Microsoft just announced the Surface RTX Spark Dev Box at Build 2026, and for the first time in years, a Surface device made me stop scrolling my X feed for all the Build 2026 news and stare at the quirky design.

The compact AI-focused desktop is designed for developers building local AI workloads, fine-tuning models, running agentic workflows, and experimenting with on-device inference without constantly depending on cloud compute. Microsoft says the system delivers up to one petaflop of AI compute using NVIDIA’s RTX Spark superchip with 128GB of unified memory.

But the specifications aren’t what intrigues me.

The Surface RTX Spark Dev Box looks bizarre in the best possible way. Microsoft built the machine using a monolithic aluminum chassis covered in roughly 1,000 ventilation holes, which functions as both the thermal system and the visual identity of the product. Microsoft references the “1,000 teraflops” of compute performance as the inspiration behind the number of holes in the industrial grille design.

And yes, it absolutely reminds me of Apple’s infamous “cheese grater” Mac Pro design from 2019. Except here, the holes actually feel more purposeful because this thing genuinely needs cooling for sustained AI workloads.

Surface has been missing this kind of hardware personality for a long time.

After Panos Panay left Microsoft, the Surface lineup slowly shifted into making safer, more mainstream hardware. Recent Surface Laptop and Surface Pro generations have been premium and polished, but they are in no way outstanding when compared to other premium Windows PCs from the likes of Dell and Lenovo. Sadly, Surface stopped feeling experimental.

But I’m delighted to see the Surface RTX Spark Dev Box change that notion.

Of course, the performance claims need to be treated carefully until real-world benchmarks arrive. However, seeing Microsoft build something unapologetically niche and visually aggressive feels refreshing.

Specifications of the Surface RTX Spark Dev Box

Surface RTX Spark Dev Box is built entirely around local AI development

Microsoft says the new Dev Box ships with a developer-optimized version of Windows 11 Pro configured specifically for AI workloads.

According to the official Surface documentation, developers will be able to start coding “on day one” because Microsoft preconfigures the system with Visual Studio Code, GitHub Copilot in Windows Terminal, WSL, PowerShell 7, Git, Python, Node.js, GPU passthrough, CUDA support, and multiple developer-focused settings already enabled.

Even Windows 11 is modified around developer workflows.

Dark mode is enabled by default. Widgets are removed. Do Not Disturb is enabled. PowerShell 7 becomes the default shell. Microsoft says the entire idea is to eliminate the hurdles of setting up so developers can begin working immediately after first sign-in.

Most AI developers today spend hours configuring environments, debugging dependencies, fixing paths in Python, setting up CUDA drivers, installing WSL packages, and dealing with all the headaches in the container before even touching a model.
Microsoft appears to be trying to turn Windows into a ready-made AI workstation instead of just an operating system.

At the center of the system is NVIDIA’s RTX Spark superchip, combining a Grace CPU and Blackwell RTX GPU into one package. Microsoft claims the machine can run AI models with over 120 billion parameters locally using up to 128GB of unified memory. The company also says the system supports local inference with one million token context windows and enough sustained compute to fine-tune models that previously required expensive cloud GPU instances.

Again, all of that sounds incredibly ambitious. No independent performance numbers exist yet, and people should absolutely remember what happened with Microsoft’s earlier Snapdragon-powered Windows Dev Kit 2023.

Microsoft heavily promoted the Qualcomm developer system before quietly cancelling it later. Surface RTX Spark Dev Box is still awaiting FCC approval, too, so some caution is warranted here.

Still, from a design and positioning standpoint, Microsoft clearly understands where AI development is heading. Cloud AI costs are becoming absurdly expensive. Developers are looking for hybrid workflows where smaller models, inference tasks, and experimentation happen locally while larger workloads move to the cloud only when necessary.

Microsoft repeatedly used the phrase “unmetered intelligence” during Build 2026. The idea is that local AI compute reduces token costs, lowers cloud dependency, and allows developers to iterate faster. The Dev Box looks like a product designed by people who understand those workflows.

