https://digestai.news/category/hardwareHardware & Compute stories from Digest AI, with sources.enhttps://digestai.news/story/microsoft-cancels-uk-ai-data-center-project-due-to-power-delayshttps://digestai.news/story/microsoft-cancels-uk-ai-data-center-project-due-to-power-delaysSat, 24 Oct 2026 17:00:00 GMTMicrosoft has terminated its anchor-tenant agreement for a planned AI data center in Loughton, Essex, developed by Nscale. The decision, made on October 18, 2026, stems from UK Power Networks informing the project that adequate grid power would not be available until the early to mid-2030s. The facility was intended to open in late 2026 or early 2027 and was designed to supply computing power to…https://digestai.news/story/cerebras-explains-how-to-run-3t-models-on-wse-3t-chipshttps://digestai.news/story/cerebras-explains-how-to-run-3t-models-on-wse-3t-chipsThu, 22 Oct 2026 00:00:00 GMTCerebras published a technical guide detailing how to deploy large language models on its WSE-3T wafer-scale chips. The company used the open-source Kimi K3 model, which has approximately 3 trillion parameters, as a case study. The post explains that a single WSE-3T wafer contains 44 GB of on-chip memory, which is insufficient for the terabytes of weights required by frontier models. To solve…https://digestai.news/story/nvidia-h200-rental-costs-rise-about-160-since-january-report-showshttps://digestai.news/story/nvidia-h200-rental-costs-rise-about-160-since-january-report-showsSat, 24 Oct 2026 16:19:00 GMTRenting an Nvidia H200 chip now costs approximately 160% more than it did at the start of 2026, according to recent data covering January 1 through mid-October. The price hike highlights ongoing demand for AI computing power that continues to exceed available supply. Other Nvidia chips have also seen rental price increases this year, including the B200 (up about 120%), the RTX PRO 6000 (up…https://digestai.news/story/nvidia-ai-revenue-grew-117-broadcom-grew-221-last-quarterhttps://digestai.news/story/nvidia-ai-revenue-grew-117-broadcom-grew-221-last-quarterSat, 24 Oct 2026 07:30:00 GMTNvidia and Broadcom are both benefiting from the surge in AI data center spending, but they serve different parts of the market. Nvidia sells general-purpose GPUs and complete server systems, leveraging its CUDA software platform which many developers already use. This allows customers to buy flexible computing solutions for training and inference without custom engineering. Broadcom takes a…https://digestai.news/story/zagita-tool-doubles-rtx-30-40-gpu-compute-driving-used-prices-uphttps://digestai.news/story/zagita-tool-doubles-rtx-30-40-gpu-compute-driving-used-prices-upSat, 24 Oct 2026 22:16:55 GMTA new open-source Linux tool called Zagita has unlocked the full BF16/FP16 compute rates on NVIDIA's RTX 30 and RTX 40 series GPUs. By using a kernel module that applies specific speed selectors before the driver binds, the tool doubles the instruction issuing speed for FP/BF16 FMA operations on GA102, GA103, GA104, and GA106 chips without requiring vBIOS flashing or hardware mods. Users report…https://digestai.news/story/nvidia-and-spil-plan-15x-expansion-of-co-packaged-optics-capacity-by-mhttps://digestai.news/story/nvidia-and-spil-plan-15x-expansion-of-co-packaged-optics-capacity-by-mSat, 24 Oct 2026 14:23:00 GMTNvidia CEO Jensen Huang announced at GTC Berlin that the company and packaging partner Siliconware Precision Industries (SPIL) will expand co-packaged optics (CPO) optical-engine capacity by 15 times by mid-2027, with a further doubling to 30 times by end of 2028. Huang said optical networking demand in AI data centers could grow 15 times as copper interconnects hit physical limits on power,…https://digestai.news/story/investors-watch-amd-google-and-amazon-challenge-nvidia-in-ai-chipshttps://digestai.news/story/investors-watch-amd-google-and-amazon-challenge-nvidia-in-ai-chipsFri, 23 Oct 2026 11:30:00 GMTNVIDIA currently dominates the AI chip market, but investors are closely monitoring its competitors as they seek a larger share of the industry. Major