NVIDIA is taking the RTX Spark launch out of the keynote hall and onto the street. The company confirmed it will host a public “RTX Spark Launch Celebration” in Berlin on Friday, October 16, 2026, the same day the first RTX Spark laptops begin shipping to retail. The event was first reported by TechPowerUp, with corroborating coverage from ComputerBase, Gamestar, Eurogamer Germany and GamesMarket. Unlike NVIDIA’s typical invite-only press briefings, this one is open to anyone 18 or older who registers in advance, and it doubles as a hands-on showcase for the Grace-Blackwell chip NVIDIA is betting will redefine what a Windows laptop can do with AI running entirely on-device.
The timing is not an accident. RTX Spark preorders opened October 7, 2026, and NVIDIA used that same day to stage a joint announcement with Microsoft in San Francisco, where CEO Jensen Huang and Microsoft CEO Satya Nadella held a fireside chat about what the companies are calling a “new beginning” for Windows PCs built around local AI agents. Berlin closes the loop: a street-level party timed to the exact day consumers can walk into a store and buy the hardware the two executives spent weeks previewing.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
NVIDIA Opens the Doors to the Public for RTX Spark’s Berlin Debut
According to the reporting, the party takes place at XPERION, the gaming and esports venue tucked inside the Saturn electronics store at Alexanderplatz 3 in central Berlin. The queue opens at 18:00 local time, with doors at 18:45. Entry is free, but attendance is not guaranteed on a walk-up basis: NVIDIA is requiring advance registration through the event-ticketing platform Luma, and capacity is explicitly described as limited, particularly for the stage program. The age cutoff is 18, which lines up with the goodie bags that include a GeForce NOW subscription and likely reflects standard venue and alcohol-service policy for an evening event rather than any content restriction tied to RTX Spark itself.
Choosing a retail electronics store in the middle of Alexanderplatz, one of Berlin’s busiest public squares, is itself a statement. NVIDIA’s past hardware reveals have mostly lived inside convention centers (GTC) or closed studio broadcasts. Putting an RTX Spark celebration in a shopping district, with registration open to the general public rather than press and partners only, signals that NVIDIA wants this launch read as a consumer event, not just an enterprise AI story. That matters for a company whose stock and reputation over the past two years have been built overwhelmingly on data-center GPU demand rather than retail PC hardware.
What’s on the Program: Creators, CD Projekt Red, and Demo Stations
Coverage of the event describes a night built around three elements: a stage program, hands-on demo stations, and a laptop raffle. On stage, NVIDIA has booked creators AlexiBexi, FalcoPunch, and Alex WÃĪtzel, alongside a talk from CD Projekt Red, the Polish studio behind The Witcher 3 and Cyberpunk 2077. Pairing an AI-PC hardware launch with a major games studio is a deliberate move: NVIDIA’s RTX Spark pitch leans heavily on local generative AI, but the company still needs gamers in the room, and CD Projekt Red’s back catalog is a reliable showcase for ray tracing and DLSS upgrades on new silicon.
The demo stations are where attendees are expected to get real hands-on time with early RTX Spark laptops and compact desktops from the first wave of OEM partners, testing both gaming performance and the on-device AI workloads NVIDIA has been promoting since the October 7 San Francisco announcement. That announcement, detailed on NVIDIA’s own blog, emphasized running large language models locally without a cloud connection, a point the Berlin demo stations are built to prove in person rather than in a slide deck.
Five Jensen Huang-Signed Laptops and a Green Berlin TV Tower
The giveaway is the headline hook for anyone deciding whether registering is worth the trip. NVIDIA is raffling off five ASUS ProArt P14 laptops personally signed by Jensen Huang, a collector’s-item angle the company has used sparingly and mostly at flagship events like GTC. Every attendee, win or lose, walks away with a goodie bag that includes a six-month GeForce NOW Ultimate membership, which NVIDIA reportedly values at more than âŽ130, plus food and drinks covered for the night, while supplies last.
