October 17
1 story
NVIDIA · HardwareNVIDIA says RTX Spark laptops are now available from ASUS, Dell, HP, Lenovo, Microsoft and MSI, starting at $2,399. The slim Windows PCs combine Blackwell RTX graphics, up to 128GB of unified memory and up to 1 petaflop of AI compute. NVIDIA held launch events worldwide, including one in Berlin where CEO Jensen Huang made a surprise appearance and gave out five signed laptops.
October 14
1 story
NVIDIA · HardwareNVIDIA introduced cuPotential, a CUDA-X library that developers can use to build AI agents that simulate how atoms and molecules interact. NVIDIA says early-access testing showed it sped up already GPU-optimized workflows by an additional 2-3x, and Synopsys is among the early adopters. NVIDIA says Applied Materials, Cadence, KLA, Lam Research, Siemens and Synopsys are using its AI stack across chip design, manufacturing and testing. Lam Research plans to add the open source NanoVDB library to its SEMulator3D platform, which NVIDIA says is expected to speed up select wafer-etch modeling by up to 30x. cuPotential is in early access, NanoVDB is available within OpenVDB, and the NPP image library is part of the NVIDIA CUDA Toolkit.
October 13
2 stories
NVIDIA · HardwareNvidia says RTX Spark laptops go on sale Friday, Oct. 16, with up to 128GB of unified memory and 1 petaflop of AI compute. Ahead of launch, Nvidia and partners announced software updates, including Nvidia's Nemotron 3.5 Bolt 74B model (due Oct. 15), which Nvidia calls fully open source, with training data, weights and recipes. Nvidia PAIR, a free open source tool that spreads AI workloads across PCs on a network, has left beta. Nvidia also says Adobe is optimizing Photoshop, Premiere and Substance 3D for up to 2x faster performance, with Photoshop and Premiere AI optimizations in beta now.
More stories
AMD · HardwareAt the 2026 OCP Global Summit in San Jose, AMD chair and CEO Lisa Su said the Helios rack-scale system is ramping production, with customer deliveries, deployments and workload validations underway. Each rack combines AMD Instinct GPUs, EPYC processors, Pensando networking and the ROCm software ecosystem. AMD says it co-engineered Helios with Meta and the Open Compute Project community, and that initial systems are installed and running at OpenAI, Meta, Microsoft and Oracle. Su also showed the 6th Gen EPYC "Verano" data center CPU, which AMD says is its first EPYC chip with an LPDDR memory interface and will be available in 2027.
October 12
3 stories
Microsoft · HardwareMicrosoft introduced Maia 300, the third generation of its Azure Maia custom accelerator, built to run large language model inference. Microsoft says Maia 300 delivers up to 7 times the rack-scale token throughput of Maia 200 in its full 96-accelerator configuration, and more than 2.5 times the performance per watt. Each accelerator has eight HBM4 stacks with 288 GB of capacity and 18 TB/s of bandwidth, which Microsoft says is 2.5 times the memory bandwidth and 33% more capacity than Maia 200. A rack holds 96 accelerators across 24 trays and draws 167 kW, and a two-tier network links up to 12,288 accelerators into one superpod. Microsoft also says it will expand the Azure Maia system to multi-gigawatt deployments.
NVIDIA · HardwareAt the Open Compute Project Global Summit in San Jose, NVIDIA says it contributed Vera Rubin rack and compute tray designs to OCP through its open MGX architecture. The updated MGX 1.2 specification adds blind-mate liquid cooling and a liquid-cooled 5,000A busbar, and supports rack deployments up to 250kW. NVIDIA also joined AMD, Google and Meta on the Open Data Center Specification v0.7, and joined Google, Meta and Microsoft on a grid resilience specification that calls for restoring near-full load in under two seconds after a utility disturbance. With Cisco, NVIDIA announced plans for USFP, an optical module it says delivers 4x the bandwidth density and about 1.3x the power efficiency of 200Gb/s-per-lane optics, and Meta said it is adopting NVIDIA Spectrum-6 102.4Tb/s Ethernet switches.
