April 06, 2026

What TOPS reveals about performance, efficiency, and long-term readiness

AI for business is no longer limited to pilot programs or experimental teams. It’s showing up in analytics dashboards, workflow automation, content generation, and everyday productivity tools that employees increasingly expect to “just work”. In enterprise environments, this shift is already influencing device strategies as organizations look to standardize on AI-ready hardware.

As these capabilities move closer to the device, business laptops are being asked to do more than traditional computing. That shift is forcing IT leaders to evaluate a newer performance metric that’s gaining attention: TOPS (Tera Operations Per Second), a measure of how much AI processing a device can handle per second.

AI workloads are moving to the device

AI adoption continues its rapid acceleration across industries, driven by anticipated productivity gains and the growing availability of AI-enabled software. Many of these experiences rely on inference, running trained models locally to deliver real-time results without sending data back and forth to the cloud.

To support this shift, select Surface for Business devices include a dedicated neural processing unit (NPU), enabling supported AI workloads to run locally with improved efficiency and more consistent performance. ¹ For many organizations, this can support faster response times, lower latency, and more consistent performance across hybrid environments.

As AI workloads move closer to users, laptops need to handle them efficiently. That’s where metrics like TOPS AI begin to matter, not as a marketing number, but as a signal of how prepared a device may be for sustained, on-device AI use.

What TOPS actually measures

TOPS, which stands for Tera Operations Per Second, measures how many operations a processor can perform in one second when handling AI workloads. In practical terms, TOPS reflects how quickly a system can process AI inference tasks such as real-time data analysis, image recognition, or natural language processing.

Modern business laptops increasingly rely on specialized components like an NPU (neural processing unit), which functions as an AI accelerator alongside the CPU and GPU. These AI accelerators are designed to handle repetitive, parallel operations more efficiently than general-purpose processors, often using less power in the process.

On Surface Copilot+ PCs, the NPU works alongside other system components to support hardware accelerated AI experiences, helping certain supported features run efficiently without placing sustained load on general purpose processors. ²

Understanding AI TOPS allows IT leaders to compare devices more realistically, but it’s important to be viewed in context.

Where TOPS affects performance at scale

TOPS ratings indicate a device’s capacity for on-device AI processing and can help contextualize how well it may handle AI-enabled workloads. For business users, this can translate to faster AI-assisted features, smoother multitasking, and fewer performance bottlenecks when multiple AI-enabled applications are running at once. In enterprise collaboration scenarios, this can help sustain experiences such as AI-enhanced video and audio features on Surface devices, even when multiple applications are running concurrently. ³

From an IT perspective, local AI processing can reduce reliance on constant cloud connectivity by limiting the need to send data back and forth for inference tasks. In some environments, this may also help manage bandwidth demands or cloud compute usage.

For enterprise IT teams, pairing on device AI with Surface’s built-in security and manageability, such as Windows Autopilot and cloud based management through Microsoft Intune, helps maintain control and consistency as AI usage expands across device fleets. ⁴ Over time, that shift can contribute to more consistent user experiences and more predictable performance as AI workloads expand.

Microsoft notes that running AI inference directly on devices can reduce latency and cloud compute costs for common, high-volume tasks, while keeping processing available even when connectivity is limited. That said, TOPS alone doesn’t determine whether a laptop is right for your environment.

Avoiding common TOPS misconceptions

A higher TOPS number doesn’t automatically mean a better overall business laptop. Sustained AI performance depends on several factors working together: thermal design, battery capacity, power management, and how workloads are distributed across the CPU, GPU, and NPU. That systems level balance is why Surface for Business devices are designed and tested with Windows and security features together to support sustained performance, reliability, and enterprise readiness beyond any single metric.

For some roles, occasional AI-assisted tasks may not require the highest TOPS rating available. For others, such as data-heavy teams or roles using continuous AI-driven workflows, AI readiness becomes more critical. The key is matching TOPS capability to real workload demands rather than chasing specs in isolation.

How IT leaders can evaluate AI readiness

For IT leaders planning refresh cycles, TOPS can serve as one part of a broader readiness check. Questions worth asking include:

• Does the device include an NPU or dedicated AI accelerator?
• How does AI performance hold up on battery power?
• Are drivers, firmware, and operating systems optimized for AI workloads?
• Will the laptop support future AI features as software evolves?
• Can the device support enterprise‑grade deployment, security, and lifecycle management alongside AI acceleration?

Modern business laptops, including Surface for Business devices with integrated NPUs, are beginning to incorporate dedicated AI acceleration to support these emerging on-device workloads. Evaluating these factors together helps reduce the risk of early obsolescence and supports longer-term planning as AI becomes more embedded in daily work.

Is your device strategy ready for AI-powered work?

As AI continues to shape how employees work, business laptops are becoming active participants in that shift, not just endpoints. Understanding metrics like TOPS can help organizations make more informed decisions about performance, efficiency, and longevity. Surface Copilot+ PCs bring together on device AI acceleration, Windows security, and enterprise management capabilities to help support AI-powered work as it continues to evolve.

If you’re reviewing your current device strategy or planning ahead for broader AI adoption, this is a good moment to assess whether your hardware foundation aligns with where AI-enabled work is headed next to help keep you ahead of the curve.