NVIDIA is also reportedly lighting up Berlin’s iconic Fernsehturm, the TV Tower that dominates the city’s skyline near Alexanderplatz, in NVIDIA’s signature green for the occasion. It is a relatively low-cost, high-visibility marketing move, the kind of gesture that turns a product launch into a landmark photo opportunity for a city whose skyline rarely gets commercially rebranded. Combined with the free entry and public registration, the full package functions less like a traditional press launch and more like a one-night consumer activation designed to generate social coverage well beyond the hardware press that would normally write about a new NVIDIA chip platform.
What RTX Spark Actually Is, and Why NVIDIA Is Throwing It a Party
RTX Spark is NVIDIA’s platform for slim Windows laptops and small-form-factor desktops built on an Arm-based NVIDIA Grace CPU paired with a Blackwell RTX GPU, a combination internally referred to during development as N1X. Per NVIDIA’s official RTX Spark product page and the October 7 NVIDIA Newsroom announcement, the flagship configuration pairs an up-to-20-core Grace CPU with an up-to-6,144-core Blackwell RTX GPU, connected over a 600 GB/s link, with as much as 128GB of unified LPDDR5X memory shared between CPU and GPU. NVIDIA rates the top configuration at roughly 1 petaflop of FP4 AI performance, enabled by fifth-generation Tensor Cores with native NVFP4 support, plus fourth-generation RT cores, DLSS 5, hardware-accelerated AV1 encode/decode, and full support for NVIDIA’s CUDA software stack.
That unified-memory design is the real pitch. Rather than splitting a fixed amount of VRAM from system RAM the way a conventional gaming laptop does, RTX Spark lets large AI models claim most of that shared 128GB pool when needed. NVIDIA’s own materials point to running models in the range of 125 billion parameters locally, citing an example like Qwen 3.8 Flash Next with a 51-billion n-gram component, entirely on-device and without a cloud API call. On the gaming side, NVIDIA claims AAA titles can hit 1440p at 100-plus frames per second with DLSS 5 enabled, a figure our earlier RTX Spark gaming performance analysis found still trails a discrete RTX 5090 desktop card, which is expected given RTX Spark targets efficiency and portability rather than peak discrete-GPU throughput.
| Event Detail | Information |
|---|---|
| Event name | RTX Spark Launch Celebration |
| Date | Friday, October 16, 2026 |
| Venue | XPERION, inside Saturn, Alexanderplatz 3, Berlin |
| Queue opens | 18:00 local time |
| Doors open | 18:45 local time |
| Entry cost | Free, registration required via Luma |
| Age restriction | 18+ |
| Capacity | Limited, especially for the stage program |
| Stage guests | AlexiBexi, FalcoPunch, Alex WÃĪtzel; talk from CD Projekt Red |
| Main giveaway | Five ASUS ProArt P14 laptops signed by Jensen Huang |
| Goodie bag | 6-month GeForce NOW Ultimate membership (valued over âŽ130) |
| Extra touch | Berlin TV Tower lit NVIDIA green |
From Preorder to Retail: RTX Spark’s Rollout Timeline
RTX Spark’s path to market has moved fast by NVIDIA’s own standards for a brand-new PC platform. Preorders for RTX Spark laptops opened October 7, 2026, immediately after the Huang-Nadella fireside chat in San Francisco. Retail availability for laptops follows just nine days later, on October 16, the same date as the Berlin party, meaning the event functions as both a launch celebration and a de facto retail availability announcement. NVIDIA has said compact RTX Spark desktops will follow in November 2026, giving the company a second wave of coverage and store traffic once the initial laptop cycle settles.
That compressed nine-day gap between preorder and retail is unusually tight for a platform this new, and it echoes a pattern NVIDIA set earlier in the year with the DGX Spark line. As we covered when NVIDIA’s DGX Spark 64GB launched at $4,999 amid a broader memory-pricing surge, the company has leaned on fast preorder-to-ship cycles across its Spark-branded hardware to capture demand before component costs climb further, a dynamic that remains live given continuing DRAM and LPDDR5X price pressure across the industry.
