AI PC Revolution a Bust? Analyst Declares Supercycle Dead on Arrival

A leading market analyst has noted that Micron’s disappointing Q3 earnings and its future outlook for the next fiscal year might indicate that the anticipated boom in AI-driven PCs and smartphones may not materialize. The company’s struggles are primarily due to a softer than expected market for memory products in the PC and smartphone sectors. Additionally, various reports from other market research firms suggest that the expected AI PC revolution has not taken off as predicted.

In its recent financial disclosure, Micron, a major American memory producer, reported a Q3 revenue of $8.709 billion, slightly below the expected $8.721 billion. More concerning was Micron’s projection for the second fiscal quarter of 2025, forecasting revenues of only $7.9 billion against the anticipated $8.98 billion, causing a more than 16% drop in its stock price at the time of reporting.

Daniel Newman, a semiconductor analyst, referred to the shortfall and the forecast as a “huge miss” in a post on X. He emphasized that this is not the “end of the road for AI investments,” highlighting companies like Nvidia, which has seen significant growth to become one of the world’s largest firms due to the rise of AI chips earlier this year.

While high-bandwidth memory (HBM) is poised to become a significant market for Micron, expected to grow from a $16 billion market this year to $100 billion by 2030, the company’s main revenue stream still comes from manufacturing memory chips for PCs and smartphones.

Newman pointed out, “However, the core business is shrinking as PC and smartphone sales decline AND Micron is grappling with customer inventory that is moving slowly, leading to lower booking/sell-through this quarter and the next,” and concluded, “The bad news is that the AI PC and AI smartphone ‘supercycle’ has largely been a letdown.”

In 2024 and 2023, there was considerable optimism that the PC industry would be revitalized by the demand for AI-equipped PCs, thanks to their new AI-powered features. However, it appears that these expectations have not been met. A report by IDC Research in September indicated that AI is not driving the demand for AI PCs, which is instead fueled by the general desire for hardware upgrades, as new chips also come with improved CPU and GPU cores.

Trendforce also released a report last month with a similar finding: customer interest in AI PCs is lacking. The firm predicted that the transition to Windows 11 PCs from older Windows 10 systems would likely spur more sales than AI features in the coming year.

Qualcomm is also facing challenges with its new Snapdragon X chips for Copilot+ laptops, as demand appears to be weak for AI PCs. In the third quarter of this year, Qualcomm only managed to secure 0.8% of the PC market, selling just 720,000 units. Intel’s interim co-CEO Michell Johnston Holthaus reported last week that return rates for Snapdragon X laptops were relatively high, an assertion Qualcomm has denied.

It seems Qualcomm is looking forward to the release of upcoming Snapdragon X PCs that aim to lower the entry price to $700 without sacrificing performance on the NPU, a critical component for AI capabilities on Qualcomm’s PC chips. The cheapest Snapdragon X laptops currently cost $1000, so a $300 price drop is significant.

However, if there were strong demand for AI PCs, it would seem premature for Qualcomm to shift focus to lower price points so quickly, especially if these upcoming PCs maintain the high-end NPU found in the Snapdragon X Elite and Plus models. Focusing on more budget-friendly market segments makes sense if there’s a lack of interest in high-end AI PCs, but it also reveals Qualcomm’s struggles to scale up.

The AI PC faces significant challenges because there’s currently little need to run AI locally on a PC. AI software that operates directly on a user’s device tends to attract only tech enthusiasts. Meanwhile, popular services like ChatGPT are cloud-based and don’t necessitate a brand-new AI PC.

As AI capability becomes a standard feature in PCs, having AI hardware will no longer be a unique selling point, similar to how multicore CPUs, integrated graphics, and solid-state drives have become commonplace. Consumers are unlikely to pay extra for an AI PC, but they also won’t likely purchase a non-AI PC, as that would mean using outdated technology that predates the AI revolution. In essence, AI will become a basic expectation.

However, the waning interest in AI PCs and smartphones doesn’t necessarily signal that the entire AI industry is on the brink of collapse. “The good news is that HBM is selling well and expanding rapidly, indicating that demand for AI chips remains strong,” Newman asserts. “Companies like NVIDIA, Broadcom, AMD, Marvell, etc., need not panic. HBM is on a rapid growth trajectory.”

While it’s not particularly positive news that consumers aren’t excited about running AI locally on PCs and smartphones, there remains a significant market for AI in the cloud, especially since mobile devices readily facilitate connectivity to online services. A decline in demand for HBM would be a far more concerning indicator of trouble, as HBM is crucial for AI chips designed for data centers.

Yet, if AI proves to be a bubble that eventually bursts, this could be an early warning sign: Currently, there is no explosive demand for locally run AI on end-user devices.