tech companies like AMD, Google, and Amazon are developing their own hardware to reduce reliance on external suppliers and lower costs for their customers. AMD is pushing its Instinct AI chips for demanding workloads, while Google uses custom TPU…https://digestai.news/story/google-and-voltus-expand-virtual-power-plant-deal-for-1-070-megawattshttps://digestai.news/story/google-and-voltus-expand-virtual-power-plant-deal-for-1-070-megawattsSat, 24 Oct 2026 04:00:00 GMTGoogle and energy company Voltus announced an expanded agreement on Thursday to create a large virtual power plant across the PJM electric grid, which serves about 67 million people in 13 states and Washington, D.C. The project targets an additional 1,070 megawatts of capacity by coordinating thousands of smaller energy resources at homes and businesses, such as smart thermostats and batteries.…https://digestai.news/story/cisco-and-nvidia-unveil-usfp-module-for-ai-factory-networkshttps://digestai.news/story/cisco-and-nvidia-unveil-usfp-module-for-ai-factory-networksSat, 24 Oct 2026 17:11:00 GMTCisco and NVIDIA have announced the Ultra-Dense Small Form Factor Pluggable (USFP), a high-density optical module designed for AI factory networks. The technology facilitates data transmission speeds of 400 gigabits per second (Gb/s) per lane, doubling the current 200Gb/s per lane capabilities. This allows for an overall capacity of 12.8 terabits per second (Tb/s) while maintaining a compact…https://digestai.news/story/ciena-ceo-says-ai-demand-may-outpace-supply-for-its-optical-networkinghttps://digestai.news/story/ciena-ceo-says-ai-demand-may-outpace-supply-for-its-optical-networkingSat, 24 Oct 2026 15:40:01 GMTGary Smith, chief executive of networking equipment maker Ciena, told the Wall Street Journal that demand for AI‑driven workloads is growing faster than the company can currently produce its optical networking chips. Smith, who survived the 2001 telecoms crash, described his outlook as “suitably paranoid, but optimistic,” noting that the surge in AI training and inference traffic is putting…https://digestai.news/story/tom-s-hardware-premium-weekly-roundup-covers-tp-link-surface-laptop-ulhttps://digestai.news/story/tom-s-hardware-premium-weekly-roundup-covers-tp-link-surface-laptop-ulSat, 24 Oct 2026 12:25:00 GMTTom's Hardware published its weekly Premium roundup for the week ending October 24, 2026. The edition summarizes several subscriber-only articles. One piece examines TP-Link's regulatory challenges in the U.S., noting the company's pledge of nearly $1 billion in U.S. R&D and manufacturing and a letter of intent with Key Tronic to assemble Wi-Fi 8 routers in Mississippi. Another article reviews…https://digestai.news/story/nvidia-rtx-5090-shipment-470-000-stolen-from-brazil-airport-card-listehttps://digestai.news/story/nvidia-rtx-5090-shipment-470-000-stolen-from-brazil-airport-card-listeThu, 22 Oct 2026 16:18:47 GMTMultiple sources report that hundreds of Nvidia GeForce RTX 5090 graphics cards have been diverted to Chinese modification facilities where they are dismantled for AI server use. Technicians remove the GPU dies and 32GB GDDR7 memory chips from retail cards, including Founders Edition and premium models like the Asus ROG Astral RTX 5090, and reinstall them on custom PCBs with dual-slot blower…https://digestai.news/story/sk-hyper-seeks-350b-to-fund-5gw-ai-data-center-buildout-in-south-koreahttps://digestai.news/story/sk-hyper-seeks-350b-to-fund-5gw-ai-data-center-buildout-in-south-koreaFri, 23 Oct 2026 11:02:02 GMTSK Hyper Co., the AI data center unit of SK Telecom, is in talks with tech companies and financial investors to secure $350 billion, equivalent to roughly 462.91 trillion won, CEO Jung Suk-keun told Bloomberg. The proposed figure covers equipment costs and the initial 5-gigawatt phase of a planned 15-gigawatt buildout in South Korea targeted for completion by 2035. The sum reflects an estimated…https://digestai.news/story/nvidia-amd-apple-local-ai-hardware-compared-for-inference-speedhttps://digestai.news/story/nvidia-amd-apple-local-ai-hardware-compared-for-inference-speedFri, 