← News

HARDWARE

74 stories

  1. Apple's new 14-inch MacBook Pro with M6, open and angled on a dark background with a glowing M6 chip mark behind it, in Apple's official launch imageHARDWAREApple's M6 lands in a $1,999 MacBook Pro and a $1,499 iMac, with AI bolted into every GPU coreApple announced the 14-inch MacBook Pro and the 24-inch iMac with its new M6 chip on Tuesday, Oct. 27, 2026, with preorders starting Wednesday, Oct. 28 and availability Friday, Nov. 6.
  2. Close-up of the underside of Anduril's BlueChip laptop, engraved 'BlueChip, Anduril Industries, Designed in California, Built in America' with a U.S. flag, in Anduril's official launch photoHARDWAREAnduril and Qualcomm unveil BlueChip, a U.S.-built laptop for government and defenseAnduril and Qualcomm unveiled BlueChip on Tuesday, Oct. 27, 2026, a new product line starting with a laptop designed and assembled in the U.S. for government, defense, first responders and public safety.
  3. Apple HomePod mini 2 color lineup: white, blue, green, pink, and midnight speakers arranged on a light surfaceHARDWAREApple's HomePod mini 2 brings Siri AI to the $149 speaker can, and preorders are openApple announced HomePod mini 2 on Oct. 13, 2026, with preorders at $149 and availability beginning Friday, Oct. 30; Siri AI arrives on the speaker in a future software update.
  4. Two black Axon Body 5 body-worn cameras with green status lights, front and angled viewsHARDWAREAxon's Body 5 bodycam has 25x the AI power, and the Pro model scans licenses for warrantsAxon announced Axon Body 5 on Oct. 24, 2026, its newest body-worn camera, which the company says delivers 25 times the AI processing power of earlier-generation Axon cameras and comes for the first time in two versions, Body 5 and Body 5 Pro.
  5. Four Microduck biped robots from Pollen Robotics in cream, graphite, lavender and sky colorwaysHARDWAREThe robot duck Jensen Huang played with onstage is selling so fast it won't make ChristmasPollen Robotics, the Hugging Face robotics team behind the $399 Microduck, now tells shoppers on its official store that it "can't promise Christmas delivery" for new orders and estimates four to six months to ship, days after Nvidia CEO Jensen Huang demoed the robots onstage and Hugging Face CEO Clem Delangue said preorders were crossing 20,000.
  6. Waymo's "Hello Toronto!" launch graphic showing a white Waymo-branded Jaguar I-PACE robotaxi with roof sensors in front of the Toronto skyline and CN TowerHARDWAREWaymo is coming to Toronto, its first Canadian city, with robotaxi rides targeted for 2028Waymo on Friday, Oct. 16, 2026, said it is laying the groundwork for a fully autonomous ride-hailing service in Toronto, starting with human-driven mapping and testing in the coming days and aiming to welcome its first riders in 2028.
  7. NVIDIA RTX Spark laptop shown in exploded layers with AI creative software on screen and glowing Grace-Blackwell superchip visible under a translucent deckHARDWARENVIDIA's RTX Spark AI laptops are on sale today, starting at $2,599NVIDIA on Friday, Oct. 16, 2026, said RTX Spark Windows laptops are available today from major PC makers, packing a Blackwell RTX GPU and Grace CPU into slim notebooks meant to run local AI agents, creative apps, and games on one machine.
  8. Amazon Bee Frontier wearable AI device and Bee Eclipse magnetic phone attachment on a light surfaceHARDWAREAmazon's Bee launches two new always-listening AI devices, starting at $75Amazon on Thursday, Oct. 15, 2026, introduced Bee Frontier, a next-generation personal AI wearable, and Bee Eclipse, a credit-card-sized magnetic phone attachment, the first new Bee hardware since Amazon acquired the wearable maker — with Frontier available now at $74.99.
  9. U.S. District Judge Edward J. Davila of the Northern District of California in his black judicial robe and a red-and-blue striped tie, official court portraitHARDWAREOpenAI says an engineer took 100+ files on its secret device back to AppleOpenAI told a federal court in San Jose on Wednesday, Oct. 14, 2026, that a former employee who came from Apple, worked at OpenAI for six months and has since returned to Apple sent more than 100 highly confidential files about at least one unreleased OpenAI device to his personal accounts — and later presented OpenAI's product plans to Apple's senior leadership. Apple says it has found nothing to support the claims. At a hearing the same day, U.S. District Judge Edward Davila said he is "highly likely" to grant Apple an injunction in its own case against OpenAI.