The duality between the Surface RTX Spark Dev Box and the Surface Laptop Ultra

The Surface Laptop Ultra looks fantastic. In many ways, it feels like Microsoft’s answer to the MacBook Pro, and while the inspiration is obvious, I honestly think Microsoft’s version looks cleaner and more refined. The hardware design is premium, modern, and distinctly Surface, even if it still follows the familiar high-end MacBook-ish formula.

But then you look at the RTX Spark Dev Box, and it feels like it came from an entirely different era of Microsoft hardware design philosophy.

One product is polished, mainstream, and commercially safe. The other is aggressively niche, covered in ventilation holes, and unapologetically weird, which is understandable as it is positioned for devs.

But while the Surface Laptop Ultra may be the more practical product, the Dev Box is the one that reminds me of the era when Surface hardware constantly experimented with unusual ideas, strange form factors, and ambitious industrial design.

That’s the version of Surface I’ve missed the most.

The best-designed Surface devices showed Microsoft at its most ambitious

Microsoft built the Surface brand around hardware experiments that were years ahead of the rest of the PC industry. Some products succeeded commercially, while others struggled because Windows or Intel hardware was not fully ready for the ideas Microsoft wanted to push.

Even so, it’s the designs and the ideas from the Surface line that made the PC industry interesting as opposed to the mainstream Windows PC and MacBook Pros that have grown to be boring with every passing year. A Surface inspires you to do greater things, at least, that’s what I felt years ago.

Surface Pro (2013)

Back in 2013, the original Surface Pro completely changed how I thought about Windows hardware. A full PC inside a tablet with a kickstand and detachable keyboard sounded impractical at the time, yet Microsoft somehow made the concept feel usable for everyday work.

The integrated kickstand became one of the defining Surface design elements, and nearly every Windows tablet manufacturer eventually copied some version of it.

Surface Pro X (2019)

Surface Pro X pushed the Surface aesthetic even further with ultra-thin bezels, ARM hardware, LTE connectivity, and magnetic Slim Pen storage built directly into the keyboard. The device looked futuristic long before Windows on ARM matured enough to support the experience properly.

Even today, Surface Pro X still looks more modern than many current Windows laptops.

Surface Laptop (2017)

The original Surface Laptop deserves far more credit than it usually receives. Alcantara keyboards sounded like a terrible idea on paper, but the device looked clean and premium compared to the endless wave of generic aluminum laptops available at the time (even MacBooks).

Microsoft managed to make Surface Laptop feel warm and distinctive instead of cold and industrial.

Surface Laptop Go (2020)

Surface Laptop Go took that same design language and compressed it into a smaller, more affordable machine. Compact premium laptops barely existed in the Windows ecosystem back then, and Microsoft somehow made an entry-level Surface device feel desirable (just like what the MacBook Neo did 6 years later).

Surface Book (2015)

Surface Book is one of the wildest laptop designs Microsoft has ever shipped.

The fulcrum hinge looked unlike anything else on the market, while the detachable display transformed into a standalone tablet with the GPU housed inside the keyboard base. Closing the laptop left a visible gap between the display and keyboard, which admittedly looked strange, but the engineering ambition behind the product was impossible to ignore.

Surface Laptop Studio (2021)

Surface Laptop Studio carried that same experimental energy forward with its floating pull-forward display mechanism. The screen could transition between laptop, stage, and tablet modes in a way that still feels unique today. I still remember how my eyes glimmered when I watched Panos Panay’s legendary presentation on my normal laptop, wishing to buy it in a couple of years, but it never received an update!

I haven’t ever gotten a similar feeling since then, and I watch keynotes of all major device launches. Microsoft clearly wanted to rethink how creative professionals interacted with Windows hardware instead of simply building another premium notebook.

Surface Studio (2016)

Surface Studio remains my favorite Surface product Microsoft has ever created.