The OEM Lineup Behind RTX Spark
NVIDIA’s own announcement lists Acer, ASUS, Dell, HP, Lenovo, Microsoft, MSI, and Gigabyte as RTX Spark partners, though not all eight are shipping at once. The initial retail wave centers on ASUS, Dell, HP, Lenovo, Microsoft, and MSI, with Acer and Gigabyte following afterward. That is a notably broad partner list for a first-generation Arm-based Windows platform, wider than Qualcomm’s initial Snapdragon X rollout and a sign NVIDIA pushed hard to line up simultaneous launch partners rather than stagger them one OEM at a time.
Individual OEM configurations vary widely in price and spec. We’ve tracked several of these separately: Dell’s XPS 16 RTX Spark Creator Edition launched at $3,799 with an 18-core configuration and a 99Wh battery, while leaked pricing pegged HP’s RTX Spark laptop at $2,999. Microsoft’s own entry builds RTX Spark into the Surface Laptop Ultra, and the company separately shipped a developer-focused Surface RTX Spark Dev Box at $5,999 with 128GB of RAM, aimed squarely at AI engineers rather than consumers. The spread between a roughly $2,599 entry laptop and a near-$7,000 maxed-out workstation shows RTX Spark is being positioned as a tiered platform, not a single SKU.
| RTX Spark Milestone / Spec | Detail |
|---|---|
| Platform codename | N1X (Grace CPU + Blackwell RTX GPU) |
| CPU | Up to 20-core NVIDIA Grace CPU |
| GPU | Up to 6,144-core Blackwell RTX GPU |
| CPU-GPU link | 600 GB/s |
| Unified memory | Up to 128GB LPDDR5X |
| AI performance | Up to 1 petaflop FP4 |
| Tensor cores | 5th-gen with native NVFP4 support |
| TDP range | 45W to 80W |
| Preorders opened | October 7, 2026 |
| Laptop retail launch | October 16, 2026 |
| Compact desktops | November 2026 |
| First-wave OEMs | ASUS, Dell, HP, Lenovo, Microsoft, MSI (Acer, Gigabyte to follow) |
Pricing Is Still Finding Its Footing
Even as preorders opened and the Berlin event date locked in, pricing clarity lagged behind the specs reveal. TechRadar’s coverage of the October 7 announcement noted that detailed pricing was still largely unavailable at the time NVIDIA confirmed RTX Spark would not throttle performance when running on battery, a notable claim for an Arm-based chip aimed at thin-and-light laptops. Pricing has since filled in OEM by OEM rather than through a single NVIDIA-set MSRP: entry configurations with an 18-core CPU, 24GB of unified memory, and 512GB of storage have landed around $2,599, while fully loaded 20-core, 128GB, 2TB configurations, such as a top-end ASUS ProArt model, run as high as roughly $6,999.
That OEM-by-OEM pricing approach mirrors how NVIDIA has generally let partners set final consumer pricing on GPU-based laptops for years, but it is a departure from the single-price-point clarity Apple offers with its own silicon lineup. For shoppers comparing RTX Spark machines across six or more brands, the practical research burden falls on configuration-by-configuration comparison rather than one simple price tier, which is likely why NVIDIA wants hands-on demo stations in Berlin: a shopper who can physically compare a $2,599 entry unit against a $6,999 workstation on the same night has a much easier time justifying the spread than one reading spec sheets online.
