Just before Google I/O, Google held a preview event for Android 17 at midnight on May 13. Unexpectedly, during the event, Google unveiled an entirely new product line—Android PCs—without prior warning. Unlike Chromebooks, Android PCs are positioned as premium devices with productivity as their core selling point. Google is no longer satisfied with the entry-level market and aims to capture greater share in the PC segment beyond netbooks.
The concept of AI PCs has gained tremendous popularity in recent years, with countless PC chip and device manufacturers highlighting the AI capabilities of their products and repeatedly emphasizing how AI is transforming PC usage scenarios. The emergence of Android-powered computers has unveiled a new approach to AI PCs: one that no longer relies on traditional desktop operating systems, where cloud-based AI is not an add-on but the core, enabling all related functionalities.
(Source: Google)
If Android PCs succeed, cloud PCs are likely to become the definitive solution for the AI era.
Current AI PCs are not yet truly "AI"
Currently, AI PCs in the PC industry are more like traditional PCs wrapped in an AI layer. In terms of chips, Intel and AMD have added dedicated AI computing units to PC processors to enhance their on-device AI capabilities. Regarding systems and ecosystems, terminal manufacturers are building their own AI applications within their operating systems, including proprietary PC management tools and agents, while also integrating external large models.
However, these AI PCs are essentially still traditional Windows computers, with AI serving more as an added enhancement. Moreover, the AI scenarios implemented on AI PCs are mostly based on cloud-based AI, including document summarization and editing, image generation, and various "lobster" tools.
Although chip manufacturers have consistently promoted their chips' local AI capabilities and emphasized scenarios involving heterogeneous computing with CPU, GPU, and NPU for deploying open-source models, in reality, the AI computing power available from consumer-grade PC chips remains severely limited—after all, not every consumer owns a 5080 GPU or has 32GB of memory or more.
(Source: JD.com)
Under these circumstances, a standard consumer-grade PC struggles to truly run large local models and therefore cannot effectively handle more complex AI tasks.
Recently, OpenClaw went viral, causing the Mac mini to sell out and prices to rise. However, most people are using cloud-based models to “raise shrimp,” and various lobster deployment guides frequently mention which AI offers cheaper tokens and how to reduce token consumption.
(Source: Gitbook)
But this raises a new question: if AI PCs still rely on cloud-based AI to deliver AI experiences, what is the actual hardware value of an AI PC?
After all, theoretically, a traditional PC without an AI chip premium can become an AI PC simply by connecting to cloud-based AI over the internet.
In fact, we could go even further by drastically reducing PC hardware specifications—any device with a screen, keyboard, and internet connectivity could become a cloud AI computer. The rapid advancement and widespread adoption of AI seem to be giving this not-so-new concept of “cloud computers” a breakthrough opportunity.
Cloud PCs + AI—are they the future of AI PCs?
Cloud computers are not unfamiliar to us. Several years ago, the booming cloud gaming industry was essentially built on the foundation of cloud computers. At that time, with the widespread adoption of 5G, its low latency and high throughput capabilities were seen as the ultimate solution for popularizing cloud computers.
But the reality is harsh—cloud gaming has never gained significant traction. Google’s cloud gaming service, Stadia, launched in 2019, was shut down less than three years later. According to reviews and user feedback from overseas media, Stadia required extremely high network quality to achieve a smooth experience comparable to local gaming platforms—such as using a wired connection with high-speed broadband at home; even Wi-Fi significantly degraded performance, let alone more volatile mobile networks like 5G.
(Source: Google)
However, cloud gaming is highly sensitive to network latency, while online AI is much more tolerant. As ordinary users, we are already accustomed to AI taking time to "think" when answering questions or completing tasks, and we do not expect immediate feedback from AI as we do with games.
Ultimately, the bottleneck in AI response speed isn't internet speed, but computing power. Even if you install a local large model, it still requires sufficient inference time to generate an answer.
Therefore, we believe that cloud computers are naturally suited for AI PCs. Meanwhile, Google’s Android PCs are building AI PCs using a model distinct from traditional PCs. On Android PCs, AI is not an add-on but a core feature. Google notes that most AI tools today operate as standalone apps, requiring users to copy data into AI interfaces to access AI functionality. In contrast, Android PCs integrate AI throughout the entire system—most visibly, when the mouse pointer moves anywhere, AI appears there, capturing and directly processing information such as text, images, and code near the pointer.
(Source: Google)
Additionally, the implementation approaches for Android PCs are highly diverse. Google primarily provides product concepts and implementation frameworks, while the actual hardware development is carried out by partner manufacturers. According to Google’s announced partnerships, these collaborators are primarily categorized into two groups: chipmakers and device manufacturers. The chipmakers include Intel, Qualcomm, and MediaTek, while the device manufacturers include HP, Lenovo, Acer, ASUS, and Dell.
From a chip brand perspective, it’s clear that Google doesn’t care what architecture chips Android PCs use—whether x86 or ARM. After all, at this stage, AI capabilities on Android PCs still rely heavily on cloud-based Gemini, making local hardware processing power relatively less critical.
In addition, internet and cloud service providers have consistently offered cloud computer services and are evolving toward AI PCs.
Taking Alibaba as an example, in 2024, it launched the Yinying AI Cloud Computer, featuring powerful cloud-based hardware configurations and robust support for large models. By 2026, the Yinying AI Cloud Computer was further upgraded to provide comprehensive support for OpenClaw shrimp farming, enabling one-click deployment, direct integration with Qwen, and seamless connectivity with communication tools such as DingTalk, Feishu, and WeChat.