The introduction video alone still stands out years later because Microsoft transformed a giant all-in-one PC into a drafting table using a gravity-defying hinge system. Combined with Surface Dial support, the entire experience felt completely different from anything Apple or other PC makers were attempting at the time (and even now).

Microsoft should seriously consider reviving Surface Studio using RTX Spark hardware because architects, designers, creators, and AI developers would probably love a modern Studio built around local AI compute.

Surface Duo (2020)

Surface Duo deserves recognition because, despite the software problems, the hardware itself was beautiful. I still prefer the dual-screen approach over many modern foldables because Surface Duo felt thin, elegant, and mechanically satisfying in a way most foldables still struggle to achieve.

The hinge engineering alone made the device feel special. Also, I feel the original version looked better than the Duo 2.

Surface Neo (2019)

Surface Neo was another Surface product I genuinely wanted to exist. Modern, efficient chips from Qualcomm, Intel Lunar Lake, or Panther Lake could finally make that form factor practical today.

Microsoft may have cancelled Neo too early. Well, now another company has made the “Neo” name ubiquitous.

Surface Hub 2 (2018)

Surface Hub was equally ambitious because giant collaborative displays running Windows always felt incredibly “Microsoft.” The product line represented Microsoft’s vision for workplace collaboration long before hybrid work became mainstream.

I still think killing Surface Hub was a mistake.

Surface lost its identity for a while

Older Surface devices always tried to reinvent what PCs could look like. Some ideas failed. Some were impractical. Some launched before Windows itself was ready for the hardware.

But at least Microsoft was trying.

Windows and Intel processors were usually the limiting factors. Battery life was poor. Performance consistency struggled. Touch-first experiences were okay-ish. Many of those ambitious form factors arrived before the software ecosystem could properly support them.

Ironically, modern ARM chips, better efficiency, stronger GPUs, local AI processing, and Microsoft’s renewed focus on native Windows performance might finally make some of those old Surface ideas viable again.

Surface RTX Spark Dev Box feels like the first sign of Microsoft rediscovering that confidence. And even if the product itself ends up being niche, I hope Microsoft keeps going down this path again. Because boring laptops are everywhere.

The post Microsoft’s new Surface RTX Spark Dev Box is like the cheese grater Mac Pro, and that’s why I love it appeared first on Windows Latest

Modern IoT and event-driven applications demand more than reliable messaging. They require the ability to handle larger payloads, efficiently route data at scale, and automatically adapt to changing workloads. Today, we're excited to introduce a new set of capabilities in Azure Event Grid Namespace (MQTT Broker & Event Broker) that help organizations build more scalable, resilient, and cost-efficient connected solutions. With support for MQTT v5 Subscription Identifiers, larger messages(1MB), and Auto scale-up and scale-down, customers can simplify architecture, reduce operational overhead, and accelerate deployments across manufacturing, automotive, retail, energy, and other connected industries.

Event Grid Standard: MQTT v5 Subscription Identifier (GA)

What it solves

In many MQTT applications, one client subscribes to several topics at the same time. When a message arrives, the application needs a quick way to tell which subscription triggered it and what should happen next.

• Which subscription triggered this message?
• Which business rule should process it?

What’s new

MQTT v5 Subscription Identifier is now generally available in Event Grid MQTT broker. Each subscription can have an ID, and that ID is included when the message is delivered.

Each subscription can have an ID, and when a message is delivered, that ID comes with it.

Why it matters

• Faster message routing
• Cleaner application logic
• No need to parse topic strings

Simple example

Example subscriptions:

Subscription ID 101 = factory/line1/temperature

Subscription ID 202 = factory/line1/vibration

When a message arrives

If the message comes in with Subscription ID 101, the app immediately knows it is temperature data and can send it to the correct processing path without inspecting the topic string.