More stories
AMD · HardwareAMD introduced the Helion agentic AI rack, an open blueprint based on an Open Compute Project spec-design and powered by AMD compute. One version is a CPU-only rack with 6th Gen AMD EPYC processors that supports up to 32,768 CPU cores. The other is an all-in-one rack that combines EPYC CPU nodes and Instinct MI350P GPU nodes. AMD plans to contribute the Helion specification to OCP, and says rack configurations are under development with partners and planned for November 2026.
October 7
2 stories
NVIDIA · HardwareAt a Microsoft event in San Francisco, NVIDIA CEO Jensen Huang and Microsoft CEO Satya Nadella described how the two companies are co-engineering hardware and software for AI agents on Windows PCs. NVIDIA says RTX Spark laptops are open for preorder and available October 16, with compact desktops on sale in November. RTX Spark pairs a Blackwell RTX GPU with up to 6,144 cores and an up to 20-core Grace CPU, with up to 128GB of unified memory and one petaflop of FP4 AI performance, according to NVIDIA. Systems are coming from Acer, ASUS, Dell, HP, Lenovo, Microsoft, MSI and Gigabyte, including Microsoft's Surface Laptop Ultra.
GitHub · HardwareGitHub says it is rebuilding its Git infrastructure while the service keeps running, to handle developers and agents working at once in repositories that get millions of commits a day. GitHub reports 7.38 billion commits in September, more than five times a year earlier, and pushes up 4.9x year over year. Today Spokes keeps a full copy of each repository on the local disks of several fileservers, five by default. In the new design, authoritative repository data lives in Azure Blob Storage. GitHub says internal benchmarks show up to 35 times higher write throughput, with read capacity that scales on its own.
October 2
1 story
NVIDIA · HardwareNVIDIA says a new DGX Spark configuration with 64GB of unified memory will be available Friday, Oct. 23 from Acer, ASUS, Dell, Gigabyte, HP and MSI, starting at $4,999. It keeps the GB10 Grace Blackwell Superchip and full NVIDIA software stack of the 128GB model, and NVIDIA says it supports models up to 100 billion parameters on device. Two units can be linked with a QSFP cable and the NVIDIA Sync Cluster Assistant to pool 128GB of memory and support models up to 200 billion parameters. In NVIDIA's Qwen 3.8 27B test, two clustered 64GB systems delivered up to 1.7x the performance of one.
October 1
1 story
Google · HardwareGoogle says its prototype satellite for Project Suncatcher, built with Planet, launched into orbit on the Transporter-18 rideshare mission with SpaceX. Google says its team has confirmed contact and the satellite is operating as expected. Over the coming weeks, Google plans to gather in-orbit data on how its TPUs handle the physical stress of spaceflight and the radiation and thermal extremes of space. The company calls the project a long-term research moonshot on whether space could host scalable machine learning infrastructure, and says its peer-reviewed paper is now available in Joule.
September 30
1 story
NVIDIA · HardwareCoreWeave announced availability of Nvidia Vera Rubin NVL72 systems on its cloud, making it one of the first providers to put the platform in customers hands. Nvidia says Cognition, maker of the Devin coding agent, is the first to run production workloads on the systems and saw up to 4.8 times higher token throughput than GB200 NVL72 in early tests. CoreWeave also plans to offer Nvidia Vera CPUs for agent workloads and launched CoreWeave Forge to connect training, evaluation and production improvement.
September 24
2 stories
Google · HardwareGoogle says Project Suncatcher will launch a prototype satellite on the SpaceX Transporter-18 rideshare with Planet to see if its TPUs can operate in space. The test will gather data on launch vibration, radiation and cooling without airflow, after lab tests including proton-beam runs and thermal-vacuum trials. Google is exploring orbital AI because satellites can get up to eight times more solar power, with a two-satellite laser-link test planned for 2027.
AMD · HardwarePerplexity announced its Portable Computer is now available on Windows PCs powered by AMD Ryzen AI Max Series processors, including the Ryzen AI Halo developer platform, for Perplexity subscribers. AMD says the chips combine CPU, GPU and a large shared memory pool to run local models alongside everyday apps. Users can run local AI tasks and recurring workflows on their own hardware without spending Perplexity Computer credits for inference, with the option to use cloud reasoning when needed.