Google and Microsoft Battle for AI PC Supremacy: Is the Cloud the Ultimate Future?

Google and Microsoft Clash Over AI PCs: Is Local Compute a Premium Gimmick While Cloud PCs Represent the Ultimate Form?

Just ahead of the google I/O conference, Google held a pre-briefing for Android 17 in the early hours of May 13. Unexpectedly, the event saw the company quietly unveil an entirely new product line: Android PCs. Distinct from Chromebooks, Android PCs are positioned at the higher end of the market and place Productivity at the core of their value proposition. Google is clearly no longer satisfied with dominating the entry-level segment and is moving to seize a larger share of the PC market beyond simple netbooks.

The concept of the AI PC has gained significant trACTion in recent years. Countless PC chip and terminal manufacturers have been emphasizing the AI capabilities of their products, tirelessly touting the new changes AI brings to PC usage scenarios. The emergence of Android PCs, however, presents the industry with a new blueprint for the AI PC: one that no longer relies on traditional desktop operating systems, where cloud AI is not just an accessory but the central core, deriving all related functionalities from this foundation.

If Android PCs succeed, the cloud computer could very well become the DeFinitive answer for the AI era.

Current AI PCs Are Not "AI" Enough
The AI PCs currently prevalent within the industry more closely resemble a traditional PC wrAPPed in an AI veneer. On the hardware side, both Intel and AMD have added independent AI computing units to their PC processors to bolster On-device AI capabilities. On the system and ecosystem side, terminal manufacturers are building their own AI applications within the OS, including proprietary PC managers and intelligent Agents, while also integrating external large language models.

However, these AI PCs are fundamentally still traditional Windows machines; AI acts merely as an enhancement, the icing on the cake. Moreover, the vast majority of AI scenarios realized on these AI PCs rely on cloud-based AI, including document summarization and editing, Image Generation, and various "agent" tools.

Although chip manufacturers consistently promote the local AI capabilities of their Silicon and emphasize the deployment of open-source models using CPU, GPU, and NPU heterogeneous computing, the reality is that the AI computing power offered by consumer-grade PC chips remains inherently limited. After all, not every consumer owns a RTX 5080 graphics card or 32GB of RAM as Standard.

Under these circumstances, an aveRAGe consumer PC struggles to truly run large-parameter local models and is consequently unable to handle genuinely complex AI tasks. The recent viral sensation of openclaw, for instance, led directly to the Mac mini selling out and experiencing price hikes. Yet, the vast majority of users are relying on cloud models to run their "clAWS," with various deployment tutorials frequently mentioning which AI provider offers cheaper tokens and how to minimize Token Consumption.

This situation gives rise to a new question: If AI PCs still depend on cloud AI to realize their AI scenarios, then what exactly is the value proposition of the AI PC hardware itself? Theoretically, a traditional PC without the added premium of an AI Chip could be transformed into an AI PC simply by connecting to the internet and accessing cloud AI. We could even be more radical and drastically strip down the PC’s hardware configuration; as long as it has a screen, a keyboard, and network connectivity, it could function as a cloud AI Computer. The rapid development and proliferation of AI seem to offer a breakout opportunity for the cloud PC—a concept that is not exactly new.

Cloud PC + AI: The True Future of the AI PC?
The cloud PC is not an unfamiliar concept. The cloud gaming boom of a few years ago was essentially realized through a form of cloud PC. At that time, the full-scale rollout of 5G, with its low latency and high throughput characteristics, was viewed as a panacea for popularizing cloud computers. Reality, however, proved less promising; the cloud gaming concept never truly ignited. Google's own cloud gaming service, Stadia, launched in 2019, was shut down in less than three years. According to media reviews and user feedback from overseas, achieving a smooth experience comparable to a local gaming platform on Stadia demanded extremely high network quality—requiring a wired connection to a local high-speed broadband, with even Wi-Fi significantly degrading the experience, let alone the more variable 5G mobile network.

But while cloud gaming is highly sensitive to network latency, online AI is far more tolerant. As ordinary users, we have grown accustomed to AI needing a moment to "think" when answering questions and processing tasks, and our expectations for AI-generated results are not as immediate as our reactions in a game. Ultimately, the bottleneck for AI response speed lies not in network speed but in computing power. Even if you install a local large language model, it still requires sufficient inference time to generate an answer.

Therefore, we believe that the cloud PC form factor is inherently suited for AI PCs. Google’s Android PC is crafting an AI PC using a model distinct from traditional PCs. On an Android PC, AI is not an accessory but the core functionality. Google has stated that most AI tools today are standalone apps requiring users to copy data into the AI interface to utilize its features. The Android PC, however, integrates AI deeply into every corner of the system. Most intuitively, the AI appears wherever the mouse pointer goes, capturing Information like text, images, and code near the pointer to process and manipulate it directly.

Moreover, the implementation pathways for Android PCs are highly diverse. Google provides the product vision and realization Framework, while the hardware itself needs to be built by partner manufacturers. According to Google's announced list of partners, they fall into two main categories: chips and terminals. The former includes Intel, Qualcomm, and MediaTek, while the latter includes HP, Lenovo, Acer, Asus, and Dell. A look at the chip brands reveals that Google is indifferent to whether an Android PC uses x86 or Arm architecture. This is because, as it stands, the realization of AI scenarios on Android PCs still heavily depends on the cloud-based Gemini, making the on-device hardware's computing power relatively less critical.

Furthermore, internet and cloud service providers have been continuously offering cloud PC services and are evolving them in the direction of AI PCs. Alibaba, for example, launched its Wuying AI Cloud Computer in 2024, which not only boasted powerful cloud hardware configurations but also provided robust support for large models. By 2026, the Wuying AI Cloud Computer underwent further upgrades, offering comprehensive support for deploying OpenClaw agents, one-CLIck deployment, direct integration with Tongyi Qianwen, and connectivity with communication tools like DingTalk, Feishu, and WeChat.

