Microsoft opened Surface Laptop Ultra preorders on October 7, with availability planned to begin October 16. That is the new milestone: the company introduced the hardware on May 31. A Canadian preorder now presents a concrete choice of processor, memory, storage, keyboard and Windows edition.
The useful question is whether the exact configuration supports the work you intend to run. This is documentation-based reporting and editorial analysis. Inquory has not tested a device, selected a store configuration, checked out an order or measured application performance.
The paper cards, stylized closed computer and unfinished checklist separate listed configuration choices from unresolved workflow checks. They do not depict an actual product, replaceable memory, internal hardware or a tested setup.
AI-generated editorial illustration · InquoryWhat the Canadian listing actually separates
On October 12, Microsoft's Canadian consumer configurator displayed a Platinum model with an 18-core CPU, 5,120-core GPU, 24 GB unified memory, 512 GB SSD and Windows 11 Home at CAD $3,899.99. It also displayed an 18-core/5,120-core Platinum model with 32 GB memory, 1 TB SSD and Windows 11 Pro at CAD $4,849.99. Those are separate listed configurations, not checkout quotes or verified delivery dates.
The page assigns Home to 24 GB configurations and Pro to 32–128 GB configurations. A memory selection therefore changes more than the memory number. Record the operating system alongside the processor, storage and keyboard before comparing an offer.
The Canadian business listing starts at CAD $4,099.99 and says its 24–128 GB configurations include Windows 11 Pro. We have not established equivalent consumer and business SKUs. The two starting prices cannot establish the cost or value of a like-for-like upgrade. Listing prices and configuration status can change; refresh the exact row and order terms before a purchase decision.
A local-model headline leaves a workflow to verify
Microsoft's October 7 announcement says memory varies by configuration and that the amount addressable by the GPU is less than total system memory. Its AI compute headline uses theoretical sparse FP4 performance. That does not establish the speed, quality or memory needs of a particular application and model.
Write down the model, precision or quantization, context size and runtime you intend to use. Ask the application or runtime supplier for support covering that setup on the selected hardware. Treat an unanswered requirement as an open question, rather than converting a maximum specification into a successful workflow.
Keeping a model on the machine also does not, by itself, establish that every step stays offline. Downloads, sign-in, extensions, synchronization and optional services are separate questions to check in the tools you choose. This article makes no privacy, security or cloud-cost claim for an untested setup.
Check the application and the things plugged into it
Microsoft's Windows on Arm emulation documentation says emulation supports user-mode code, not drivers; kernel-mode components must be compiled for Arm64. That is general platform documentation, not acceptance proof for an Ultra configuration or every application, plugin and peripheral you own.
Use exact app and plugin versions in the compatibility question. For a device that requires a driver, obtain the supplier's Arm64 support statement. A connector that physically fits does not answer the driver, software or supported display-mode question.
Microsoft links its tested-peripherals documentation from the launch announcement. We did not qualify an Ultra-specific row in that page's testing matrix. The document says its list is not comprehensive, so this observation establishes neither blanket compatibility nor incompatibility. Check your exact dock, display, cable and required mode with applicable supplier documentation.
| Decision | Evidence to record | Question still open until verified |
|---|---|---|
| Canadian configuration | Processor, memory, SSD, Windows edition and keyboard from the exact listing | Final order terms and delivery for the selected offer |
| Local model | Model, runtime, precision and context requirements | Behaviour and performance in the intended workload |
| Application chain | App and plugin versions; applicable platform support | Whether the complete workflow works together |
| Connected equipment | Exact dock, display, cable and any driver requirement | Whether the required mode is supported on this setup |
A practical decision before a preorder
Imagine Camille, a fictional Canadian designer who needs one rendering application, a specific plugin and a dock driving two displays. Camille can compare the listed configurations now, but the decisive next step is to obtain support information for that application chain and display setup. The example is an editorial scenario, not a customer test or recommendation to buy.
Before committing to an offer:
- Save the exact Canadian configuration and the time its price and status were checked.
- List the application, plugin, model and peripheral versions that the work depends on.
- Mark each requirement as documented, unresolved or actually tested; keep those states separate.
- Resolve a requirement that could prevent the work before treating a preorder as a suitable replacement.
The separate RTX Spark Dev Box has US-only preorder terms and planned November shipping in the October 7 announcement; those terms do not establish Canadian laptop availability. For Laptop Ultra, October 16 remains a planned availability milestone, not evidence that an order has arrived. The news gives buyers a configuration to investigate. A completed workflow still needs its own evidence.
For another example of separating a product promise from a practical trial, see our cloud-gaming connection and data test.