  10. Salience Labs optical circuit-switch photonic chip on a green board with fiber ribbon arrays attached, labeled SalienceHARDWARESalience Labs raises $125 million to put optical switches inside AI data centersSalience Labs, the Oxford and Münster spin-out building photonic optical circuit switches for AI networking, announced an oversubscribed $125 million Series B on Wednesday, Oct. 14, 2026, led by Bullhound Capital with AMD Ventures, Intel Capital, SBI Group, Hudson River Trading and the British Business Bank among the new money.
  11. Apple HomeView smart display on a wooden living room credenza, its 7-inch screen showing a 68-degree thermostat and home controlsHARDWAREApple unveils HomeView, a $299 smart display built around Siri AI that knows who is looking at itApple on Tuesday, Oct. 13, 2026, introduced HomeView, a 7-inch smart home display powered by Apple Intelligence with Siri AI, the conversational new Siri. It recognizes household members by their faces and shows each one their own content, and it doubles as a speaker, FaceTime screen and smart home dashboard. HomeView costs $299 in tabletop and wall-mounted versions, with pre-orders starting today and availability Nov. 13. Apple also announced a new Apple TV 4K and HomePod mini 2 with Siri AI and a line of LG and Schneider Electric smart home accessories.
  12. A row of AMD Helion agentic AI racks, each black door lettered AMD Helion, in an AMD product renderHARDWAREAMD unveils Helion, a rack built for AI agents that can run on CPUs aloneAMD on Monday, Oct. 12, 2026, introduced Helion, an open rack-scale blueprint for agentic AI built on an Open Compute Project spec, at the OCP Global Summit in San Jose. Hyperscalers can get a CPU-only rack with up to 32,768 cores of its 6th Gen EPYC processors dedicated to running agents, while enterprises can get one rack that combines EPYC CPUs, Instinct MI350P GPUs and general-purpose compute. AMD says partner-built configurations are planned for November 2026.
  13. HARDWARESabi raises $50 million to build a baseball cap that reads brain signals and turns imagined words into AI promptsSabi, a San Francisco startup building a brain-reading baseball cap for talking to AI without typing or speaking, announced on Friday, Oct. 9, 2026, a $50 million seed round led by Khosla Ventures. Accel, Initialized Capital, former OpenAI chief product officer Kevin Weil, DST Global, Collaborative Fund and Ascend also took part. The company plans to show an early prototype at CES 2027.
  14. NVIDIA graphic comparing GPT-6 Astra Standard vs Ultrafast response times on NVIDIA AI infrastructureHARDWARENVIDIA says GPT-6 Astra Ultrafast on Blackwell is up to 8× faster than StandardOn Oct. 1, 2026, NVIDIA said GPT-6 Astra Ultrafast — running on Blackwell GPUs with OpenAI-tuned inference software — is available now and can generate tokens up to eight times faster than Astra Standard.
  15. Official Google Project Suncatcher logo lockup on yellow-to-magenta gradientHARDWAREGoogle's Project Suncatcher prototype satellite is in orbit with TPUs aboardGoogle Research said Thursday its Project Suncatcher prototype satellite — built with Planet and launched on SpaceX's Transporter-18 rideshare — is in contact and operating as expected, the first flight test in a moonshot asking whether space can host scalable machine-learning infrastructure.
  16. Muse Gadgets lineup on Meta card — Home Link puck, ESP32 boards, Raspberry Pi, e-ink displayHARDWAREMeta open-sources Muse Gadgets so you can put Muse on DIY hardwareMeta on Friday published Muse Gadgets — open-source ESP32 and Linux device SDKs (Apache 2.0) so people can build their own Muse-connected hardware — and opened a claim list for Muse Home Link, a USB-C home-network dongle Meta says it manufactured in a batch of 5,000.
  17. NVIDIA DGX Spark partner OEM units — Acer ASUS Dell Gigabyte HP Lenovo MSI — green frameHARDWARENVIDIA ships a $4,999 64GB DGX Spark for local AI amid the memory crunchNVIDIA on Friday said DGX Spark will be available with 64GB of unified memory from Acer, ASUS, Dell, Gigabyte, HP and MSI starting Oct. 23 at $4,999, keeping the GB10 Grace Blackwell Superchip and full NVIDIA AI stack while offering a cheaper on-ramp than the 128GB model.