23 Oct 2026 16:30:00 GMTA new comparison evaluates local AI hardware from NVIDIA, AMD, and Apple, finding that the best choice depends on workflow compatibility rather than raw speed alone. The NVIDIA DGX Spark offers the safest software bet with its 128GB unified memory and CUDA ecosystem, allowing existing PyTorch and TensorRT code to run without modification. However, its 273 GB/s memory bandwidth is the slowest of…https://digestai.news/story/amazon-stores-5-billion-in-nvidia-servers-due-to-power-shortageshttps://digestai.news/story/amazon-stores-5-billion-in-nvidia-servers-due-to-power-shortagesFri, 23 Oct 2026 16:54:00 GMTAmazon is holding billions of dollars worth of Nvidia AI servers in warehouses because its data centers lack sufficient electricity to power them, according to a Business Insider report. Internal emails reviewed by the outlet indicate that Amazon Web Services (AWS) expected to receive approximately $5 billion worth of Nvidia GB300 server racks in August, with projections for an additional $3…https://digestai.news/story/semianalysis-tests-trainium3-pytorch-stack-finds-better-profiling-thanhttps://digestai.news/story/semianalysis-tests-trainium3-pytorch-stack-finds-better-profiling-thanFri, 23 Oct 2026 19:06:44 GMTSemiAnalysis engineers spent weeks testing Amazon's Trainium3 racks with the new TorchNeuron backend and report a native PyTorch experience that works in eager mode with torch.compile and standard distributed APIs. They trained Qwen3-32B on 16 Trainium2 chips at 23.89% MFU (eager) and 30.47% MFU (compiled), and tested gpt-oss 20B on Trainium3 with a 27-minute first-step compilation time.…https://digestai.news/story/nvidia-launches-openshell-reference-architecture-for-safe-enterprise-ahttps://digestai.news/story/nvidia-launches-openshell-reference-architecture-for-safe-enterprise-aWed, 21 Oct 2026 02:47:00 GMTNVIDIA announced the OpenShell Reference Architecture at GTC Berlin on October 21, 2026, a full-stack blueprint for building, containing, observing, and enforcing safe AI agents across enterprise infrastructure and edge systems. The architecture combines NVIDIA Vera Rubin infrastructure for model inference, Vera CPUs for agent harnesses running the open-source OpenShell secure runtime, and…https://digestai.news/story/hpe-plans-first-moon-supercomputer-benchmarks-with-spaceborne-computerhttps://digestai.news/story/hpe-plans-first-moon-supercomputer-benchmarks-with-spaceborne-computerWed, 21 Oct 2026 10:30:00 GMTHewlett Packard Enterprise (HPE) announced it will run the first high-performance computing benchmarks on the Moon using its Spaceborne Computer-4 module. The fanless, graphics-card-sized blade will ride an Astrolab FLIP rover to the lunar south pole aboard a Voyager Technologies Griffin-1 lander, launching on a SpaceX Falcon Heavy in late 2026 under NASA's Commercial Lunar Payload Services…https://digestai.news/story/nvidia-says-1gw-ai-factory-costs-50-billion-unveils-dsx-podhttps://digestai.news/story/nvidia-says-1gw-ai-factory-costs-50-billion-unveils-dsx-podWed, 21 Oct 2026 03:59:00 GMTAt Nvidia's GTC Berlin keynote, CEO Jensen Huang stated that a 1-gigawatt AI factory costs about $50 billion and could generate roughly $40 billion in annual revenue, with delays potentially costing operators about $10 billion in foregone revenue. He introduced DSX POD, a reference design for Vera Rubin AI factories built in modular containers, which Nvidia says can reduce deployment time from…https://digestai.news/story/nvidia-bets-cell-towers-become-ai-servers-by-2027https://digestai.news/story/nvidia-bets-cell-towers-become-ai-servers-by-2027Wed, 21 Oct 2026 03:11:00 GMTNokia and Nvidia are collaborating with telecom operators to integrate artificial intelligence directly into radio access networks, a concept known as AI-RAN. By combining Nvidia’s CUDA platform and Aerial RAN Computer (ARC) with Nokia’s anyRAN software, the partners aim to build AI-native infrastructure that improves signal processing rather than adding it as an afterthought. This approach