HP’s AI PC shipments have surged to over 25% of its personal systems mix, a quarter ahead of internal plans, as the Windows 11 upgrade cycle and higher average selling prices drove a 6% revenue increase in the company’s fiscal third quarter. CEO Enrique Lores told investors that the combination of a commercial refresh and on-device AI demand delivered double-digit sequential revenue growth for the AI PC category, while traditional printer revenue slipped 4% amid competitive pricing and tariff pressures.

The results, covering the period ending July 31, confirm a simple but lucrative arithmetic: a mix shift toward premium, AI-enabled SKUs is improving profitability even as unit growth remains modest. HP reported total net revenue of $13.9 billion, with Personal Systems revenue reaching $9.9 billion—up 6% year-over-year—and Printing at $4.0 billion. The company said AI PCs now represent more than 25% of its unit mix, carrying a price uplift of 5% to 10% compared with non-AI equivalents.

The arithmetic of a mix shift

“We continue to shift our mix to higher value segments such as AI PCs, commercial premium and services,” Lores said on an earnings call. Unit growth was around 5%, but the larger share of higher-priced configurations pushed average selling prices higher, delivering margin expansion. HP’s management highlighted three specific drivers: AI PCs’ rising ASPs and double-digit revenue growth, the Windows 11 refresh converting older commercial fleets, and services and premium products contributing to overall profitability.

This strategy—trading volume for margin—is playing out at scale while HP also navigates geopolitical supply chain risks. The company has aggressively relocated production for North America-bound products outside China, moving assembly to Vietnam, Thailand, Mexico, and U.S. facilities. CFO Karen Parkhill noted that nearly all products sold in North America are now built outside China, a move that helped preserve margins despite rising trade-related costs. At the same time, HP saw double-digit revenue growth in China for Personal Systems, underscoring uneven global demand.

What makes a PC an ‘AI PC’?

The industry broadly defines an AI PC as a system equipped with a dedicated neural processing unit (NPU) or inference accelerator alongside a CPU and GPU, enabling local AI model execution. Microsoft’s Copilot+ initiative sets baseline NPU performance thresholds around 40 trillion operations per second (TOPS) for optimal on-device experiences. HP’s AI PC portfolio aligns with these specs, leveraging silicon from Intel, AMD, and Qualcomm.

But the hardware is only half the story. Microsoft has been building the software scaffolding with DirectML, the Windows Copilot Runtime, and toolchains that let developers target NPUs via popular frameworks like ONNX and PyTorch. This abstraction layer reduces the friction for independent software vendors (ISVs) to optimize applications for local inference. Adobe, for example, is integrating generative AI features into Creative Cloud that can run on PC NPUs, while Zoom is working on AI-assisted capabilities that benefit from on-device processing. Security vendor CrowdStrike has prototyped using NPUs to accelerate memory scans and endpoint detection, potentially lowering latency and cloud costs.

HP argues these examples prove organic demand—not just channel stuffing. “The key thing behind the AI PCs is the fact that over the next quarters, more and more software applications are going to take advantage of the capabilities that the AI PCs have,” Lores said. He noted that shifting workloads locally “means it will be faster and also it will reduce some of the cloud costs.”

The developer path to NPU acceleration has become smoother thanks to Microsoft’s open standards push. DirectML allows developers to target GPUs and NPUs with a common API, while the Windows Copilot Runtime exposes local models for tasks like text summarization and image generation. PyTorch and ONNX runtime integration means models from Hugging Face can feasibly run on-device with minimal porting effort. Qualcomm, Intel, and AMD have each shipped drivers and execution providers that expose their NPUs through these frameworks. The result is that proof-of-concepts that were cloud-only a year ago are now being tested on client hardware.

Windows 11 refresh fuels commercial momentum

The Windows 11 upgrade cycle remains a powerful tailwind. With support for Windows 10 ending in October 2025, enterprises are scrambling to replace aging fleets. HP estimates that slightly more than 50% of the installed base has migrated to Windows 11, indicating a substantial runway of refresh activity. Commercial customers are naturally gravitating toward systems that offer not just OS compatibility but also AI readiness, hence the jump in premium PC sales.

Lores told analysts he expects PC growth to remain in the mid-single digits for the second half of the fiscal year and into fiscal 2026, driven by sustained Windows 11 demand and AI PC uptake. The holiday consumer season is also expected to provide a seasonal lift.