  18. Desk graphic — U.S. Army creates FASCOM autonomous systems commandHARDWAREU.S. Army creates FASCOM to buy and field autonomous systemsThe U.S. Army is creating an Army Futures and Autonomous Systems Command and naming a portfolio acquisition executive for autonomy, according to a memo first reported by Axios on Friday, one day after Defense Secretary Pete Hegseth highlighted the Meridian and Agincourt robot-warfare projects.
  19. Bull logo on the facade of its Angers, France supercomputer factoryHARDWAREFrench Bull doubles Angers supercomputer output after €80M rebuildFrench state-owned supercomputer maker Bull said it doubled production at its Angers factory to 12 racks a month after an €80 million rebuild, formally opening the new production building on Friday as Europe races to expand AI compute.
  20. Lime and earth desk graphic with Amiga robot inset — Bonsai World satellite-to-3D sim for rugged physical AIHARDWAREBonsai Robotics ships Bonsai World to train farm robots before they arriveBonsai Robotics said Friday it launched Bonsai World, a simulation and world-model tool that turns satellite imagery of farms and mines into 3D environments so autonomous machines can train on conditions they have not yet physically encountered.
  21. Synopsys–TSMC OIP 2026 release graphic: agentic AI workflows, CoWoS IVR, 2nm IP, COUPE opticsHARDWARESynopsys and TSMC deepen AI chip-design stack with agentic EDA on A14Synopsys announced new collaborations with TSMC spanning A14 design enablement with agentic AI capabilities, AI-assisted analog, digital, and multi-die workflows, CoWoS support for integrated voltage regulators, and joint enablement for TSMC-COUPE co-packaged optics.
  22. Person wearing Meta VR Glasses, Meta’s ~100g VR glasses announced at Connect 2026HARDWAREMeta unveils Meta VR Glasses — ~100g VR in a glasses form factorAt Meta Connect, Meta introduced Meta VR Glasses — a ~100-gram VR glasses system with a clip-on compute puck, IMAX Enhanced cinema, immersive sports, and a private multi-screen workspace — going on sale Spring 2027 for $1,299.99.
  23. Official Meta Connect collage of Ray-Ban Meta AI glasses styles including Audio and Gen 3 framesHARDWAREMeta launches Ray-Ban Meta Audio and Gen 3 AI glasses; Muse comes to glassesAt Meta Connect, Meta introduced Ray-Ban Meta Audio — its first camera-free audio AI glasses — plus Ray-Ban Meta (Gen 3) and confirmed Muse, its personal AI agent, is coming to Meta AI glasses for hands-free tasks and goals.
  24. SPRIND × NADI AI-Native Chip Design Challenge claim card: EUR40M over 20 monthsHARDWARESPRIND and NADI put €40M into AI-native chip designGermany’s SPRIND and the Netherlands’ NADI opened a joint AI-Native Chip Design Challenge — up to €9.6M per team across two stages — to use AI agents to design energy-efficient chips for AI training and inference.
  25. NVIDIA DLSS 5 3D-Guided Neural Rendering promotional still for NBA 2K27 on GeForce RTX 50 SeriesHARDWARENVIDIA DLSS 5 brings 3D-guided neural rendering to gamesNVIDIA detailed DLSS 5 with 3D-Guided Neural Rendering, a final neural stage that adds lighting and material detail while the game engine’s geometry stays in place, shipping in NBA 2K27 on GeForce RTX 50 Series, alongside ACE speech models and in-game inferencing updates for local AI.
  26. Intrinsic Core OMTS open machine tending reference solution graphicHARDWAREGoogle’s Intrinsic open-sources Intrinsic Core under Apache 2.0Google’s Intrinsic announced Intrinsic Core, an open-source set of ROS-compatible robotics capabilities under the Apache 2.0 license, plus an Open Machine Tending Solution reference design, at ROSCon 2026 in Toronto.
  27. T-Head booth at Apsara showing Zhenwu V900 AI training and inference chip displayHARDWAREAlibaba unveils Zhenwu V900 AI chip and 5–10T Qwen roadmap at ApsaraAt Apsara Conference in Hangzhou, Alibaba announced the Zhenwu V900 AI training and inference processor, said Qwen 4 is in training with Qwen 4.5 and Qwen 5 aimed at 5–10 trillion parameters, and set a 20 GW Alibaba Cloud capacity target by 2032.