(Source: Alibaba Cloud)
Another point to note is that AI giants are engaged in a frantic arms race to build AI infrastructure, becoming the primary driver behind rising storage prices. Moreover, there is no sign of storage costs decreasing in the near term. As a result, upgrades to consumer-grade PCs will be further constrained; attempting to develop AI PCs using traditional PC iteration models will become increasingly difficult. Rather than investing heavily in local AI configurations with clearly capped computational power, it makes more sense to offload AI tasks directly to the cloud.
Times have changed—how should PC manufacturers respond?
The AI transformation of PCs is an irreversible trend, and all players across the PC industry are diligently exploring how to board the AI PC wave—each playing different roles and advancing AI PCs in distinct ways.
First are the chip manufacturers, which continue to emphasize AI computing power in consumer-grade chips and build AI use cases around them. More importantly, Intel and AMD are actively pushing forward in the server market, continuously vying for orders from AI giants.
After all, AI manufacturers need to build AI infrastructure, which naturally involves large-scale procurement of AI chips. Besides NVIDIA, the main companies capable of fulfilling these orders are traditional CPU brands like Intel and AMD.
AMD's latest earnings report showed that its "data center" business segment generated $5.8 billion in revenue during the first fiscal quarter, accounting for more than half of total revenue. Additionally, both Intel and AMD are unable to meet order demand, and AMD has already sought assistance from other foundries, such as Samsung, beyond TSMC.
(Source: AMD)
Secondly, there are terminal manufacturers, including traditional PC brands such as Lenovo, ASUS, and HP, as well as emerging brands like Huawei, Xiaomi, and Honor. Currently, their development of AI PCs is primarily based on the traditional architecture of Intel/AMD chips combined with the Windows operating system, enhancing AI capabilities through software integrations such as PC management tools and intelligent agents.
At the same time, smartphone brands have another advantage in the AI PC space: they can integrate PC products with their own hardware ecosystem, including smartphones, automotive systems, wearables, smart home devices, and more, enabling seamless AI functionality across all devices. For example, Xiaomi’s “Super Xiao Ai”—a tool that combines intelligent agents, AI assistants, and voice assistants—can be found across all devices within Xiaomi’s ecosystem.
(Source: Xiaomi)
Additionally, Apple is a unique player in the AI PC space. Although Apple Intelligence was announced early, its implementation has been sluggish, leaving the AI integration on Macs in an awkward position. However, Apple’s advantage in the PC market remains unmatched, thanks to its unparalleled integration of hardware and software and its absolute control over the M-series chips and macOS.
Recently, Apple increased the production of the MacBook Neo from 5 million to 10 million units and has been willing to pay premium prices to maintain the production of the A18 Pro chip. Due to the success of this laptop, according to LuoTu's Q1 online laptop market data, Apple has become the second-largest PC brand in China by market share, behind only Lenovo.
(Source: Loto)
Amid the surge in storage prices, the budget-friendly MacBook has become remarkably appealing. Frankly, the MacBook Neo was initially not expected to succeed and seemed more like a way to clear out A18 Pro inventory. This demonstrates that Apple is capable of creating successful budget PCs. Once it builds a solid user base, a MacBook powered by Apple Intelligence has the potential to catch up and lead in the AI PC era.
Finally, Microsoft, as the dominant player in PC systems, cannot be overlooked. Microsoft’s actions regarding AI PCs primarily involve three areas: defining AI PC hardware standards, system rearchitecture, and hardware architecture diversification.
Microsoft requires AI PCs to have more than 40 TOPS of computing power and more than 16 GB of memory, and has introduced the Windows Copilot Runtime at the system level, integrating multiple small models. Additionally, Windows offers AI-powered features such as live captions and Recall.
(Source: Microsoft)
Another crucial point is that Copilot leverages GPT’s large model technology and Bing’s online capabilities, and is deeply integrated into Windows, Edge browser, and Office 365, fully utilizing its ecosystem advantages. This is primarily achieved through cloud-based AI capabilities.
In conclusion
The emergence of Android PCs challenges the traditional PC form that has remained unchanged for years. It represents an alternative vision for PC development in the AI era: lightweight on local devices, heavy on the cloud.
Today, with storage costs remaining high and local consumer-grade computing power hitting its limits, this approach—breaking down hardware barriers and entrusting core productivity directly to cloud-based large models—is undoubtedly more imaginative.
Of course, the transformation of PC form factors driven by AI has only just begun. Microsoft and traditional PC manufacturers won't sit idle—they continue to emphasize the importance of on-device computing power, while fully integrating cloud-based AI. Apple, too, will continue to capture market share by leveraging its integrated hardware-software ecosystem and downward market strategy. The future PC market will no longer be about mere hardware specification competition, but rather a comprehensive battle centered on cloud empowerment, AI-driven restructuring of the underlying operating system, and cross-device ecosystem integration.
Whether Android PCs can become the ultimate solution still depends on overcoming challenges such as network stability, data privacy, and user habit migration. But one thing is certain: AI has fundamentally reshaped the definition of the PC.
The PCs of the future may no longer require expensive graphics cards or large amounts of memory—just a screen and an internet connection to the cloud can unlock full productivity. A brand-new era of AI-powered cloud computers is on the horizon.
This article is from "Lei Technology".