel on AI‑driven, spec‑driven SoC verification featuring leaders from Moores Lab, Breker, Arm, and Intel. Participants highlighted that AI can translate raw specifications into intermediate representations with roughly 90‑95% accuracy, but emphasized the need for...
Bluetti AC240 Review: One of the Most Durable Power Stations I've Tested
The ZDNET review crowns the Bluetti AC240 as a top‑tier portable power station, highlighting its IP65 rating and 1,536 Wh LiFePO₄ battery. It delivers 3,600 W of continuous power, can charge to 80 % in just 45 minutes, and supports up to 1,200 W of...
Mirrorcle Announces Renewed Orders From Taara at OFC 2026
Mirrorcle Technologies announced renewed production orders from Taara at OFC 2026, extending a decade‑long partnership that underpins Taara's Lightbridge terrestrial laser‑communication system. The Lightbridge product delivers 20 Gbps over 20 km using Mirrorcle's MEMS fine‑steering mirrors, now moving into volume production. Mirrorcle...
Nokia Rebuilds Its Optical Engine, One Building Block at a Time
Nokia’s $2.3 billion acquisition of Infinera is bearing fruit as the company unveils a modular "building‑block" optical engine featuring four new DSPs and interchangeable indium‑phosphide, silicon‑photonic, and lithium‑niobate front ends. The approach lets customers mix and match components to serve 13...
IPhone 17 vs 17e: 15 Reasons to Spend the Extra $200
Apple launched the budget‑friendly iPhone 17e at $599, while the standard iPhone 17 starts at $799. The price gap translates into a suite of upgrades, including a smoother ProMotion display, longer battery life, enhanced camera optics, and advanced spatial‑computing hardware....
UK Startups Offered Use of National AI Supercomputers
Britain’s Sovereign AI Unit has launched the AIRR programme, allocating £500 million to give strategically significant UK AI startups access to national supercomputing resources. Around 10 % of the capacity of the Isambard‑AI and DAWN supercomputers—valued at roughly £20 million per year—will be...
Microwave Quantum Network Shows Resilience Against Heat-Related Disturbances
Researchers in Shenzhen have built a superconducting microwave quantum network that remains coherent despite thermal noise, using radiative cooling and tunable couplers to purge heat photons. The system transmits quantum states through a channel warmed to up to 4 K and...
Salicru Strengthens Indonesia’s Railway Infrastructure
Salicru has expanded its footprint in Indonesia’s railway sector by deploying uninterruptible power‑supply (UPS) solutions for PT Kereta Api Indonesia (PT KAI) across key hubs in Java and Sumatra. Working with local partner PT BINA SERVICES, the company installed new...
Nvidia's CEO Goes Full Principal Skinner in Response to DLSS 5 Backlash — Says It's the Gamers Who Are 'Completely...
Nvidia CEO Jensen Huang defended the newly unveiled DLSS 5 at GTC 2026, dismissing gamer criticism as "completely wrong." The technology claims to fuse generative AI with developer‑controlled rendering to enhance graphics on RTX 5000 GPUs. Early demo footage, however, sparked backlash for...
Hesai Brings LiDAR to E-Scooters as Per-Unit Costs Plummet
Hesai has slashed LiDAR prices to about $200 per unit, enabling its FTX blind‑spot sensor to be fitted on Niu Technologies' new NXT2 e‑scooter. The partnership marks a turning point for affordable perception technology, previously limited to high‑end EVs and...
Samsung Agrees AI Memory Deal with AMD
Samsung Electronics signed a memorandum of understanding with AMD to supply next‑generation AI memory, including HBM4 for AMD’s MI455X accelerators, DDR5 for the Helios system, and advanced DRAM for 6th‑generation EPYC CPUs. The agreement also opens discussions about a future...
