Search
CoursesStock researchPaper tradingNews
HomeAbout
The white lattice roof of Nvidia's headquarters campus on San Tomas Expressway in Santa Clara, California

Photo: Coolcaesar / Wikimedia Commons, CC BY-SA 4.0, cropped

Microsoft and Nvidia Built a $2,599 Laptop That Runs AI Without the Cloud. Who Wins and Who Loses?

Key facts
  • Microsoft unveiled the Surface Laptop Ultra on October 7, 2026. It starts at $2,599 and goes on sale October 16.
  • It runs on Nvidia's new RTX Spark chip, which supports up to 128GB of memory. That's enough to run large AI models on the laptop itself.
  • A model with a faster chip starts at $3,700. Top versions reach $5,900. A desktop version for developers starts at $6,000.
  • Dell, HP, Lenovo and Asus are building PCs on the same chip. Dell's XPS 16 Creator Edition costs $3,800.
  • The chip is Arm-based, the same family Apple uses. It is not an Intel or AMD x86 chip.

What they announced

On Wednesday in San Francisco, Microsoft CEO Satya Nadella and Nvidia CEO Jensen Huang shared a stage to launch a new kind of Windows computer.

The Surface Laptop Ultra is built on an Nvidia chip called RTX Spark. Its selling point is simple: it can run serious AI models on the laptop, without sending anything to a data centre.

Device Starting price Notes
Surface Laptop Ultra $2,599 On sale October 16
Surface Laptop Ultra, faster chip $3,700 Up to $5,900 fully loaded
Surface RTX Spark Dev Box $6,000 Desktop for software developers
Dell XPS 16 Creator Edition $3,800 Same Nvidia chip

HP, Lenovo and Asus are also building machines on the chip.

Why memory is the headline

AI models are huge files that have to sit in memory while they work. A normal laptop has 16GB. The RTX Spark supports up to 128GB.

That's what lets it run models other laptops can't. Microsoft says the Ultra creates AI images 4.3 times faster and AI video 6.2 times faster than a MacBook Pro with Apple's M5 Pro chip. Those are Microsoft's own tests, so wait for independent reviews.

Microsoft also changed Windows 11. A new feature called Execution Containers walls off AI "agents," the programs that carry out tasks for you, so they can't touch the rest of your computer.

Cloud AI vs local AI

This is where the story reaches investors.

Today: you type a question. It travels to a data centre full of Nvidia chips. A company such as OpenAI or Microsoft pays for that computing, and charges you a subscription or a per-use fee.

With a local model: the laptop does the work. You paid for the hardware once. There's no fee per question, and your files never leave the machine.

Microsoft described a mix of both. Simple jobs run free on the laptop. Hard jobs still go to a powerful, expensive cloud model.

Who could win

Nvidia. It dominates data centre chips. This puts it inside PCs as the main processor, a market it has never led. It gets paid whether AI runs in the cloud or on your desk.

Microsoft. PC sales have been dull for years. A real reason to upgrade helps Windows and Surface. And if routine AI work moves to the customer's own hardware, Microsoft spends less on data centres to serve it.

Memory makers. A laptop with 128GB uses eight times the memory of a normal one. That's more demand for Samsung, SK Hynix and Micron, in a market already short of supply.

PC makers. Dell, HP and Lenovo get to sell $3,000-plus machines.

Who could lose

Intel and AMD. Their x86 chips have powered Windows PCs for 40 years. RTX Spark is Arm-based. If the most profitable laptops switch, that hurts.

Apple. Developers who run AI on their own machines have favoured the MacBook Pro. Microsoft is going straight at them with up to $1,000 off for trading one in.

OpenAI and other cloud AI sellers, a little. If a free local model handles your everyday tasks, that's fewer paid requests. This matters more this week, after a report that OpenAI's revenue is lower than investors believed.

Reasons to stay calm

What it means for you

If you own an S&P 500 or Nasdaq fund: you already own Microsoft, Nvidia, Apple, Intel and AMD. Winners and losers both sit inside your fund. That's the point of investing in AI without picking winners.

If you pick stocks: one product launch rarely changes a company's path. Watch whether businesses actually buy these in volume over the next few quarters.

If you're shopping for a laptop: unless you run AI models or edit video for a living, a normal laptop is fine. You don't need 128GB to use ChatGPT in a browser.

Frequently asked questions

What is the Microsoft Surface Laptop Ultra?

It is a Windows laptop unveiled on October 7, 2026, built around Nvidia's Arm-based RTX Spark chip. It supports up to 128GB of unified memory so it can run large AI models directly on the device. Microsoft calls it the most powerful Surface it has made.

How much does the Surface Laptop Ultra cost and when is it out?

It starts at $2,599. A version with a more powerful chip starts at $3,700 and fully loaded models reach $5,900. It is available starting October 16, 2026. Microsoft is offering up to $1,000 off for trading in a MacBook Pro.

What does it mean to run AI locally?

Most AI tools send your request to a distant data centre, which does the work and sends back an answer. A local AI model runs on your own computer. There is no per-use fee, it works offline, and your data stays on your machine. It needs a lot of memory and a powerful chip, which is why these PCs are expensive.

Is the AI PC bad news for OpenAI?

It could trim demand at the edges. If everyday tasks can be done by a free model on your laptop, fewer requests go to paid cloud services such as ChatGPT. The most capable models still need data centres, so this is a slow shift and not a replacement.

Which stocks are affected by the Microsoft and Nvidia AI laptop?

Nvidia gains a new market in PC processors. Microsoft gets a reason for customers to upgrade Windows PCs. Intel and AMD face a new rival in high-end laptops, and Apple's MacBook Pro has a direct competitor. Memory makers benefit because each machine needs so much memory.

Primary sources

Data & disclaimer: This article is for educational purposes only and is not financial or investment advice. Figures reflect data available as of Oct 9, 2026, and conditions change — always confirm current pricing, rates and rules with the provider before you act. Written by Elizabeta Dimoska. See our editorial standards and disclosure.

ED
About the author

Elizabeta Dimoska

Founder and writer of RiskStock. Self-directed investor covering ETFs, long-term investing, tax-advantaged accounts (TFSA, RRSP, Roth IRA, 401(k)), retirement, macro, and markets — in plain English, with every claim tied to a primary source. Not a licensed financial advisor; RiskStock is educational. See our editorial standards.

More from Elizabeta Dimoska →

Comments

Want More Like This?

Get our weekly newsletter with market recaps, educational explainers, and honest takes — delivered every Sunday.

Subscribe Free