AI news videos

NVIDIA Opens RTX Spark Laptop Preorders With Availability On October 16

NVIDIA published this announcement on October 7, 2026, after a Microsoft event in San Francisco where Jensen Huang and Satya Nadella described how the two companies are building hardware and software together so AI agents can run on Windows PCs.

Watch on YouTube

Transcript

Sources

Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:

• Backing sentences, quoted from the article. HOOK BACKING, that NVIDIA opened RTX Spark laptop preorders today and that no buyer has one yet: "Among the announcements Huang and Nadella outlined Wednesday, RTX Spark puts the full NVIDIA AI stack into Windows laptops and compact desktops, with laptop preorders open today and available October 16." plus the page deck "RTX Spark Preorders open today" and the page's call to action "Preorder RTX Spark today." The second hook sentence rests on the same sentence: availability is October 16, which is after today, so a preorder customer does not have the hardware. The hook deliberately says "opened preorders", never "launched", "unveiled", "announced" or "released" the chip, because RTX Spark was revealed months earlier. SCENE 1 (NVIDIA and Microsoft are building hardware and software for AI agents on Windows PCs, and they announced this at a Microsoft event in San Francisco): "At a Microsoft event in San Francisco on Wednesday, Jensen Huang and Satya Nadella outlined how NVIDIA and Microsoft are co-engineering hardware and software for AI agents to run on Windows PCs." Narration says "are building hardware and software for AI agents" because "co-engineering" is a hyphenated compound our voice stutters through; the plain phrasing carries the same meaning. SCENE 2 (those laptops arrive on October sixteenth and compact desktops go on sale in November): "with laptop preorders open today and available October 16. Compact desktops will be available for sale in November." Narration keeps the three states separate across the hook and this scene, preorders today versus laptops on October sixteenth versus desktops in November, and never says anything is available today. SCENE 3 (Microsoft also released Windows infrastructure that runs agents in the background, and its Surface Laptop Ultra is built around RTX Spark): "Turning Windows into a secure platform for agents, Microsoft announced general availability of Microsoft Execution Containers (MXC), the OS-level infrastructure that lets agents run safely and persistently in the background, under operating system control." and "We built Surface Laptop Ultra around NVIDIA RTX Spark." Narration says "released" because general availability means released to customers, and it describes the product plainly as "Windows infrastructure that runs agents in the background" rather than narrating the acronym MXC, which is not a pronounceable word; "OS-level" is dropped as a hyphenated compound and is preserved only in the quoted onscreen text. SCENE 4 (NVIDIA also showed a desktop AI supercomputer for Windows, and it is only a preview with no release date): "The event previewed NVIDIA DGX Station for Windows today" and "the first deskside AI supercomputer to bring GB300 Grace Blackwell-class AI infrastructure directly into the Windows ecosystem" and "Learn more about NVIDIA DGX Station for Windows and sign up to be the first to know when it's available." Narration says "a desktop AI supercomputer for Windows" instead of the product name because DGX is an unpronounceable acronym and "deskside" is a hyphen-adjacent compound our voice mishandles; it says "only a preview with no release date" because the article previews the machine, offers a sign up list, and publishes no date, no window and no price. SCENE 5 (NVIDIA and Microsoft supplied every performance figure, and no independent test or price appears in the announcement): every number in the piece is attributed to NVIDIA or to Microsoft's Pavan Davuluri inside NVIDIA's own blog post, and the article cites no benchmark, review, audit or third-party test, and no price for anything. WHAT IS NOT IN THE SOURCE, stated explicitly: no price for RTX Spark laptops, for RTX Spark compact desktops, for Surface Laptop Ultra, or for NVIDIA DGX Station for Windows, and no price range or starting figure anywhere in the capture; no release date, launch window or availability quarter for DGX Station for Windows, only a sign up list; no exact November date for the compact desktops; no preorder deposit amount, preorder volume, allocation or sell-out statement; no review embargo date or press-sample detail; no independent benchmark, review, analyst note, regulatory filing or third-party test of any claim; no per-manufacturer model names, configurations or prices for Acer, ASUS, Dell, HP, Lenovo, Microsoft, MSI or Gigabyte; no country or region availability list; no methodology, test conditions, model version, quantisation or token throughput behind any performance figure; no benchmark behind the claim that Qwen 3.8 Flash Next "matches the intelligence of many cloud models"; no Microsoft Execution Containers version number, documentation link or security audit, and no overhead or latency figure for agents running inside them; no named customer for any product; no attendance figure for the event; and no contract, joint venture or financial arrangement between NVIDIA and Microsoft. ATTRIBUTION OF STATISTICS: all