Analysts see rapid adoption, but definitions vary

Research firms paint a bullish picture, though their numbers differ. Gartner forecast that AI-capable PCs would account for 31% of the global market by the end of 2025 and could reach 54% in 2026. By 2027, it may become difficult to buy a PC without AI capabilities. Gartner also expects 40% of software vendors to prioritize support for AI running directly on PCs by 2025, up from just 2% in 2024, with small language models deployed locally.

HP’s own 25% mix, achieved a quarter early, places it ahead of its own roadmap and suggests that the company is capturing a high-margin slice of an accelerating trend. However, comparing internal mix with market-wide forecasts is tricky because HP’s figures may reflect its specific channel and regional strengths.

Printing decline adds urgency to the PC pivot

The printer business—once HP’s cash cow—is under structural pressure. The 4% revenue decline to $4.0 billion was blamed on a softer office market and aggressive pricing competition. Lores admitted that enterprise customers are prioritizing PC upgrades over print investments, a trend unlikely to reverse. HP’s long-term printer strategy will likely involve deeper verticalization or a greater tilt toward contractual print services rather than transactional hardware sales, but for now, the personal systems division is carrying the profit load.

The honest caveats

For all the enthusiasm, significant uncertainties remain.

  • No killer app yet. Many current AI features are incremental—faster image processing, local speech recognition, AI noise cancellation—not transformative enough to force immediate fleet-wide replacement. The broader “killer app” that compels mass commercial replacement remains emergent rather than universal.
  • Pricing sensitivity. A 5–10% price premium may be material for cost-conscious buyers. Enterprise procurement teams will weigh whether the productivity gains and potential cloud savings justify the higher upfront cost over a typical three- to four-year lifecycle.
  • Security and data governance. On-device inference reduces cloud exposure but introduces new attack surfaces—model storage, local caches, and NPU drivers. CrowdStrike’s NPU acceleration experiments highlight both promise and new hardening requirements. IT teams must treat AI PCs as a security evolution, not just a hardware upgrade.
  • Supply chain and geopolitics. HP’s manufacturing pivot reduces tariff risks but adds complexity and near-term costs. Currency fluctuations and commodity inflation could compress margins even as hardware becomes more valuable.

What IT buyers and channel partners should do

For enterprise IT teams, the current moment calls for a strategic reevaluation. If a significant portion of the fleet still runs Windows 10 or older hardware, the Windows 11 end-of-support deadline creates a natural refresh window. But before ordering thousands of AI PCs, run targeted pilots with workloads that genuinely leverage NPUs—local transcription, summarization, endpoint detection—and measure latency, cloud cost offsets, and user feedback.

Treat AI PC security as strategy. Update endpoint protection, telemetry, and device hardening plans to account for local AI artifacts. The shift from cloud-dependent to hybrid AI workloads means rethinking data governance and model integrity.

For channel partners and resellers, the opportunity lies in selling total cost of ownership, not just specs. Bundle managed services, AI-specific security, and device lifecycle management to capture higher margins. HP’s own services growth signals where the profit pools are moving.

Real trend or first-wave hype?

HP’s fiscal Q3 results provide hard evidence that AI PCs are not a paper tiger. The company is executing a coherent strategy—push high-value SKUs, ride the Windows 11 refresh, and monetize nascent AI demand—and the payoff is visible in topline growth and margin expansion. With supply chains realigned and software ecosystem support building, HP is well positioned to ride the wave.

Yet the journey from early adoption to mainstream standard is far from complete. The killer app that forces universal upgrade cycles remains elusive. ISVs must move from proofs-of-concept to shipping features that demonstrably improve workflows, and enterprises must quantify ROI beyond marketing claims. Analyst forecasts suggest a tipping point within two years, but history shows that technology transitions often take longer than the bullish scenarios.

Looking ahead, HP’s checklist includes sustaining AI PC mix above 25%, translating the 5–10% ASP uplift into durable margin improvement, and watching for ISVs that ship NPU-accelerated features with measurable ROI. The Windows 11 migration’s second half will also be closely monitored—the company’s estimate that only half the installed base has converted suggests there is still plenty of commercial upgrade spending ahead. Whether that spending flows predominantly toward AI PCs or more traditional SKUs will reveal how much extra value the market truly assigns to on-device AI.

For now, the message is clear: AI PCs are already a profitable commercial category. HP has proven that integrating NPUs and selling the vision of on-device AI can boost ASPs and margins today. The next test is whether the software ecosystem and enterprise demand will follow at the pace needed to sustain that momentum. One quarter’s data point does not a revolution make, but it sure looks like the start of a significant shift in the PC industry’s value proposition.