  28. RealSense depth-sensing camera product close-upHARDWARECognex to buy RealSense for ~$500M cash in robotic perception pushCognex announced a definitive agreement to acquire RealSense, the depth-sensing and robotic-perception company originally founded at Intel, for approximately $500 million in cash, with closing expected in Q4 2026.
  29. Kodiak autonomous truck with PrePass Safety Alliance partnership graphicHARDWAREKodiak AI and PrePass connect driverless trucks to state roadside screeningKodiak AI and PrePass announced a collaboration that routes Commercial Vehicle Safety Alliance Enhanced Inspection results for Kodiak’s autonomous trucks into PrePass roadside screening so states can authorize weigh-station bypass, starting in Texas and Louisiana.
  30. Bad Signal claim diagram — Boston Dynamics Atlas RMAC training center GeorgiaHARDWAREBoston Dynamics opens Atlas training center at Hyundai MetaplantBoston Dynamics opened the Robotics Metaplant Application Center (RMAC) inside Hyundai Motor Group Metaplant America near Savannah, GA, to train Atlas humanoids on real auto-plant tasks.
  31. Einride autonomous freight Pod with Einride and NVIDIA logos at a warehouse dockHARDWAREEinride taps NVIDIA Hyperion for heavy-duty autonomous truckingEinride AB announced a strategic collaboration with NVIDIA to adapt the NVIDIA DRIVE Hyperion platform for heavy-duty autonomous trucking across a freight network already serving shippers in the U.S., Europe, and the Middle East.
  32. Bad Signal claim diagram — ams OSRAM Thin-Film VCSEL for AI scale-up optical interconnectsHARDWAREams OSRAM unveils Thin-Film VCSEL for AI scale-up linksPhotonics company ams OSRAM announced a next-generation 850 nm Thin-Film VCSEL platform for AI scale-up optical interconnects, citing error-free 32 Gb/s operation near 0.25 pJ/bit and a connectorized multi-core fiber demonstrator with BizLink at ECOC 2026 in Malaga.
  33. Bad Signal claim diagram — Light-O1 turns internet video into robot motionHARDWARELight-O1 turns internet video into robot motionLight Origins launched Light-O1, a whole-body intelligence model pretrained on human actions recovered from internet video, with demos retargeted to LightBot and Unitree G1, and released a 6-billion-parameter Light-O1-Preview checkpoint on Hugging Face under Apache 2.0.
  34. Bad Signal claim diagram — Japan factory machine IDs for physical AI dataHARDWAREJapan plans factory machine IDs to feed physical AINikkei Asia reports Japan’s government and leading machinery makers will work together to collect machine-learning data for physical AI such as autonomous robots, assigning unique IDs to individual pieces of factory equipment so information can be gathered across manufacturers.
  35. Puneet Chandok with Indian officials at a plaque for the Microsoft India South Central cloud regionHARDWAREMicrosoft names India South Central its AI hub for AsiaMicrosoft said its India South Central datacenter region in Hyderabad is live as a strategic hub for Asia and the Global South, with AI-ready hardware across its India cloud regions and mechanical cooling it describes as using effectively no water.
  36. MediaTek Dimensity CX C10 Max chip beside the product nameHARDWAREMediaTek unveils Dimensity CX C10 Max for Googlebook laptopsMediaTek introduced the Dimensity CX C10 Max, the first chip in its new Dimensity Compute Experiences family, built on TSMC's 3-nanometer process for an upcoming premium Googlebook laptop with on-device Gemini features.
  37. Opening page of the European Commission delegated regulation on a data-centre rating schemeHARDWAREEU proposes energy and water rating labels for big data centresThe European Commission proposed rules requiring European data centres of 500 kilowatts and above to disclose energy and water efficiency under a new EU labelling scheme, as Brussels plans to roughly triple data-centre capacity to support AI while worrying about grids and water.
  38. White ugo humanoid robot with blue LED eyes holding camera equipmentHARDWAREIFR: about 7,000 humanoid robots sold worldwide in 2025About 7,000 humanoid robots were sold worldwide in 2025 for industrial and professional service use, according to figures compiled by the International Federation of Robotics and reported by Reuters — a small base versus hundreds of thousands of conventional industrial robots, with many humanoids bought by labs or companies collecting AI training data rather than doing factory work.