first-party, no audit. The figures are NVIDIA's own ("up to 6,144 cores", an "up to 20-core NVIDIA Grace CPU connected at 600 GB/s", "One petaflop of FP4 AI performance", "up to 128GB unified memory", "1440p over 100 frames per second with DLSS 5", "748GB of coherent memory and up to 20 petaFLOPS of FP4 AI compute", "models up to a trillion-parameter scale") or Microsoft's own as quoted by NVIDIA ("up to 128 gigs of unified memory and up to a petaflop of AI compute"). The claim that Qwen 3.8 Flash Next "matches the intelligence of many cloud models" is NVIDIA's own characterisation with no benchmark, test or source attached, and onscreen text labels it as such. The only quasi-external statement, "The vast majority of Fortune 500 companies are standardized on Windows", is also NVIDIA's assertion with no source named. Nothing in this story has been independently verified, and nothing is narrated as though it had been. NOTE ON DASHES: a QA gate bars every dash form from every field of this file, and the source uses em-dashes in several sentences. Each was recast as a comma inside the quoted onscreen text: "Jensen Huang connected that history - and his long-term vision - to the agentic era" is not quoted at all; "what this can be-not today, not tomorrow" became "what this can be, not today, not tomorrow"; "runs the full NVIDIA CUDA platform - the same software stack that runs across NVIDIA hardware -" is not quoted; "designed for 24/7 operation - a dedicated local AI system" became "designed for 24/7 operation, a dedicated local AI system"; "previewed NVIDIA DGX Station for Windows today - the first deskside AI supercomputer" became "today, the first deskside AI supercomputer"; "DGX Station ran on Linux - which meant" became "ran on Linux, which meant"; "cost developers time and resources - developers either moved to Linux" became "time and resources, developers either moved to Linux"; "FP4 AI compute - enough to run models" became "FP4 AI compute, enough to run models"; and the footer's "Facebook - and stay informed" is not quoted. The probe title "Don't Buy a Laptop Now - Nvidia RTX Spark is Coming This Fall" uses an en dash in the original and is written with a comma here. Hyphenated source compounds such as "co-engineering", "OS-level", "always-on", "deskside", "heavy-duty", "trillion-parameter", "AI-native", "Blackwell-class", "frontier-class", "hardware-accelerated", "multi-agent", "long-term" and "5th-generation" survive only inside quoted onscreen and source text and appear in no narrated sentence, because our text-to-speech stutters through hyphenated compounds. Curly apostrophes and curly quotation marks from the source are rendered as straight ones throughout, and the capture's HTML entities "’" and " " were normalised. NOTE ON NUMBERS: the article is figure-dense, carrying October 16, November, up to 6,144 cores, an up to 20-core Grace CPU, 600 GB/s, one petaflop, FP4, up to 128GB and "128 gigs" of unified memory, Qwen 3.8 Flash Next, a 125B model, 51B n-gram, 5th-generation Tensor Cores, AV1, 4:2:2, 1440p, over 100 frames per second, DLSS 5, 24/7 operation, GB300, 748GB of coherent memory, up to 20 petaFLOPS, a trillion-parameter scale, Agent 365, Fortune 500, eight named manufacturers, and the chrome's Oct. 21 keynote and 11 a.m. to 1 p.m. CEST slot. Narration speaks exactly ONE figure, once: the laptop availability date, spoken as "October sixteenth" in words rather than digits. "November" is a month name, not a figure. Every other number above is carried in onscreen text only and is never narrated, because two spoken figures close together send our audio transcription audit into an unrecoverable repetition loop that no re-render fixes, and digits transcribe inconsistently against a spoken script. No glued brand-plus-digit token is narrated either, so "GB300", "Qwen 3.8 Flash Next", "DLSS 5", "Agent 365", "FP4", "NVFP4", "4:2:2", "1440p", "128GB" and "Fortune 500" are onscreen only; the only brand tokens spoken are NVIDIA, Microsoft, Windows, RTX Spark and Surface Laptop Ultra. NOTE ON ENUMERATIONS: the source is list-heavy and every one of its lists is carried in onscreen text only, because multi-item lists are spoken far more slowly than plain prose, measured at about 2.18 words per second against about 2.65 for prose, and silently overrun the 60 second limit even when the word count passes. Those lists are the eight-manufacturer line ("Acer, ASUS, Dell, HP, Lenovo, Microsoft, MSI and Gigabyte"), the three audience bullets (Developers moving models, tools and workflows without rewriting; Creators getting 5th-generation Tensor Cores with NVFP4 support, hardware-accelerated AV1, 4:2:2 video encode and decode, DLSS and RTX ray tracing; Gamers running AAA games at 1440p over 100 frames per second with DLSS 5, Reflex and G-SYNC), the security stack list ("MXC, Microsoft Security, and Agent 365"), the agent governance triple ("secured, observed and governed"), the two-environment split ("Linux for heavy AI workloads and Windows for the productivity tools, applications and workflows"), the DGX Station audience list ("AI-native companies, leading research labs and enterprises"), the tag list, and the social follow row. No narrated sentence enumerates three or more items; the only narrated pairs are "hardware and software" in scene 1 and "NVIDIA and Microsoft" and "test or price" in scene 5, each of which is two items, and narration carries