It is also worth noting that AI giants are engaged in a frenzied arms race over AI infrastructure, acting as a primary culprit behind rising storage costs. Furthermore, there is no short-term prospect of storage prices dropping. This dynamic further constrains hardware upgrades for consumer-grade PCs. If the industry persists with the iterative model of traditional PCs to build AI PCs, progress will become increasingly difficult. Rather than piling on the high cost of localized AI Configurations with a clear ceiling on compute power, it is more straightforward to delegate AI tasks directly to the cloud.

Times Have Changed: How Should PC Manufacturers Respond?
The AI transformation of the PC is an irreversible megatrend. Players across the entire PC industry chain are racking their brains, contemplating how to board the AI PC ship. Their roles vary, and so do their approaches to advancing the AI PC.

Firstly, chip manufacturers continue to stress the AI computing power of consumer-grade chips and build AI scenarios around them. More importantly, both Intel and AMD are making sustained efforts in the server market, aggressively pursuing orders from AI giants. After all, AI companies require massive procurement of AI chips for their infrastructure buildout. Beyond NVIDIA, the remaining major players capable of fulfilling these orders are the traditional CPU brands like Intel and AMD. AMD's latest financial report revealed that its data center segment contributed $5.8 billion in revenue in the first fiscal quarter, accounting for over half of the total. Moreover, both Intel and AMD are struggling with production capacity unable to meet order volumes, with AMD already seeking assistance from other wafer foundries like Samsung, beyond TSMC.

Secondly, there are the terminal manufacturers, including traditional PC brands like Lenovo, Asus, and HP, as well as emerging players like Huawei, Xiaomi, and Honor. Currently, their development of AI PCs largely relies on the traditional architecture of Intel/AMD chips plus the Windows system, enhancing a PC’s AI capabilities by embedding software such as PC managers and intelligent agents. In the realm of AI PCs, smartphone brands also possess a distinct advantage: they can seamlessly connect PC products with other diverse devices within their hardware ecosystem, such as phones, vehicles, wearables, and smart Home Appliances, allowing AI capabilities to flow effortlessly across devices. Using Xiaomi as an example, its "Super Xiao Ai"—a tool integrating the roles of intelligent agent, AI assistant, and voice assistant—can appear on various devices within the Xiaomi ecosystem.

Apple, meanwhile, is a unique entity in the AI PC arena. Apple Intelligence was announced early on, but its rollout has been disAstrously slow, leaving the AI integration of the Mac in a rather awkward position. Apple's advantage in the PC domain remains its unparalleled integration of software and hardware, with absolute control over its M-series chips and the macOS operating system. Recently, Apple increased the production of the MacBook Neo from 5 million to 10 million units and is sustaining production of the A18 Pro chip at high cost. Driven by the success of this notebook, according to data from Runto Technology for Q1 online notebook sales, Apple has become the second-largest PC brand in the Chinese market, trailing only Lenovo.

Against the backdrop of soaring storage prices, the budget-friendly MacBook has shown astonishing appeal. Frankly speaking, the MacBook Neo was not initially favored and was seen more as a means to clear out A18 Pro chip inventory. This reflects Apple's capability to create a successful, low-cost PC. Once it establishes a solid user base, a MacBook enhanced by Apple Intelligence has the potential to catch up from behind and take the lead in the AI PC era.

Finally, Microsoft, as the gatekeeper of the PC operating system, cannot be overlooked. Microsoft’s strategy for the AI PC encompasses three main aspects: defining hardware standards for AI PCs, restructuring the system, and diversifying hardware architecture. Microsoft mandates that an AI PC must possess over 40 TOPS of compute power and at least 16GB of RAM. It has introduced the Windows Copilot Runtime deep within the Windows operating system, integrating multiple Small Language Models. Concurrently, Windows now provides AI-powered features such as live captions and Recall.

Crucially, Copilot leverages the large model technology of GPT and Bing's internet connectivity, integrating deeply into the Windows OS, the Edge browser, and Office 365, fully exploiting Microsoft's ecosystem advantages. This, significantly, is primarily achieved through cloud-based AI capabilities.

Conclusion
The emergence of the Android PC poses a challenge to the conventional PC form factor that has remained unchanged for years. It represents a different product philosophy for PC development in the AI era: light on local resources, heavy on the cloud. In an era of persistently high storage costs and bottlenecks in local consumer-grade computing power, this approach—breaking down hardware bARRiers and directly delegating core productivity to cloud-based large models—undoubtedly possesses greater imAGInative potential.

Of course, the transformation of the PC form factor catalyzed by AI is only just beginning. Microsoft and traditional PC manufacturers will not stand idly by; they are still emphasizing the importance of on-device computing power, even as they comprehensively integrate cloud AI. Apple, too, will leverage its tightly integrated ecosystem and a strategy of market entry with lower-cost models to continue competing for market share. The PC market going forward will no longer be an arms race confined purely to hardware specifications. Instead, it will be a comprehensive contest involving leveraging the cloud, architecting fundamental AI capabilities into the OS, and competing across device ecosystems.

Whether the Android PC will become the ultimate definitive answer still needs to withstand the tests of network stability, data privacy, and user habit migration. But what is certain is that AI has already profoundly reshaped the very definition of the PC. The PC of the future might truly no longer require an expensive graphics card and massive mEMOry; just a screen and a network connection to the cloud could be enough to unleash productivity. A whole new era of the AI-powered cloud PC is marching towards us.