is…https://digestai.news/story/openai-scales-azure-cosmos-db-to-serve-1-billion-chatgpt-usershttps://digestai.news/story/openai-scales-azure-cosmos-db-to-serve-1-billion-chatgpt-usersFri, 23 Oct 2026 17:00:00 GMTOpenAI has detailed how it scaled its internal storage platform, Habitat, to support over 1 billion users across 40 geographic regions. The system now manages more than 500 petabytes of data, with the largest single Azure Cosmos DB account handling over 20 million queries per second. The engineering team describes a complex journey from single-tenant to multi-tenant architectures, driven by the…https://digestai.news/story/razer-basilisk-v4-pro-adds-ai-prompt-button-to-gaming-mousehttps://digestai.news/story/razer-basilisk-v4-pro-adds-ai-prompt-button-to-gaming-mouseFri, 23 Oct 2026 16:00:00 GMTRazer has released the Basilisk V4 Pro, an update to its all-around gaming mouse line that introduces a dedicated AI prompt button. The mouse features Razer's Focus Pro 50K Gen-3 optical sensor, supporting up to 50,000 DPI and 930 IPS, along with Gen-4 optical switches rated for over 100 million clicks. It offers three connectivity options: 2.4GHz wireless, Bluetooth, and wired USB-C. The device…https://digestai.news/story/apple-unveils-299-homeview-smart-home-display-with-siri-aihttps://digestai.news/story/apple-unveils-299-homeview-smart-home-display-with-siri-aiFri, 23 Oct 2026 15:11:00 GMTApple introduced HomeView, a $299 7‑inch smart‑home hub that combines a display, speaker, FaceTime, Siri AI, and a camera. Preorders began on Oct 13 for the United States, Australia, Canada, India, New Zealand, and the United Kingdom, with shipments slated for Nov 13. The device can be placed on a tabletop or wall‑mounted, but it is not sold in the European Union because Siri AI is unavailable…https://digestai.news/story/nvidia-dsx-platform-enables-companies-to-overcome-power-bottleneck-inhttps://digestai.news/story/nvidia-dsx-platform-enables-companies-to-overcome-power-bottleneck-inThu, 22 Oct 2026 05:00:25 GMTNVIDIA has introduced RTX Spark, a new generation of slim Windows laptops designed for AI workloads, gaming, and content creation, featuring the custom N1X system-on-a-chip. The laptops offer up to 128GB of unified memory, allowing the CPU and GPU to share a single high-speed memory pool for improved performance on demanding AI tasks. The N1X chip is built on TSMC's 3-nanometer process, scales…https://digestai.news/story/kazakhtelecom-plans-to-add-more-than-1-000-gpus-to-data-center-valleyhttps://digestai.news/story/kazakhtelecom-plans-to-add-more-than-1-000-gpus-to-data-center-valleyFri, 23 Oct 2026 02:00:13 GMTNVIDIA, Pony.ai, and inDrive signed a memorandum of understanding on October 22, 2026 to explore bringing Pony.ai's seventh-generation robotaxis to inDrive's ride-hailing platform in Kazakhstan, with testing targeted for the first half of 2027 in Astana. The collaboration includes NVIDIA supplying its Hyperion platform and Halos safety system, and is part of NVIDIA's broader push into…https://digestai.news/story/opinion-credo-technology-group-unveils-1-6terabit-zeroflap-active-lighhttps://digestai.news/story/opinion-credo-technology-group-unveils-1-6terabit-zeroflap-active-lighThu, 22 Oct 2026 17:45:00 GMTCredo Technology Group (CRDO -10.04%) designs active electrical cables (AEC) that let hyperscalers link servers and switches at 800 gigabits per second over copper. The company also sells the retimer and digital signal processor chips inside those cables, plus optical DSPs, PCIe retimers and licensed IP. Revenue in the quarter ended Aug. 1 more than doubled from a year earlier, and the stock has…https://digestai.news/story/lava-says-it-identified-32-nvidia-blackwell-ultra-b300-gpus-in-china-thttps://digestai.news/story/lava-says-it-identified-32-nvidia-blackwell-ultra-b300-gpus-in-china-tFri, 23 Oct 2026 12:26:23 GMTLava, an Israeli security firm, says it identified 32 Nvidia Blackwell Ultra B300 chips inside the network of China Telecom, a state‑owned telecommunications provider. The chips are barred from export by