Why Berlin? NVIDIA’s European AI-PC Push
Germany is Europe’s largest PC and gaming hardware market, and Berlin specifically has become a recurring stage for NVIDIA’s European consumer outreach, from gaming showcases to creator-economy marketing tied to GeForce NOW. Staging the RTX Spark celebration there, rather than in the U.S. where the platform was first announced alongside Microsoft, suggests NVIDIA wants its AI-PC pitch to land with European consumers and press on a dedicated timeline rather than as an afterthought to the San Francisco news cycle. It also puts RTX Spark in front of German outlets like ComputerBase, Gamestar, and Hardwareluxx, publications with deep enthusiast-PC readership that NVIDIA has historically courted for GPU reviews.
There is also a regulatory and competitive subtext. The European Union has been increasingly active in scrutinizing how AI capabilities are marketed and priced to consumers, and a hands-on public event, where ordinary attendees can test the local-AI claims themselves rather than take NVIDIA’s word for it, is a relatively low-risk way to build consumer trust in a market where AI-performance marketing claims face more scrutiny than they historically have in the U.S.
Microsoft’s Deep Stake in RTX Spark’s Success
RTX Spark is not purely an NVIDIA story. Microsoft co-announced the platform on October 7 and built its own Surface Laptop Ultra around RTX Spark silicon, which we detailed when we covered the Surface Laptop Ultra’s $2,599 starting price and local-LLM capabilities. Microsoft is pairing the hardware with software infrastructure called Microsoft Execution Containers, designed to let AI agents operate securely on-device, and tying that into Agent 365, Microsoft’s broader push to make autonomous AI agents a standard part of the Windows and Microsoft 365 experience.
For Microsoft, RTX Spark is a hardware proof point for a software thesis: that the next phase of Windows differentiation is agentic AI that runs locally rather than purely through cloud subscriptions like Copilot. For NVIDIA, Microsoft’s backing gives RTX Spark immediate legitimacy in enterprise IT departments that would otherwise be slower to adopt a brand-new Arm-based Windows platform. The Huang-Nadella fireside chat was as much about cementing that mutual dependency publicly as it was about specs.
RTX Spark vs. Mac Studio, Snapdragon X2, and the Rest of the AI-PC Field
RTX Spark’s closest conceptual rival is Apple’s Mac Studio, which has built a strong reputation among AI developers and creators for running large local models thanks to Apple Silicon’s own unified-memory architecture. RTX Spark’s pitch against Mac Studio rests on three claims: native CUDA compatibility (which Apple Silicon lacks), a comparable or larger top-end unified memory pool at 128GB, and access to the broader Windows software and gaming ecosystem that macOS does not offer. Apple, in turn, retains the advantage of tighter hardware-software integration and a multi-year head start in optimizing local AI inference on its own chips.
Within the Windows-on-Arm category, RTX Spark also competes indirectly with Qualcomm’s Snapdragon X family, which Microsoft itself uses elsewhere in the Surface lineup, as covered in our report on Surface Pro’s Snapdragon X2 Plus launch at $1,149. Snapdragon X2 Plus targets battery life and thin-and-light efficiency at a lower price point, while RTX Spark leans entirely on raw AI and graphics throughput, positioning it as a premium tier above mainstream Windows-on-Arm hardware rather than a direct price competitor. AMD’s Ryzen AI chips and Intel’s Core Ultra lineup round out the field, both competing more on integrated efficiency than on RTX Spark’s unified-memory AI ceiling.
Historical Context: NVIDIA’s Long Road From Data Center to Desktop
RTX Spark did not appear out of nowhere. NVIDIA has spent the better part of two years building out a “Spark” sub-brand that started in AI developer workstations, most notably the DGX Spark desktop units, before extending into consumer-facing laptops and compact desktops. That evolution mirrors NVIDIA’s broader strategic shift: a company whose stock has been powered almost entirely by data-center GPU demand for Hopper and Blackwell chips is now trying to translate that AI infrastructure story into a reason for ordinary consumers to buy a new laptop, something NVIDIA has not had a central hand in since its early GeForce-era partnerships with OEMs.
