Tom Butler, VP, Worldwide Commercial Portfolio and Product Management at Lenovo – Interview Series

This interview was conducted in connection with MWC 2026, where Lenovo showcased its latest advancements in AI-powered commercial computing.

Tom Butler, Vice President, Worldwide Commercial Portfolio and Product Management at Lenovo, brings more than two decades of experience within the company, having progressed through multiple leadership roles spanning director, executive director, and now vice president of its global commercial portfolio. Based in the Raleigh-Durham area, he is responsible for shaping Lenovo’s worldwide commercial product strategy, aligning hardware, software, and services to meet enterprise needs. His career reflects deep expertise in product management, go-to-market execution, and large-scale portfolio leadership, built on earlier technical and operational roles at Cisco and Ericsson, where he developed a foundation in wireless systems, enterprise infrastructure, and customer support.

Lenovo is a global technology company generating tens of billions in annual revenue and serving customers across more than 180 markets. The company offers a broad portfolio spanning PCs, smartphones, tablets, servers, and enterprise infrastructure, alongside software, AI-driven solutions, and services designed to support digital transformation. Its business is structured across key segments including intelligent devices, infrastructure solutions, and services, enabling it to operate across both consumer and enterprise ecosystems. Increasingly, Lenovo is positioning itself around AI and integrated computing experiences, combining hardware with intelligent software and cloud-based capabilities to support modern workplaces and large-scale enterprise environments.

You’ve spent more than two decades at Lenovo, rising through product marketing, strategic planning, and portfolio leadership roles to now oversee the worldwide commercial notebook portfolio. How has that journey shaped your perspective on what truly defines an “AI PC” versus a traditional business laptop with AI features layered on top?

I’ve spent most of my career inside commercial PCs, so I tend to be pretty strict on definitions. An AI PC isn’t a traditional laptop with a few features added on, it’s a system designed from the ground up to run AI workloads locally, efficiently, and securely.

What matters to me is what the device can do natively. Can it run meaningful AI experiences without relying solely on the cloud? Can it adapt to the user, their data, their workflows? And can it do that in a way that meets enterprise expectations for security and manageability?

After years in this space, I’ve learned to look past the hype. If it doesn’t fundamentally change how people work, it’s not really an AI PC yet.

Increasingly, we’re seeing this as part of a broader shift toward hybrid AI architectures, where the device handles more locally and only goes to the cloud when needed.

As the leader responsible for ThinkPad, ThinkBook, and commercial software, how are you rethinking the commercial notebook roadmap to account for AI workloads that are increasingly running locally on-device rather than exclusively in the cloud?

As more AI workloads run locally, the roadmap needs to prioritize sustained on-device performance, not just peak specs. In practice, that means expanding NPU-enabled configurations across ThinkPad and ThinkBook and treating the NPU as a core resource alongside the CPU and GPU. Tasks like meeting summaries, transcription, intelligent search, and content creation are increasingly optimized for local acceleration.

It also requires a shift in how we think about software. Enterprise customers don’t want isolated AI features, they want integrated experiences that work across workflows, devices, and environments.

This also aligns with a broader shift we’re seeing from cloud-first to more hybrid AI models. Enterprises are starting to think about a “local-first, cloud-when-needed” approach, where the PC or edge device handles more of the workload, and the cloud is used more selectively. That has real implications not just for performance, but for cost, control, and how IT environments are designed. That’s why we’re focused on open, flexible architectures that support multiple models and evolving ecosystems, rather than locking customers into a single approach.

Enterprises are under pressure to modernize their device fleets, but refresh cycles are tightening. What’s the strongest business case today for CIOs to invest in AI PCs instead of extending the life of existing hardware?

The business case today is becoming clearer: CIOs are being asked to modernize for a world where productivity and security expectations have shifted, and older PC fleets can’t always reach that new baseline through software updates alone.

Performance is part of the story, but it’s no longer the most interesting part. What’s changed is that AI workloads are now part of everyday productivity, and older fleets simply weren’t designed for that.

The real ROI comes from enabling new ways of working, local AI that keeps sensitive data on the device, better collaboration experiences, and reduced dependency on constant connectivity. There’s also a cost dimension that’s becoming more important. Data centers are increasingly reserved for revenue-generating workloads, so if you can move testing, inference, or early-stage AI tasks to the client device or edge, you’re reducing unnecessary cloud usage and lowering overall cost.

At the same time, enterprises want to modernize without disruption. Pairing new devices with lifecycle services, phased deployments, and “prove-it-fast” pilots helps make refresh cycles feel deliberate and practical rather than reactive.
 