U.S. controls, according to a report by The Information, and were spotted using Nvidia’s DCGM Exporter tool, which can be accessed beyond the data‑center network and sometimes lacks password…https://digestai.news/story/asus-offers-proart-p16-laptop-with-nvidia-rtx-spark-starting-at-2-800https://digestai.news/story/asus-offers-proart-p16-laptop-with-nvidia-rtx-spark-starting-at-2-800Thu, 22 Oct 2026 16:00:00 GMTThe reviewer examined ASUS’s new ProArt P16, the first notebook to ship with NVIDIA’s RTX Spark system‑on‑chip. In benchmark testing the laptop lagged behind an earlier ASUS A16 on PCMark 10 Applications by about 3,000 points and fell slightly short in Geekbench 6 single‑thread CPU scores. However, its GPU outperformed a MacBook Pro with an M5 Max by 4,700 points in Geekbench 6, and it posted…https://digestai.news/story/intel-says-ai-demand-outpaces-wafer-capacity-can-only-double-outputhttps://digestai.news/story/intel-says-ai-demand-outpaces-wafer-capacity-can-only-double-outputFri, 23 Oct 2026 11:45:14 GMTIntel’s leadership, including CTO Pushkar Ranade, warned that AI compute has been growing three‑times per year for the past decade and that the company could at most double its leading‑edge wafer production with the current global stock of chip‑making tools. The shortage is driven by a mismatch between surging AI workloads and the limited number of EUV lithography machines from suppliers such as…https://digestai.news/story/ai-can-speed-up-hardware-prototype-cycles-for-startupshttps://digestai.news/story/ai-can-speed-up-hardware-prototype-cycles-for-startupsFri, 23 Oct 2026 10:57:31 GMTAI is beginning to accelerate parts of the hardware build cycle, especially for startups that rely on rapid 3D‑printed prototypes. The 2026 Stanford AI Index reported that 88% of surveyed organisations used AI in at least one business function in 2025 and 70% used generative AI somewhere in the organisation. For hardware teams the focus is how much time AI can shave between deciding to change a…https://digestai.news/story/nvidia-secures-45-000-grace-gpus-for-kazakhstan-aws-launches-serverleshttps://digestai.news/story/nvidia-secures-45-000-grace-gpus-for-kazakhstan-aws-launches-serverlesThu, 22 Oct 2026 23:46:00 GMTNvidia announced that its partner Firebird has obtained 45,000 Grace Blackwell GPUs for a new AI data centre in Kazakhstan, with U.S. export authorization. The first phase is being built in Data Center Valley together with Kazakhtelecom and is slated for commercial launch in the third quarter of 2027. After the second phase, the facility is expected to reach roughly 500 megawatts of AI capacity.…https://digestai.news/story/nvidias-rtx-spark-superchip-outperforms-apple-and-amd-in-local-ai-benchttps://digestai.news/story/nvidias-rtx-spark-superchip-outperforms-apple-and-amd-in-local-ai-bencThu, 15 Oct 2026 12:00:21 GMTNVIDIA announced the worldwide launch of RTX Spark laptops on October 16, 2026, from manufacturers including ASUS, Dell, HP, Lenovo, Microsoft, and MSI. Priced starting at $2,599, the systems feature local AI agent support, creative applications, and gaming capabilities. Configurations vary by region, with two main laptop models: the RTX Spark N1X (6,144-core GPU, 20-core CPU, up to 128GB…https://digestai.news/story/nvidia-adds-samsung-ibiden-unimicron-amkor-and-spil-to-substrate-supplhttps://digestai.news/story/nvidia-adds-samsung-ibiden-unimicron-amkor-and-spil-to-substrate-supplWed, 21 Oct 2026 17:54:00 GMTNvidia announced at its GTC Berlin event that it is expanding its silicon substrate and packaging supply chain to keep up with AI hardware demand. The company said it will work with Samsung’s Electro‑Mechanics unit, Japanese substrate maker IBIDEN, Taiwanese player Unimicron, Korean packaging firm Amkor and optics outfit SPIL. The plan calls for a 2.5x increase in substrate capacity over the…https://digestai.news/story/mit-uses-ferveret-cooling-claiming-up-to-35-more-ai-compute-for-nucleahttps://digestai.news/story/mit-uses-ferveret-cooling-claiming-up-to-35-more-ai-compute-for-nucleaThu, 22 Oct 2026 22:15:29 GMTMIT’s