  39. Faraday Future All-New Futurist humanoid robot product still, front view with one hand raisedHARDWAREFaraday Future puts nine new EAI robots on sale at its 919 launchFaraday Future said it launched nine new Embodied AI robot configurations across five models at its 919 event — including the All-New Futurist humanoid and Master Mini from $9,990 — plus education, research, security, and inspection solution packs, with orders open the same day.
  40. CXMT product shot: branded ChangXin Memory Technologies DRAM package on a circuit boardHARDWARECXMT says G5 DRAM platform is in mass productionChangXin Memory Technologies (CXMT) said its fifth-generation DRAM manufacturing platform has entered mass production and showed two 24-gigabit LPDDR5X chips already in volume production for smartphones and other portable devices.
  41. Official Waymo in Singapore graphic showing a Waymo all-electric Jaguar I-PACE against the Singapore skylineHARDWAREWaymo to bring autonomous rides to Singapore in 2028Singapore's Land Transport Authority said Waymo will become the city-state's newest autonomous-vehicle operator. Waymo plans to map and validate local roads with manually driven cars in 2027, then aims to open commercial autonomous ride-hailing in 2028.
  42. Huawei Connect keynote stage: Ascend 960 SuperPod / Atlas 960E NPO announcement on the main screenHARDWAREHuawei unveils Ascend 960 SuperPod with NPO at Connect 2026At Huawei Connect 2026 in Shanghai, Huawei’s China newsroom said rotating chairman David Wang unveiled the Ascend 960 SuperPod — framed as the industry’s first SuperPod using near-packaged optics (NPO / Hi-ONE) — and same-day wires report the Ascend 960DT training chip is now targeted for Q1 2027 and the 960PR inference chip for Q3 2027.
  43. Emerald AI chart showing AI cluster demand response vs total grid power demand in PhoenixHARDWAREEmerald AI, Google, NVIDIA launch AEMA for flexible AI data centersNVIDIA’s company blog said Emerald AI, Google and NVIDIA launched the AI Energy Management Alliance (AEMA) to advance AI data centers that can dynamically manage electricity use in response to grid conditions — aiming for faster interconnection while supporting reliability and affordability.
  44. NVIDIA Vera Rubin NVL72 rack-scale AI system official product renderHARDWARENVIDIA Vera Rubin NVL72 debuts on MLPerf Inference — up to 3.7x GB300NVIDIA’s company blog said Vera Rubin NVL72’s first MLPerf Inference v6.1 preview submission delivered up to 3.7x better throughput than GB300 NVL72 on Qwen3-VL, and up to 2.5x on DeepSeek-R1, with GB300 also showing 99% multi-rack scaling efficiency.
  45. Rune company product page: solar-farm arrays with modular RELIC-style DC compute units on a yellow dashed bus, DC BUS LIVE metrics, Fastest time to compute / 100% clean headlineHARDWARERune raises $40M Series A for RELIC solar-powered modular AI computeRune unveiled RELIC (Renewable Energy Linked Intelligent Compute), a modular compute system that installs directly at solar generation sites to turn wasted solar power into AI compute without a grid connection. Alongside the launch, the company announced a $40 million Series A led by Spark Capital, with Union Square Ventures, Lowercarbon Capital, Activate Capital, Committed Capital, Timeless Partners, and Logos Fund also participating, bringing total funding to $53.5 million. Rune says RELIC units can be forklift-deployed in about 60 minutes and bring GPUs online in as little as six weeks; it also disclosed a live installation at a 200MW solar facility in Texas.
  46. nubia NaviX Ultra phones in four colors beside Doubao Phone Assistant branding and listed China prices starting ¥5999 (¥5499 after subsidy)HARDWAREnubia ships NaviX Ultra — Doubao agent phone on sale from ¥5499ZTE’s China newsroom said nubia’s NaviX Ultra — framed as the world’s first AI agent smartphone running ByteDance’s Doubao Phone Assistant consumer edition — is formally on sale, with list prices from ¥5999 (¥5499 after China consumer subsidy) across three storage/RAM SKUs on ZTE Mall, JD, Tmall, Douyin, and Pinduoduo.