Segway Navimow Robotic Lawn Mowers Earn TÜV Certification for Minimal Lawn Impact
Segway Navimow’s X420 and i206 AWD robotic lawn mowers have secured TÜV Rheinland’s Lawn Care certification, the first for any robotic mower. The certification confirms Level 2 minimal lawn impact after a simulated year of high‑frequency mowing. Both units demonstrated exceptional...
HyperLight Releases 400G-per-Lane TFLN Photonic ICs
HyperLight Corporation has launched a new family of 400‑gigabit‑per‑lane thin‑film lithium niobate (TFLN) photonic integrated circuits. The PICs deliver low insertion loss, low drive voltage, and ultra‑wide electro‑optic bandwidth, enabling energy‑efficient 400G optical links for AI‑focused networking. Built on HyperLight’s...
Hitachi Unveils 800 VDC Architecture for AI Data Centers
Hitachi announced an 800‑volt DC (800 VDC) power and control architecture, fully integrated into the Vera Rubin DSX reference design and compatible with NVIDIA Omniverse DSX Blueprint. The solution uses an OpenUSD‑based 3D simulation that mirrors the entire electrical chain from...
POET and LITEON to Co-Develop Optical Modules for AI Applications
POET Technologies and Taiwan‑based LITEON Technology have announced a strategic partnership to co‑develop optical communication modules for AI‑driven data centers. The collaboration will combine POET’s patented optical interposer platform with LITEON’s optoelectronic and power‑management expertise to produce compact, thermally optimized...
Northern Virginia’s Data Center Alley Processes 70% of Global Internet Traffic
Northern Virginia’s “Data Center Alley,” home to 570 facilities, processes an estimated 70% of global internet traffic. The region’s advantages—affordable power from natural gas and nuclear, abundant water, and a dense fiber network—have attracted major cloud providers. Energy demand from...
Samsung Expects Move to Longer-Term Chip Contracts
Samsung Electronics co-CEO Jun Young-hyun announced plans to shift from quarterly or annual memory‑chip deals to three‑ to five‑year contracts, aiming to stabilise supply amid a shortage projected to last until 2030. The move aligns with the rise of custom...
Semiconductor Market Revenue Surpasses $830 Bn in 2025 as AI Demand Fuels Industry Growth
The global semiconductor industry broke the $830 billion revenue barrier in 2025, marking roughly a 20% increase over 2024. Artificial‑intelligence workloads, especially in data‑center processors, GPUs and high‑bandwidth memory, were the primary growth engine. Memory and data‑center segments posted the strongest...
Coherent Demos InP Technology Innovation at OFC
Coherent Corp used its OFC 2026 booth to unveil a broad suite of indium phosphide (InP) components, including high‑power continuous‑wave lasers, electro‑absorption‑modulated lasers, and high‑speed photodiodes. The portfolio targets co‑packaged optics, silicon‑photonic pluggables and 1.6‑terabit transceivers, delivering up to 400 G...
Vislink Debuts DragonFly V 5G Dual-Modem Video Transmitter
Vislink Technologies has introduced the DragonFly V 5G Dual‑Modem, a miniature bonded video transmitter that combines two cellular links with Wi‑Fi. The device supports live streaming up to 1080p 50/59 from compact cameras, helmet‑mounted rigs, and UAVs. It integrates with Vislink’s LinkMatrix...
Southeast Asia’s AI Data Centre Gold Rush Tests Power Grids in the Tropical Heat
Southeast Asia’s data‑centre boom, driven by AI workloads, is testing the region’s power grids and cooling limits. Singapore alone houses over 70 facilities with 1.4 GW capacity, where cooling can account for up to 40% of total energy use. Operators are...
A Quantum Leap for the Turing Award
Charles Bennett and Gilles Brassard, pioneers of quantum information theory, have been awarded the ACM A.M. Turing Award. Their 1979 collaboration produced the BB84 protocol, the first quantum key distribution scheme, and laid the groundwork for quantum teleportation and the...