Common use cases

• Manufacturing lines with hundreds of sensors
• Automotive telemetry platforms
• Multi‑tenant IoT solutions
• Rule engines and stream processors

Event Grid Standard: 1MB Message Support (GA) - Coming Soon

Applies to both MQTT and non-MQTT workloads.

What it solves

As devices and applications become more capable, the events they send are becoming richer too.

In many real-world scenarios, a single message may need to include things like:

• Sensor batches
• Telemetry snapshots
• Diagnostics data
• Configuration payloads

Before this, customers often had to split large messages into smaller pieces or store the payload somewhere else and send only a reference.

What’s new

Event Grid Standard Namespace now supports larger event payloads upto 1MB for both MQTT and non-MQTT scenarios. That means customers can send more complete information in one message and reduce complex workarounds in design.

Why it matters

• Fewer architectural components
• Simpler producers and consumers
• Lower latency for real‑time decisions

Example

Before: A factory device sent the main payload to storage and then published a small event that pointed to that file.

Now: The same device can publish the full event payload directly through Event Grid, making the flow easier to build and easier to operate.

Example: A smart machine can send a larger diagnostic package in one event, including the machine ID, a batch of readings, and a timestamp, instead of splitting the information across multiple services.

Common use cases

• Smart manufacturing telemetry
• Vehicle diagnostics and firmware data
• Batch sensor readings from edge devices
• Rich business events for analytics

Event Grid Standard: Autoscale Up and Down (Preview) - Coming Soon

Applies to both MQTT and non-MQTT workloads.

What it solves

Event-driven systems rarely operate at a constant rate. Traffic patterns fluctuate throughout the day, with predictable peaks, seasonal surges, and unexpected bursts that can quickly overwhelm static infrastructure.

Think about:

• Morning factory startup
• Fleet check‑ins at the top of the hour
• Flash sales or system bursts

Manually sizing infrastructure leads to:

• Over‑provisioning (higher cost)
• Under‑provisioning (missed events)

What’s new

Azure Event Grid Standard Namespace now supports Autoscale Up and Down, automatically adjusting capacity based on real-time workload demand. Whether you're running MQTT-based IoT workloads or event-driven applications using HTTP, AMQP, or CloudEvents, Event Grid dynamically scales resources to match traffic patterns without manual intervention.

As demand grows, Event Grid automatically adds capacity to absorb increased traffic and maintain performance. When activity subsides, resources scale back down to optimize efficiency and reduce costs.

Why it matters

 With Autoscale, customers can focus on building applications instead of managing infrastructure and reap benefits like:

• Scale seamlessly during traffic spikes without pre-provisioning for peak demand.
• Reduce costs during quieter periods by automatically scaling capacity down.
• Maintain consistent performance and reliability across changing workloads.
• Eliminate manual tuning and capacity planning, reducing operational overhead.
• Optimize resource utilization while supporting millions of devices and high-volume event streams.

Use case: A Modern Event-Driven Manufacturing Platform

A smart manufacturing company uses Microsoft to connect factory machines, sensors, edge gateways, and enterprise applications through a single scalable eventing platform. Devices publish rich telemetry and diagnostic payloads using MQTT, while cloud applications send operational and business events using HTTP. The company has steady traffic during the day, a large spike during shift changes, and much lower usage overnight. With autoscale, Event Grid adjusts capacity automatically, so performance stays strong when traffic is high and costs stay more efficient when traffic is low.

Common use cases

• Shift‑based manufacturing systems
• Connected vehicle platforms
• Event‑driven commerce backends
• Analytics pipelines with bursty data

Building IoT platforms, manufacturing and automotive solutions, or high-scale event-driven applications?

Azure Event Grid MQTT Broker now delivers larger messages, subscription Identifiers, and autoscale up/down, helping you simplify operations, improve performance, and optimize costs all on a single eventing platform.

Event Grid Basic: Powering Event Driven Payments with Stripe (GA)

Developers can seamlessly route Stripe events into Azure Event Grid to build scalable, event driven architectures without managing webhooks or custom brokers. Event Grid acts as a fully managed event broker between Stripe and Azure services.