  47. XGRIDS Lixel L3 product release graphic showing the handheld spatial scanner with 5 mm relative accuracy badgeHARDWAREXGRIDS Lixel L3 claims 5 mm spatial capture accuracy at INTERGEOXGRIDS introduced Lixel L3 at INTERGEO 2026, a handheld spatial scanner the company says delivers validated 5 mm relative accuracy — and about 2.5 cm real-time absolute accuracy — by fusing LiDAR, true-color 4K HDR vision, integrated RTK, and a built-in display. From one scan it outputs true-color point clouds, meshes, and photorealistic 3D Gaussian Splatting models (company).
  48. HP ZBook Ultra G3a 16-inch mobile workstation open on a deskHARDWAREHP unveils ZBook Ultra G3a laptop built to run large AI models on-deviceHP Inc. announced the HP ZBook Ultra G3a 16”, a thin-and-light mobile workstation pitched as a “personal AI factory” for professionals. The company says the AMD-powered system can run large language models with up to about 300 billion parameters locally, comes with Perplexity Computer preloaded (including Autodesk Revit support), and is expected to be available in October 2026. Pricing will come closer to launch (company).
  49. Delos Data Nonstop AI funding and interconnect press graphicHARDWAREDelos Data raises over $100M for Nonstop AI inference networkingDelos Data said it closed more than $100 million from Matrix, Playground, Socratic Partners, Capricorn’s Technology Impact Fund, Matter Venture Partners, IAG, and industry investors to launch Delos Nonstop AI™. Alongside the raise, it introduced a Nonstop AI Data Interface and Reference Architecture aimed at lower-latency, more resilient data movement for agentic AI inference across mixed hardware.
  50. Polaris Electro-Optics FenGlass silicon photonics optical chip for AI interconnectsHARDWAREPolaris Electro-Optics raises $50M to scale FenGlass optics for AI data centersPolaris Electro-Optics said it raised $50 million in Series B financing led by Walden Catalyst Ventures to move its FenGlass electro-optic modulator platform from proven devices into large-scale commercial deployment for AI interconnects. The company says FenGlass integrates a proprietary ferroelectric nematic glass onto finished silicon photonic wafers and has demonstrated 400 Gbps per lane modulation with Telcordia reliability tests completed.
  51. Crusoe and Perplexity partnership announcement graphic with both company logosHARDWARECrusoe and Perplexity ink a multi-year deal for train-to-serve AI cloudCrusoe said Perplexity will run its full model lifecycle on Crusoe Cloud — training frontier models on dedicated NVIDIA GB300 NVL72 clusters with InfiniBand, then serving them in production through Crusoe Managed Inference. Crusoe will also roll Perplexity Enterprise Pro and Max to its 1,800 employees.
  52. RobotPlusPlus working-at-height industrial robot from the Series C announcementHARDWARERobotPlusPlus raises Series C to scale working-at-height industrial robotsBeijing-based RobotPlusPlus (ROBOT++) said it closed a Series C worth hundreds of millions of RMB — described as tens of millions of USD — led by Qianggang Capital Fund. The company builds robots that take people off scaffolding and other hazardous structures at shipyards, oil-and-gas sites, wind farms, and building facades, and says the new capital will fund global expansion and embodied AI development.
  53. Branch Energy Series B announcement image for Arc AI-power battery systemsHARDWAREBranch Energy raises $33M to deliver AI power with Arc batteries in weeksBranch Energy said it closed a $33 million Series B led by Piva Capital and Clean Energy Ventures. The Houston company sells Arc — a self-contained battery “grid in a box” about the size of a parking space — meant to store and deliver electricity on-site for AI data centers and other power-hungry sites faster than building new plants and transmission.
  54. Agility Robotics Digit 5 humanoid robot product hero, front viewHARDWAREAgility unveils Digit 5 humanoid built to work next to peopleAgility Robotics unveiled Digit 5, the next generation of its general-purpose humanoid. The company says Digit 5 is engineered for cooperatively safe work at scale — close to people without the physical safety barriers traditional industrial robots usually need — and cites more than $300 million in multi-year orders (as of May 2026, subject to milestones). Early access is expected in the first half of 2027, with broader availability by end of 2027.
