Agentic AI PCs Target Enterprise Token Bills — at $2,600 and Up

Ten to fifteen percent of enterprise AI workloads will run on PCs within three to four years, according to Jack Gold of J. Gold Associates — and vendors are racing to build the machines that will run them.

The driver is cost. Cloud token bills are making enterprises hesitate to scale AI, and Microsoft's answer is what it calls "hybrid intelligence": hard tasks go to frontier cloud models, less demanding work runs locally on Copilot+ AI PCs.

"The future of AI is not about having just one model. It's about having the right model at the right moment — frontier capability when you need it, smaller, faster models when you want speed," said Pavan Davuluri, head of Windows and Devices at Microsoft, at an Oct. 7 PC launch event.

The PC will not replace the cloud. Preliminary AI work happens locally; tougher work — beyond what today's PCs can handle — goes to the cloud. "We'll always need to use the most advanced cloud-based frontier models for certain tasks, but… we don't need to use them for everything," wrote TECHnalysis Research Insights president Bob O'Donnell in a LinkedIn post.

Microsoft CEO Satya Nadella said hybrid intelligence will be everywhere, blurring the line between cloud and PC. "I just want to use my computing power wherever it is in sort of a seamless flow of what I'm trying to get done," he said.

The local tier matters for budgets: roughly 40% to 45% of AI workloads will stay in the cloud, Gold estimates, but companies will still need AI-accelerated hardware on the desktop. "Companies will need AI-accelerated PCs, and at some point we'll likely see a renewed emphasis on buying newer machines," Gold said.

What can an agentic AI PC actually do?

The new class of machines runs agents, generates video, writes code and chats with advanced genAI models locally. If the PC can't do the job, the task moves to the cloud.

"A year ago, such capabilities were exclusive to frontier cloud models. Now you can have these capabilities running on your PC," Davuluri said.

These PCs carry powerful GPUs that are distant cousins of data-center silicon. Nvidia's RTX Spark superchip — built on a Blackwell GPU — powers many of them, including Microsoft's new Surface Laptop Ultra.

Unlike older AI PCs, the new agentic machines run full AI models fast enough to be practical, said Gerardo Delgado, senior director of product management at Nvidia. "To us, an AI PC is the PC that will… run those local models so that you can actually keep the level of intelligence and enjoy the economic benefits," he said.

How does local-first AI cut cloud bills?

Microsoft demonstrated the mechanism with a debugging scenario. A team couldn't reproduce a software bug, so GitHub Copilot selected GPT-6 Luna in the cloud to analyze the code. The cloud model then handed off building and testing to local sub-agents running MAI Code 1.1 Flash on a PC.

The local work was free. One sub-agent fed 1.6 million tokens of code and context to the local model, which generated 10,500 tokens in response, a GitHub product manager said during the demonstration.

Microsoft compressed the 130-billion-parameter MAI Code 1.1 Flash model to make that possible. "We have taken that model to three bits of precision, reducing its size by nearly 80% while preserving coding quality," Davuluri said.

The demo used GitHub's HydraFusion feature, which moves coding tasks between local models and the cloud. Microsoft is bringing it to GitHub's Copilot app for Windows Copilot+ PCs.

HP is taking a similar path with its newly announced ZBook Ultra G3a mobile workstation, whose AMD GPU generates free tokens and cuts cloud bills. "It can be on-prem. They don't have to worry about confidential information going to cloud, or tracking tokens and getting $1,000 charges a week," said Brian Allen, manager for Global Z workstation products at HP.

The ZBook ships with Perplexity's Portable Computer service, whose agents link a local model to frontier models in the cloud. An agentic PC can handle basic AI-assisted design work in Autodesk's Revit 3D modeling software offline; for advanced work, Perplexity escalates to whichever of 19 frontier cloud models fits best. HP links to Perplexity's tools in the Windows taskbar, and Perplexity ships MCP connectors for Revit.

"I can do contextual typing of, 'hey, build this for me,' without having to know the details of that," Allen said. Revit takes years to learn, but "I can really learn how to do Revit through agentic AI in a matter of hours," he said. HP expects designers to keep proprietary work local and offload general tasks to the cloud.

What still stands in the way?

Cost, for one. The Surface Laptop Ultra starts at around $2,600 and reaches $6,000 fully configured. PCWorld senior reporter Mark Hachman noted in a review that a cheaper laptop can do the same job using cloud models, and the Surface could quickly go out of date as new models arrive.

Analysts expect PC prices to remain elevated until the end of 2027, driven by high component costs, the Iran war and an uncertain economy.

Deployment is not plug-and-play either, Gold cautioned. Companies will need to develop and customize models for these PCs depending on their enterprise applications; open-source models that can be customized could "significantly reduce AI costs and increase ROI," he said.

"For much of this to waterfall down to the average user, many different things will need to happen across component pricing, model evolution, agentic use cases, and more," wrote IDC group vice president Tom Mainelli in a LinkedIn post.

Even HP is still working out the market. Customers, Allen said, are in the "this PC is too good to be true" phase — benchmarking it against their own RAG models before they believe it.

The hybrid model is real and the economics are measurable; the friction, for now, is price and maturity.