Lumentum Showcases Optical Scale-Up Demo at OFC Using VCSELs
Lumentum demonstrated a 1060 nm VCSEL‑based optical interconnect platform at OFC 2026, targeting rack‑level AI accelerators. The demo featured a high‑density multi‑mode VCSEL array co‑packaged with a host ASIC in a fan‑out wafer‑level package. The solution promises superior bandwidth, thermal resilience above...
Hesai Joins Nvidia Halos AI Systems Inspection Lab
Hesai Technology has joined Nvidia’s Halos AI Systems Inspection Lab, the first ANSI‑accredited facility for testing AI‑driven physical systems. The partnership obliges Hesai to evaluate its LiDAR platforms against a unified framework covering functional safety, cybersecurity and AI compliance. Hesai,...
AI Takes Center Stage at the 2026 CSUN Assistive Technology Conference
The 2026 CSUN Assistive Technology Conference highlighted AI as the catalyst behind a new generation of inclusive devices. Exhibitors showcased tactile displays, AI‑driven smart glasses, and real‑time captioning solutions that empower users with visual or hearing impairments. Companies such as...
Qnap TS-H1077AFU Review: A Great All-Flash Performance for Small Spaces
QNAP’s new TS‑h1077AFU is a 1U all‑Flash NAS targeting budget‑conscious enterprises. It pairs an 8‑core AMD Ryzen 7 7700 processor with 32 GB DDR5 (expandable to 192 GB) and ten hot‑swap SATA SFF SSD bays, while dropping U.2 NVMe support for a lower price...
BenQ MA270S 5K Monitor Targets Mac Users with Thunderbolt 4 and HDR400
BenQ unveiled the MA270S, a 27‑inch 5K monitor aimed at macOS users and professional creators. The display offers a 5120 × 2880 IPS panel with 99 % DCI‑P3 coverage, HDR400, and a 70 Hz refresh rate, while integrating a Thunderbolt 4 port that supplies up...
From MP3s to Flagship Phones: Oppo’s Time Capsule
Oppo got its start as a maker of MP3 players in the early 2000s, and now it is a maker of premium cutting edge mobile products. At Oppo's Dongguan campus, all the previous products live on in this futuristic time...
M5 MacBook Pro Feels Like an Entirely New Machine
The new MacBook Pro in many ways is the same as my old one, but going from M1 to M5 and the overall speed and power makes it feel like something very different. https://t.co/65h89Z2Y0T
OPPO A6s 5G Launched in India: Price and Specifications
OPPO has launched the A6s 5G smartphone in India, offering two memory variants priced between ₹18,999 and ₹20,999. The device features a 50 MP primary rear camera, a 6500 mAh battery with 45 W SUPERVOOC fast charging, and runs on MediaTek Dimensity 6300 SoC with...
Neuromorphic Chips Could Make Data Centers Grow, Not Build
The human brain uses only 20 watts of power but performs ~1 exaFLOP (10^18 operations/sec). Today's top AI chips burn 700 watts for ~1 petaFLOP. We're still ~1,000x less efficient than biology. When neuromorphic chips close that gap, we won't...
Beijing Gives Nvidia Implicit Green Light for H200
Nvidia gets Beijing's nod for H200 chip sales, adapts Groq chip for China, sources say Not really "approval", more like, Beijing will not stand in the way of company preferences.... https://t.co/nmGLKjWY6Z
NovAtel Releases GAJT-AE3 GNSS Anti-Jam Antenna for Airborne Applications
Hexagon’s NovAtel has launched the GAJT‑AE3, a compact GNSS anti‑jam antenna aimed at airborne platforms. The unit is the first in the GAJT line to offer full multi‑constellation, multi‑frequency protection, creating up to seven nulls per band and providing multi‑jammer...