Azure Event Grid now delivers Stripe events directly to Azure subscribers, including:

• Azure Functions
• Logic Apps
• Event Hubs
• Service Bus
• And more

With Event Grid’s native integration with Microsoft Fabric, Stripe events can also be ingested as continuous streams, enabling real time transformation, enrichment, and analytics within Fabric. Learn more.

Hello, Outlook community. I’m Vicki Milton, a Principal Product Manager on the Outlook team.

Over the last year, we’ve added important capabilities across areas such as offline support, shared mailboxes, and PST files. Alongside those milestones, we've continued to deliver smaller improvements that help people work more efficiently throughout the day.

This article highlights 15 productivity features in the new Outlook for Windows that can help you stay organized, reduce routine effort, and keep important work moving.

Mail features

Email remains central to how many people manage communications, priorities, and follow-up. Outlook includes familiar tools for composing and organizing messages, along with newer capabilities that can help reduce friction and make inbox management more efficient.

1. Pin a mail: Keep important messages easy to find. The Pin feature keeps a selected email at the top of your inbox so it remains visible as new messages arrive. This can be useful for items you need to reference often or do not want to lose track of, such as travel details, approvals, or active requests. By keeping priority messages in view, Pin can reduce time spent searching and help you stay focused on current work.

2. Snooze a mail: Return messages when they are relevant again. Snooze lets you temporarily remove an email from your inbox and have it reappear at a time you choose. This can help keep your inbox focused on messages you can act on now while ensuring follow-up items come back when they are timely. It is particularly useful for requests that depend on additional information, scheduled tasks, or work you plan to handle during dedicated focus time.

3. Add multiple categories at the same time: Organize messages with fewer steps. If you use categories to manage incoming mail, Outlook makes it possible to apply more than one category in a single action. This can help when you need to capture multiple types of context, such as project, priority, or follow-up status, without reopening menus repeatedly. It is especially useful when processing a large number of messages.

4. Sweep: Reduce repetitive inbox cleanup. Sweep lets you create automatic actions for messages from a specific sender. For example, you can delete promotional mail after a set period, keep only the latest message in a thread, or move recurring updates to a folder. This can help reduce manual cleanup and keep your inbox more focused on items that need attention.

5. Schedule Send: Write on your schedule and deliver at the right time. Schedule Send lets you prepare messages when it is convenient for you and send them later at a time that works better for the recipient. This can improve visibility, support more intentional communication, and reduce the need to rely on reminders or leave messages in Drafts.

6. Simplified folder sharing: Share folders more simply. Sharing a mail folder has traditionally required extra permission steps, especially for nested folders. Now, when you share a folder, Outlook can automatically apply the visibility permissions needed for its parent folders. This can reduce setup effort, help avoid access issues for recipients, and make folder sharing easier to complete with confidence.
   

Calendar and meeting features

For many people, the workday is shaped by meetings, schedule changes, and the need to stay aligned on what comes next. Outlook includes calendar and meeting capabilities that can help simplify planning, reduce coordination overhead, and make follow-up easier.

7. Follow a meeting: Stay informed without attending live. The Follow RSVP option lets you indicate that you will not attend a meeting but still want access to the recap. This can be helpful when schedules overlap or when a meeting is useful to monitor without joining in real time. It can help you stay connected to outcomes and shared materials while keeping your calendar more manageable.

8. Save calendar views: Return to the calendar setup you need more quickly. Saved Views let you store specific calendar combinations and switch back to them without rebuilding the same view each time. This can save time for people who move frequently between personal, team, and project schedules. It also can make it easier to review the right set of calendars for different planning tasks.

9. Improved meeting tracking: Work with meeting responses more efficiently. Outlook includes tools that make it easier for organizers to review and manage meeting responses. You can sort attendee lists, search for names in the Tracking view, and copy or download response details when needed. These capabilities can be especially useful for larger meetings where attendance information needs to be reviewed quickly.