  55. Einride cab-less autonomous electric truck with Lidl / Schwarz Group branding on a German roadHARDWAREEinride and Lidl start Germany’s first cab-less Level 4 truck runsEinride and Lidl in Germany said a cab-less SAE Level 4 autonomous electric truck is now running daily public-road deliveries between a Lidl warehouse and distribution center and a nearby store under an official Federal Motor Transport Authority (KBA) permit — the first authorization of its kind in Germany. The companies plan to expand toward a multi-stop “milk run” and discuss further Schwarz Group deployments.
  56. MediaTek Dimensity 9600 Pro flagship smartphone chip card artHARDWAREMediaTek launches Dimensity 9600 Pro, claiming first 2nm phone chipMediaTek said it launched the Dimensity 9600 Pro flagship smartphone chipset, describing it as the first company announcement of a mobile SoC — the main brain of a phone — built on a 2nm process node. The company pitches a dual-NPU / Agentic AI Engine stack for on-device models (claims up to about 30 billion parameters), plus gaming and imaging upgrades, and introduced a related Dimensity 9600M for broader flagship phones. First phones with both chips are expected this quarter, the company said.
  57. EUCLYD CRAFTWERK agentic AI silicon product conceptHARDWAREEUCLYD raises more than €200M Series A for efficient AI inference chipsEindhoven-based EUCLYD said it closed a Series A of more than €200 million (~$230 million) to build ultra-efficient AI inference infrastructure — co-led by Samsung, Somerset Capital Partners, EQT’s Scaleup Europe Fund, and Innovation Industries, with participation from EIFO, imec.xpand, BOM, and Quadri. Former ASML CEO Peter Wennink joins as board chair. The company is developing craftwerk “agentic AI silicon” and craftwerk station / CWS systems aimed at cutting power and cost per token for foundation-model inference.
  58. Waymo autonomous Jaguar with GO and Nihon Kotsu branding against Tokyo Tower skylineHARDWAREWaymo aims for fully driverless Tokyo robotaxis in 2027 with GO and Nihon KotsuAlphabet-owned Waymo, Japanese taxi-app operator GO, and Tokyo taxi company Nihon Kotsu said they have agreed to prepare a commercial fully driverless ride-hailing service in Tokyo for 2027. Rides would be bookable in the GO and Waymo apps, starting with a small fleet and expanding toward about 100 vehicles across key neighborhoods — after regulatory approval and ongoing safety validation.
  59. Blue circuit board macro with electronic componentsHARDWARENVIDIA shows ~10× faster dropless MoE training in JAXNVIDIA published a developer blog showing that Transformer Engine with JAX can deliver about a 10.4× end-to-end throughput gain for dropless Mixture-of-Experts training versus an unoptimized baseline on NVIDIA GB200-class GPUs. On DeepSeek-V3, the stack moved measured performance from about 103 TFLOPS per GPU to about 1,068 TFLOPS per GPU, and the company reported about 97% scaling efficiency out to 1,024 GPUs on NVIDIA GB300 NVL72 hardware. Dropless means every routed token is processed — no capacity-based dropping or padding — which is harder for GPUs because experts get uneven token counts.
  60. Blue ethernet network cables plugged into a switchHARDWARECornelis raises $205M and launches Active Compute Fabric for AI networkingCornelis said it introduced Active Compute Fabric — an open networking architecture that puts programmable compute inside scale-up and scale-out fabrics for AI and high-performance computing — and announced about $205 million in funding to build the next generation of those products. Qualcomm Technologies will join Cornelis’ CEO on stage at AI Infra Summit to talk about networking’s role in rack-scale AI.
  61. Las Vegas Strip at night with bright casino lightsHARDWAREWaymo opens public robotaxi rides in Las VegasAlphabet-owned Waymo said it is welcoming its first public riders for fully autonomous trips in Las Vegas, its first market in Nevada. Early service uses dozens of Ojai minivans across a Strip-area zone; access starts invite-only and will widen across Clark County.
  62. Close-up of a computer circuit board with microchips and gold tracesHARDWARECognichip launches ACI Enterprise, physics-informed AI for chip designRedwood City startup Cognichip said its ACI Enterprise system — Artificial Chip Intelligence built on physics-informed models for semiconductor design — is now commercially available, naming customers including Renesas and SiTime, and claiming that in one unnamed top-20 fabless benchmark a single engineer moved a 55-page specification through micro-architecture, RTL, verification, and power-performance-area work in days instead of months; Cognichip says it has raised $93 million in total.