It is also a reminder of how fast NVIDIA’s branding has moved. The N1X codename, 6,144-core Blackwell GPU configuration, and Grace CPU pairing were previewed in earlier reporting well before the October 7 announcement made them official, and the jump from internal codename to retail shelf and public street party happened within a matter of months, a pace that is unusually quick even by the fast-moving standards of the current AI hardware cycle.
Market Impact: What a Street Party Signals About AI-PC Demand
The choice to run a public, free, hands-on launch event rather than a closed press briefing says something about where NVIDIA believes the AI-PC category currently stands: still unproven with mainstream consumers, and in need of direct demonstration rather than spec-sheet marketing. Reviewers and enthusiast press already understand petaflops and unified memory; everyday shoppers walking through Alexanderplatz do not, and a hands-on demo station answers “what can this actually do for me” far more effectively than a press release.
That consumer-education burden is the real market risk facing RTX Spark and the AI-PC category broadly. Unified memory architectures and local LLM inference are genuinely new capabilities for a Windows laptop, but they are also hard to market in a 30-second ad. Events like Berlin’s RTX Spark party function as expensive but effective substitutes for that missing consumer education, and if NVIDIA considers this one successful, expect similar city-level activations to follow in other major European and Asian markets as the compact-desktop wave arrives in November.
What to Watch Next: Five Predictions for RTX Spark
- More city-level launch parties. If the Berlin event drives strong registration and local press coverage, NVIDIA is likely to replicate the public-party format in other European or Asian cities ahead of the November compact-desktop rollout.
- Pricing will keep fragmenting by OEM. Expect the spread between entry-level (~$2,599) and maxed-out (~$6,999) configurations to widen further as Acer and Gigabyte join the lineup with their own price tiers.
- Agentic AI becomes the real battleground, not raw specs. With Microsoft’s Agent 365 and Execution Containers tied directly to RTX Spark, the competitive story over the next two quarters will shift from petaflops to which platform’s AI agents are actually useful day to day.
- Compact desktops will broaden RTX Spark’s audience in November. A smaller-form-factor desktop at a lower likely price point than the flagship laptops should pull in buyers who want RTX Spark’s AI capability without paying laptop-grade premiums.
- Collector demand for signed units. The five Jensen Huang-signed ASUS ProArt P14 laptops given away in Berlin are likely to show up in enthusiast communities and collector circles as sought-after memorabilia well beyond their retail value.
Frequently Asked Questions
What is the NVIDIA RTX Spark Launch Celebration?
It’s a public event NVIDIA is hosting in Berlin on October 16, 2026, to mark the retail launch of RTX Spark laptops. It includes a stage program, hands-on hardware demos, a laptop raffle, and free food and drinks.
Where exactly is the event taking place?
At XPERION, a gaming and esports venue inside the Saturn electronics store at Alexanderplatz 3 in central Berlin.
Do I need a ticket to attend?
Entry is free, but NVIDIA requires advance registration through the Luma event platform, and capacity is limited, especially for the stage show. Walk-ins are not guaranteed entry.
Is there an age requirement?
Yes, attendees must be 18 or older.
What can attendees win at the event?
NVIDIA is raffling off five ASUS ProArt P14 laptops signed by CEO Jensen Huang. Every attendee also receives a goodie bag with a six-month GeForce NOW Ultimate membership, valued at more than âŽ130.
What is RTX Spark, in plain terms?
It’s NVIDIA’s platform for thin Windows laptops and small desktops, combining a Grace CPU and a Blackwell RTX GPU with up to 128GB of memory shared between them, built to run AI models locally without needing the cloud, while also handling modern gaming workloads.
When can I actually buy an RTX Spark laptop?
Retail availability for RTX Spark laptops begins October 16, 2026, the same day as the Berlin event. Compact RTX Spark desktops are expected to follow in November 2026.
How much does an RTX Spark laptop cost?
Pricing varies by OEM and configuration. Entry-level models with 24GB of memory have been reported around $2,599, while fully configured 128GB models from brands like ASUS have reached roughly $6,999.