Materials in Extreme Environments Lab, led by assistant professor Ericmoore Jossou, is employing Ferveret’s Adaptive Phase Cooling platform to run AI‑driven analyses of nuclear materials. The collaboration focuses on two projects: one that applies physics‑informed deep generative models to Bragg coherent diffraction imaging data, and another that trains machine‑learning interatomic…https://digestai.news/story/nvidia-launches-vera-rubin-cpu-platform-targeting-up-to-1-5x-faster-cohttps://digestai.news/story/nvidia-launches-vera-rubin-cpu-platform-targeting-up-to-1-5x-faster-coWed, 21 Oct 2026 22:05:00 GMTNvidia announced the Vera Rubin platform, its first stand‑alone CPU built for agentic AI. The company says the chip delivers up to 1.5x faster performance per core compared with CPUs based on x86 architecture. Nvidia is on track to generate $20 billion in CPU revenue this year and estimates the market opportunity at $200 billion. For context, Nvidia’s data‑center revenue was $89 billion in the…https://digestai.news/story/nvidia-expands-confidential-computing-with-confidential-bare-metal-forhttps://digestai.news/story/nvidia-expands-confidential-computing-with-confidential-bare-metal-forWed, 21 Oct 2026 10:30:02 GMTNvidia announced the Confidential Bare Metal option for its Confidential Computing platform, letting frontier model builders run sensitive workloads on dedicated third‑party bare‑metal servers while keeping full control of models and data. The new deployment adds hardware‑rooted isolation, cryptographic attestation, and platform‑integrity checks across Nvidia’s Grace Blackwell, Rubin and other…https://digestai.news/story/nintendo-switch-2-jailbreak-created-with-openai-codex-says-modderhttps://digestai.news/story/nintendo-switch-2-jailbreak-created-with-openai-codex-says-modderThu, 22 Oct 2026 17:21:31 GMTModder Sammy Hajhamid, known as “pepsipu,” demonstrated a software‑only jailbreak for the Nintendo Switch 2 that he says was built with help from OpenAI’s Codex. The exploit, dubbed “ns2jb,” runs a custom version of Mario Kart World that includes characters such as Cinnamoroll, the red Angry Bird, Peter Griffin and Steve from Minecraft. Hajhamid’s claim about Codex’s role comes directly from…https://digestai.news/story/microsoft-launches-ai-supercomputer-infrastructure-team-led-by-danielhttps://digestai.news/story/microsoft-launches-ai-supercomputer-infrastructure-team-led-by-danielWed, 21 Oct 2026 17:00:00 GMTMicrosoft has created a new AI Supercomputer Infrastructure team to handle its most demanding AI training and inference workloads. The team is led by Daniel Firestone, a Microsoft veteran who spent over a decade deploying SmartNICs and networking components, and was responsible for the dataplane of Azure Boost, Microsoft's high-performance virtualized infrastructure for data center networking…https://digestai.news/story/nirva-raises-25-million-ahead-of-november-wearable-ai-companion-launchhttps://digestai.news/story/nirva-raises-25-million-ahead-of-november-wearable-ai-companion-launchThu, 22 Oct 2026 14:40:01 GMTNirva, a startup building a wearable AI personal companion that can give relationship advice and other assistance, announced a $25 million financing round. The funding brings the company’s total capital to $33 million. CEO Wei Lyu told Axios Pro the money will support the product’s planned launch in November. The device is positioned as a hands‑free AI partner that users can wear, extending the…https://digestai.news/story/navitas-gains-after-nvidia-800v-board-launch-ai-infrastructure-power-whttps://digestai.news/story/navitas-gains-after-nvidia-800v-board-launch-ai-infrastructure-power-wThu, 22 Oct 2026 02:14:00 GMTNavitas reported Q2 revenue of $10.5 million, a 27% year‑over‑year decline, and guided Q3 revenue to about $13.5 million, implying roughly 28% sequential growth. Shares rose 4.8% to $9 after the announcement. The company also closed its acquisition of Claros on October 6 for up to $232.8 million in cash, stock and milestones, adding integrated voltage‑regulator technology to its portfolio.…