  63. Universal Robots collaborative robot arm with controller and teach pendantHARDWAREUniversal Robots unveils Gen 7 cobot platform built for physical AIUniversal Robots — a Teradyne Robotics company and a major maker of collaborative robots, or cobots, meaning robot arms designed to work beside people — unveiled Gen 7, its seventh-generation industrial platform, at the IMTS manufacturing trade show in Chicago, calling the system “out-of-the-box AI-ready,” with three new g-Series arms, a smaller CB7 controller with more compute, an AI-ready tool flange for cameras and sensors, and PolyScope X software for edge-AI apps.
  64. NVIDIA CUDA-Q illustration of stacked quantum and accelerated computing layersHARDWARENVIDIA expands open-source CUDA-Q with a layer for fault-tolerant quantum appsNVIDIA expanded its open-source CUDA-Q platform with CUDA-Q Logical, a software layer for designing and testing applications meant for fault-tolerant quantum computers — machines that use logical qubits, which are error-checked groups of physical qubits, so they can run larger jobs.
  65. A large round silicon wafer with a mirrored surface reflecting overhead lightsHARDWAREASML says AI chip boom isn’t over as customers line up for High NA toolsDutch chip-tool maker ASML told Reuters the AI boom is still running: its roughly $200 million extreme ultraviolet (EUV) machines — the tools that print the tiniest circuits in advanced AI chips — are basically sold out through 2027, and big customers are committing to the next generation of roughly $400 million High NA tools.
  66. FUJITSU-MONAKA processor package next to a Fujitsu MONAKA rack serverHARDWAREFujitsu launches Japan-made MONAKA CPU and servers aimed at sovereign AI inferenceFujitsu said it will start global sales of FUJITSU-MONAKA, a Japan-designed processor, in November 2026, plus Japan-made servers aimed at sovereign AI inference — running already-trained models on hardware buyers can trace.
  67. Close-up of colorful source code on a computer monitorHARDWARENVIDIA opens two early ways to write GPU programs in RustNVIDIA published two early ways to write GPU programs in the Rust language that compile straight to the chip, instead of only calling code written in other languages.
  68. Dark car on a wet road at night with headlights cutting through mistHARDWARENVIDIA blog: robotaxi leaders building on its full-stack platformNVIDIA published that its modular robotaxi pipeline — training, simulation and validation, and in-vehicle compute — is being adopted across commercial-scale robotaxi programs, naming Uber, Waymo, Zoox, Mercedes-Benz, and others.
  69. d-Matrix Raptor compute tray, NVIDIA NVLink switch tray and d-Matrix MGX rack with NVLink FusionHARDWARENVIDIA blog: d-Matrix to connect Raptor XPUs via NVLink FusionNVIDIA published that inference chipmaker d-Matrix will use NVLink Fusion to connect its next-generation Raptor XPUs to NVIDIA’s AI infrastructure platform, including MGX racks and Spectrum-X.
  70. White humanoid robot with a dark visor against a black backgroundHARDWARENVIDIA blog: Skild AI’s S1 learns long robot tasks from one videoNVIDIA published that Skild AI’s S1 robot foundation model can learn previously unseen long tasks from a single video demonstration, using the video as the prompt and without updating its weights.
  71. Blue quadruped robot walking across a bright indoor hallHARDWAREArs hands-on: Unitree Go2 Pro robot dog for about $4,000Ars Technica and Understanding AI’s Timothy B. Lee published a hands-on with a Unitree Go2 Pro quadruped he paid $4,017 for after tariffs and shipping.
  72. Laboratory glassware and work surfaceHARDWAREAnthropic opens a Model Hardware Standard research preview for lab and factory devicesAnthropic opened a gated research preview of the Model Hardware Standard (MHS), a shared specification it says will let AI agents operate programmable lab and factory devices.
  73. Aerial view of a sail-shaped hotel on the GulfHARDWAREUAE revises its 5-GW AI campus after Iranian attacks, Reuters reportsReuters, citing six unnamed sources, reported that the UAE is revising a planned 5-gigawatt AI campus after Iranian attacks, while G42 said work was still progressing.
  74. High-voltage transmission towers at sunriseHARDWAREGoogle puts €13 billion into Finnish AI infrastructureGoogle said it will invest at least €13 billion in Finnish digital infrastructure over 2027–2028, calling it the company’s largest single investment in Europe.