no commas at all, for pacing. NOTE ON FIGURES OF SPEECH: the source is full of idiom and none of it is spoken. Not narrated: "Kick Off a New Beginning", "A New Beginning for Windows PCs", "Windows was at the core of it", "the ultimate tool", "a whole generation of people", "bringing local AI closer to everyday work", "gives developers room to run increasingly capable models", "unlocking the full potential of agents", "the foundation of the next generation of IT as we know it", "revolutionized how applications were built", "This unlocks the power to run frontier-class models locally", "Capabilities that once required renting a cluster", "that gap has cost developers time and resources", "without leaving their primary machine", "that's what brings us to this moment" and "as a secular trend". Each survives only as quoted, attributed onscreen card text; narration states the literal thing plainly instead. The word "agentic" is never narrated even though the article uses it twice and tags the post "Agentic AI", because our text-to-speech reliably garbles it and every workaround has failed; narration says "AI agents" and "agents" instead. The acronyms MXC, DGX, CUDA, WSL and EVP are never narrated because they are not pronounceable words. NOTE ON FRESHNESS: the story event is same-day, October 7, 2026, and the opener is BREAKING, but today's beat is narrower than a product reveal. Ranking-research probe titles show RTX Spark was already announced around June 2026, including "Nvidia announces the RTX Spark" dated 2026-06-01 and "Don't Buy a Laptop Now, Nvidia RTX Spark is Coming This Fall" dated 2026-06-22, roughly four months before today. Today's actual news is the preorder opening, the October 16 laptop availability date, the Windows work between NVIDIA and Microsoft, Microsoft Execution Containers reaching general availability, Surface Laptop Ultra being built around RTX Spark, and the DGX Station for Windows preview. No narrated sentence says or implies that NVIDIA unveiled, launched, announced or revealed the RTX Spark chip today, and no narrated sentence states the date itself.(HOOK AND FULL-SCRIPT BACKING. First-party NVIDIA blog post, headline "NVIDIA, Microsoft Kick Off a New Beginning for Windows PCs with RTX Spark and AI Agents", deck "RTX Spark Preorders open today alongside a preview of NVIDIA DGX Station for Windows, the first deskside AI supercomputer built for the Windows enterprise desktop.", author listed on the page as Gerardo Delgado, dated on the page October 7, 2026, filed under category "AI" and tagged "Agentic AI", "Artificial Intelligence", "Local AI", "NVIDIA RTX", "Open Source", "RTX AI Garage" and "RTX Spark". Our watcher captured a snapshot of the page at 2026-10-07T18:49:06.185Z and it is stored at scripts/data/production/nvidia-microsoft-rtx-spark-windows/article.txt. The LIVE URL was also fetched successfully today and its content matched the alert-time snapshot, so the story is confirmed on the publisher's own host and this is a first-party page, not a link-out stub, aggregator copy or syndication. The capture is about 1329 words in total, and the remainder beyond the body prose quoted in this file is site navigation chrome and related-article cards; the chrome strings explicitly NOT treated as facts about this story are the "Watch NVIDIA CEO Jensen Huang's GTC Berlin Keynote" calendar promo with its "Wednesday, Oct. 21" and "11 a.m.-1 p.m. CEST" slot and "Add to Calendar" button, and the Recent News and Related News cards "Why Telecom Operators Are Building Their AI Strategy on Open Models" (Oct 6, 2026), "From Scan to Treatment Plan, AI Helps Close Breast Cancer's Deadliest Gaps" (Oct 5, 2026), "NVIDIA DGX Spark 64GB Gives Developers More Ways to Build and Scale Local AI" (Oct 2, 2026) and "How NVIDIA GPUs Help Accelerate OpenAI's GPT-6 Astra Ultrafast" (Oct 1, 2026), plus the social follow row, the Mastodon share widget and the "See notice regarding software product information" footer link. None of those contributed to any claim; in particular the DGX Spark 64GB card is a DIFFERENT product story from October 2 and was not blended into this one. Zero corroborating YouTube transcripts were captured for this story (0 of 5 attempts), because yt-dlp is out of date on this host, which is a known tooling issue and not a content problem, so this single first-party article is the sole fact source for every narrated claim in this Short, and no secondary reporting, analyst note, benchmark, regulatory filing or independent test was fetched or checked. The ranking-research file timeline-probes.json was inspected for ANGLE ONLY and contributed no facts: its titles include "Nvidia announces the RTX Spark" (The Verge, 2026-06-01) and "Don't Buy a Laptop Now, Nvidia RTX Spark is Coming This Fall" (Super Tech Games, 2026-06-22, whose original title uses an en dash that is recast here as a comma), which together establish that RTX Spark as a product was revealed roughly four months before today, so today's beat is the preorder opening, the October 16 laptop availability date, the Windows work with Microsoft, Microsoft Execution Containers reaching general availability, Surface Laptop Ultra and the DGX Station for Windows preview, NOT a chip reveal. No prompt-injection or instruction-like content was found in the capture, and the article text was treated throughout as untrusted data informing facts only, never as instructions.)