AI PCs often promise improved productivity through on-device copilots and automation. In real-world enterprise deployments, where are you actually seeing measurable gains — and where is the industry still ahead of practical outcomes?

Where we see real impact is when AI is tied to specific workflows. Meeting capture, summarization, search, and device optimization are already delivering measurable time savings, especially when those workloads can run locally without latency or dependency on connectivity.

Where the industry is still early is in broad, enterprise-wide transformation. There’s no single “killer app” yet, and adoption depends heavily on training, integration, and change management, not just the hardware

Immersive form factors are being positioned as the next evolution of smart devices. From your vantage point, what does “immersive” realistically mean for business users over the next three to five years — and which use cases are mature enough to move beyond experimentation?

For business users, “immersive” has to mean practical. More usable screen space, better focus, and more natural collaboration, not something experimental.

We’re seeing that come to life through expandable displays, multi-mode designs, and modular ecosystems that extend the workspace without adding complexity.

Over the next three to five years, the most successful use cases will align with mission-critical enterprise work: document-heavy tasks, collaboration, and mobile productivity. Other technologies, such as glasses-free 3D, rings, and some spatial interaction models, will likely mature first in specialized roles like design visualization, training, or advanced collaboration before becoming more widely adopted.
 
As AI workloads increase, devices must balance performance, thermals, battery life, and security. What engineering trade-offs are most challenging when building commercial AI PCs that need to perform reliably in enterprise environments?

The hardest part is that customers want everything at once: performance, battery life, thin design, and reliability. In reality, those things compete. Bigger batteries and stronger cooling add weight and thickness, so the real challenge is finding the right balance, not chasing a single spec.

Security also has to be built in from the start. As AI workloads move closer to the endpoint, decisions about firmware integrity, hardware protections, and supply chain assurance become part of the performance equation. In enterprise environments, fast but vulnerable systems simply don’t scale.

Security and data governance remain top concerns for enterprise buyers. How does the shift toward on-device AI inference change the conversation around privacy, compliance, and risk management?

On-device AI improves the conversation around privacy, because more data can stay local to the device. That’s a big shift, especially for regulated industries. It also helps address concerns around latency and control, since sensitive workloads don’t always need to leave the endpoint.

But it doesn’t remove the need for governance. Enterprises still need clear policies around how models are used, how data is handled, and how outcomes are validated.

That’s why we focus on security below the OS, supply chain integrity, and device-level trust. As AI moves to the endpoint, that’s where risk and responsibility increasingly sit.

Many businesses are experimenting with AI agents that automate workflows and decision-making. How do you see AI PCs evolving to better support agent-driven workflows locally, and what hardware or software innovations will be required to make that seamless?

I see AI PCs evolving from simple assistive features into platforms that support agent-driven workflows. We’re moving from AI as a tool you interact with, to AI that can act on your behalf. That’s a meaningful shift.

You’ll see this evolve across devices and ecosystems, not just within a single PC. The real opportunity is coordination, where AI understands context across your work and devices, not just isolated tasks.

To make these agent workflows seamless on the device, we’ll need continued progress in on-device acceleration and efficiency. Just as important is a cross-device intelligent software architecture that can interact with models as the ecosystem evolves, along with enterprise-grade guardrails so agents can operate securely and with trust.

That evolution will depend heavily on hybrid AI architectures, where agents can operate locally when needed, while still leveraging cloud-scale models when appropriate.
 
Lenovo has historically excelled at understanding customer feedback and translating it into product strategy. What recurring themes are you hearing from enterprise customers about AI-enabled devices that the broader market may be underestimating?

The themes we hear from enterprise customers are usually more practical than the industry conversation.

First, customers want clear enterprise use cases and consistent ROI. Refresh cycles are planned years in advance, so without mission-critical outcomes many teams remain cautious.

Second, even when new devices are deployed, underutilization is common if training and onboarding aren’t built in.

Third, governance and trust are essential. Buyers want local AI options that reduce unnecessary data exposure and clear visibility into what’s happening on the device.

Finally, the fundamentals still matter. Battery life, performance, ports, reliability, and repairability. AI doesn’t replace those expectations, it raises the bar for all of them.
 
Looking ahead five years, do you believe the term “AI PC” will still exist as a category, or will AI capabilities simply become an invisible, embedded layer across all commercial devices — and what does that imply for how companies differentiate their hardware?

I think the term “AI PC” will fade over time. AI just becomes part of what a PC is.

The real differentiation shifts to experience, how personal it is, how well it’s protected, how easy it is to manage at scale, and how confidently enterprises can deploy it.

In that sense, AI becomes invisible. What customers care about is whether the device helps people work better, and whether they trust it.

Thank you for the great interview, readers who wish to learn more should visit Lenovo.