Plaud Pro Launches in India at ₹19,490
Plaud is finally selling in India through Amazon. The company is now officially selling its new Plaud Pro device for ₹19,490. The device sells for $189 in the U.S. https://t.co/O2S52C07LL
Samsung S26 Ultra Cases Fail 25W Wireless Charging
Samsung's own S26 Ultra cases can't reliably deliver the 25W wireless charging it advertises, claims dbrand ✅ Details - https://t.co/eDbr8A7clm https://t.co/8RcwPC1y4b
Scalvy and Valeo Validate Modular EV Power Architecture
Scalvy and Valeo have completed a concept evaluation of Scalvy’s distributed Power Neuron architecture for electric‑vehicle powertrains under WLTC laboratory conditions. The modular platform achieved a peak inverter efficiency of 98.3 % at 10,000 RPM, kept motor temperatures below 62 °C and power‑device...
Akamai Launches First Global Nvidia AI Grid Deployment
.@Akamai Inference Cloud deploys Nvidia AI Grid https://t.co/L2i6iabxJ7 Nvidia GTC 2026 is turning out to be a bit of a coming-out party for Akamai's Inference Cloud. Akamai's Inference Cloud is the first global implementation of Nvidia AI Grid, which routes...
CoreWeave Adds Nvidia Blackwell Ultra Instances, Integrates W&B
.@CoreWeave launches latest @Nvidia instances, adds Weights & Biases tools https://t.co/MH7YFtkal4 CoreWeave said its Nvidia HGX B300 instances, part of the Nvidia Blackwell Ultra platform, are generally available on its cloud infrastructure and said it will be among the…
Toshiba Releases AEC-Q101 Photocoupler for EV Power Systems
Toshiba introduced the TLX9920 photocoupler, a photovoltaic‑output device designed for automotive solid‑state relays (SSRs). It delivers 5,000 Vrms isolation, an 8 mm creepage distance, and operates from –40 °C to 125 °C. Qualified to AEC‑Q101, the component targets high‑voltage applications such as battery‑management systems,...
NVIDIA RTX PRO 4500 Blackwell Server Edition: 10,496 Cores, 32 GB GDDR7, and Finally a GPU that Fits in a...
NVIDIA introduced the RTX PRO 4500 Blackwell Server Edition, a single‑slot, passively cooled GPU delivering 10,496 CUDA cores, 32 GB of GDDR7 memory, and a 165 W TDP. The card targets enterprise workloads such as inference, virtual desktops, video encoding, and edge AI, offering...
NVIDIA’s AI Push Is Getting Even Bigger; Huang Raises the Target to $1 Trillion
At GTC 2026, Nvidia CEO Jensen Huang announced a new AI infrastructure outlook, targeting at least $1 trillion in orders for current and future AI chip systems by the end of 2027. This doubles the previous $500 billion forecast and reflects accelerating demand...
Money and Me: The AI Capex Spend - Who Are the Current Winners?
In this episode, Michelle Martin and AI‑investment specialist Alvin Chow dissect the $700 billion AI infrastructure spend, highlighting the massive CapEx by hyperscalers like Amazon, Google, Microsoft, Meta and Oracle and the ripple effect across hardware, memory, storage and photonics suppliers....
Parliamentary Panel Suggests MeitY Seek Expert Views on GPU Plans
India’s Parliamentary Standing Committee on Communications and IT warned the Ministry of Electronics and IT (MeitY) that high GPU hardware costs, global supply‑chain delays, and the massive power‑water footprint of data centres threaten the country’s push for large GPU clusters....
Intel Unveils Core Ultra 200HX Plus Mobile CPUs, Promising Modest Gaming Gains
Intel announced the Core Ultra 200HX Plus mobile processor family, including the 290HX Plus and 270HX Plus SKUs, in a press release today. The chips are built on an updated Arrow Lake architecture and target high‑performance laptops with discrete graphics....
Nvidia Unveils BlueField‑4 STX Architecture to Power AI‑Optimized Storage Systems
Nvidia Corp. announced the BlueField‑4 STX reference architecture at its GTC developer event, unveiling a hardware blueprint for AI‑centric storage systems. The design, featuring the BlueField‑4 DPU, Spectrum‑X switches and ConnectX‑9 SuperNICs, is slated for partner shipments in the second...