10. Meeting recap: Find follow-up materials in one place. After a Teams meeting, the calendar event in Outlook can surface a Meeting recap with links to the recording, transcript, and shared files. This can make it easier to review what was discussed, confirm details, or catch up afterward. By keeping these materials together, Meeting recap can reduce the time it takes to get oriented after a meeting.

11. Filtered views: Reduce visual clutter in your calendar. Filters let you hide meetings you are not attending and limit the distraction of declined or informational events. This can make it easier to scan your schedule, identify conflicts, and focus on the meetings that need your attention. For people with full calendars, it can help make planning more straightforward.

12. Change a recurring event: Update future meetings while preserving earlier ones. When plans change, Outlook lets you edit the current event and all following events in a recurring series. This can make it easier to adjust details such as time, location, or agenda going forward without changing the record of past meetings. It can simplify updates for organizers and reduce disruption for attendees.

Personalization and settings

Settings can play a practical role in day-to-day productivity. A few adjustments can make it easier to focus, move between accounts and calendars, and work in a way that fits your preferences. Here are several settings-related features that can help make Outlook feel more streamlined and manageable.

13. Rename your email accounts: Make the right inbox easier to recognize. If you use multiple accounts in Outlook, you can assign each one a custom name. This can help you tell accounts apart more quickly, reduce the chance of sending from the wrong inbox, and make navigation simpler as you move between accounts during the day.

14. Modern themes: Choose a look that supports comfort and clarity. Outlook includes theme and color options that let you tailor the experience to your preferences. Visual settings can influence readability and comfort, especially for people who spend much of the day in email and calendar. Options such as Dark Mode and color customization can help make the interface feel easier to use over time.

15. Keyboard shortcuts: Keep familiar ways of working. In Outlook, you can choose the shortcut style you prefer in Settings. This can help you maintain existing habits, reduce adjustment time, and complete common tasks with fewer steps. For people moving from classic Outlook or Outlook on the web, shortcut flexibility can make the transition more consistent.

These features reflect a broader effort to help people work more efficiently in the new Outlook for Windows. Whether you are managing a high volume of email, coordinating a full calendar, or tailoring the experience to match your workflow, these updates are designed to reduce effort and improve day-to-day productivity.

For more information and step-by-step guidance, see the Microsoft Support articles and the Learning Path.

{
  "id": "order-1001",
  "type": "order",
  "customerId": "customer-42",
  "orderDate": "2026-06-03",
  "status": "Shipped",
  "items": [
    {
      "productId": "product-10",
      "name": "Mountain Bike",
      "quantity": 1,
      "unitPrice": 1200
    },
    {
      "productId": "product-20",
      "name": "Helmet",
      "quantity": 1,
      "unitPrice": 75
    }
  ],
  "shippingAddress": {
    "street": "123 Main Street",
    "city": "Orlando",
    "state": "FL",
    "postalCode": "32801"
  }
}

{
  "productId": "product-10",
  "name": "Mountain Bike",
  "quantity": 1,
  "unitPrice": 1200
}

{
  "id": "customer-42",
  "type": "customer",
  "customerId": "customer-42",
  "name": "Contoso Bikes"
}

{
  "id": "order-1001",
  "type": "order",
  "customerId": "customer-42",
  "orderDate": "2026-06-03"
}

{
  "id": "ticket-9001",
  "type": "supportTicket",
  "customerId": "customer-42",
  "status": "Open"
}

/customerId

/tenantId

/categoryId

id = order-1001
partition key = customer-42

SELECT * FROM c
WHERE c.customerId = "customer-42"

SELECT * FROM c
WHERE c.status = "Pending"

tenantId|customerId

region|tenantId

customerId|orderYear

The post Data Modeling and Partitioning in Azure Cosmos DB appeared first on The Training Boss.