Jensen Huang commented that AI PC sales were below expectations, Lenovo VP responded: The key timing for the AI PC replacement wave is this year
At the 2025 CES, Jensen Huang discussed that AI PC sales did not meet expectations, attributing it to an incomplete AI ecosystem for PC adaptation. A Lenovo VP responded that although sales are stable, the wave of upgrades for AI PCs will arrive in 2025, emphasizing that AI technology is being introduced into PC systems through WSL2, driving market changes
At the 2025 CES, the two most talked-about companies both addressed the core topic of AI PCs.
After the explosive presentation of NVIDIA's RTX 5090 graphics card based on the new generation Blackwell architecture, Jensen Huang completely ignited the atmosphere at the CES opening. However, he still had to respond to the sharp topic of AI PC "sales not taking off" during post-event discussions.
Huang admitted that the market sales of AI PCs in the past year have indeed been disappointing, but he also provided a reason—it's due to the AI ecosystem's adaptation to PCs.
At the same time, Lenovo, the king of PCs, which has garnered attention through a series of new products, also responded to Huang's comments from afar. The person in charge admitted that the sales growth of AI PCs has indeed been relatively stable, and Huang was not wrong, but he emphasized—
The wave of upgrades for AI PCs is coming, with the key time point being 2025.
Is AI PC a false demand?
After NVIDIA's CES press conference, Jensen Huang responded to a series of questions, among which he clearly identified as a "good question" one regarding why AI PC sales did not take off in 2024.
Huang acknowledged the current situation, which is indeed disappointing, or in other words, growth is not as optimistic as previously expected.
However, he attributed this to the lack of a robust AI ecosystem on PC terminals.
Huang stated that AI technology initially developed in cloud environments, and AI models and their deployment applications were mostly conducted through the cloud and API calls.
However, for many potential AI PC audiences, such as designers, software engineers, creative personnel, and tech enthusiasts, the current investment in the AI ecosystem on PC terminals is far less than that in the cloud, and the demand has not been met.
Huang also emphasized that positive changes are occurring, as Windows PCs are bringing AI technologies originally designed for the cloud into PCs through WSL2 (Windows Subsystem for Linux 2), ensuring that AI capabilities available in the cloud are also entering the PC system ecosystem in large quantities, and even the entire cloud AI system ecosystem is being brought to personal computers.
Huang's viewpoint was soon posed as a question to Luca Rossi, the person in charge of Lenovo's AI PCs.
![](https://mmbiz-qpic.wscn.net/mmbiz_jpg/YicUhk5aAGtCHqzk2rl09FvA0nUQicwfRgCL9qLkb1icgsRoC8FJ1r0fbtlI3xgUx8FPfNBOBuYvibVvSFXI6EajLA/640? Luca Rossi, Executive Vice President of Lenovo Group and President of the Intelligent Devices Group (IDG), oversees a business line that includes personal computers, mobile phones, tablets, workstations, and monitors... which is one of Lenovo's core businesses.
AI PC is the core of the core.
After nearly two years of comprehensive layout, AI PC has become a new label "stuck" to Lenovo, and Lenovo cannot remain indifferent to any movements in the AI PC sector.
So how does Luca view Jensen Huang's perspective?
He said Jensen is not wrong, because the PC market in 2024 is indeed relatively stable, and data shows no significant growth.
However, Luca also emphasized that he is not worried about this, as AI PC is approaching an unprecedented catalytic moment. In addition to the edge system ecosystem issues mentioned by Jensen, he also argues for optimism from the "replacement cycle" perspective—
First, the lifecycle of Windows 11 or Windows 10 will end in October 2025, which will inevitably lead to a large demand for device replacements. Additionally, there are 400 million devices purchased during the pandemic, which have now been used for 4 to 5 years and are due for replacement.
More importantly, a series of core capabilities of AI PC are also ready and will complete the final push during the replacement cycle.
Luca stated that the issue facing AI PC is not "whether it will develop," but rather "when it will be fully popularized." He is confident about the performance of the AI PC market in 2025 and 2026, expecting to see significant growth in the AI PC market by then.
Additionally, regarding the application and deployment of AI technology, Luca believes that "hybrid AI" is the definitive trend, where some tasks need to be processed in the cloud, while others need to be completed on edge devices.
To elaborate, when processing low-latency requirements on edge devices or dealing with highly private content, users may be unwilling to upload this content to any cloud and would prefer to complete it locally. Conversely, for image generation that poses no privacy risks and requires significant computing power, the cloud will be utilized.
In summary, Luca believes that the moment when AI PC is both "well-received and profitable" is continuously approaching.
**In this process, Lenovo's PC market share will continue to be enhanced. Luca revealed that currently, one in every three activated Windows computers globally is a Lenovo computer, showing very positive momentum **
At the same time, Lenovo's PCs rank first in both the consumer and commercial markets, with the gap to the second place continuing to widen.
Luca predicts that the penetration rate of AI PCs will reach 40%-50% in the next two to three years, and could even be as high as 80%. It is foreseeable that in the next three to four years, every PC will become an AI PC.
This Lenovo VP also shared tactical approaches, clearly stating that he will not engage in a price war, but will instead offer a diverse range of innovative products to meet different layers of demand.
He believes the industry is undergoing changes. Compared to five years ago, today's technology choices are more diverse: from having only one or one and a half chip suppliers to now having three to four options; from X86 architecture to ARM architecture; and with the addition of AI Agents, this will be a whole new battlefield. Lenovo believes that differentiation can be achieved in this new world. Users can choose different chips, AI Agents, and even innovate more in device forms.
Luca also rarely showed confidence: which competitor has such a wide product line as Lenovo? From Windows to Chrome, to Android, from 6-inch screens to 30-inch all-in-one devices; from portable devices to enterprise-level cloud infrastructure solutions.
He said Lenovo is fully prepared to embrace the AI revolution.
However, there are still fundamental questions being raised on-site—
Are AI PCs or AI phones a real demand?
Luca smiled. He said he does not believe this is a pseudo-demand, and the reason is simple: AI PCs and AI phones can give humans "superpowers," right? They can allow you to accomplish tasks that previously took 25 minutes or even 2 hours in just 2 minutes. When you press a button, the machine can complete a PPT that used to take 2 hours in just 2 seconds; this is a kind of human "superpower."
It is hard to imagine such "superpowers" being a "pseudo-demand."
Will AI make PCs say goodbye to keyboards??!
Luca was also asked about how AI will continue to transform PCs.
He believes that the current pace of innovation makes many things difficult to predict, with new things and new species constantly emerging.
But for PCs, there is one trend he is very certain about: he believes that over time, the current limitation of PC form and development—the "keyboard"—may disappear. He believes that with the development of natural language technology, people will no longer have to rely on keyboards, which will unleash the potential for innovation in device forms.
In addition, Agents will also be a definite trend.
Luca hinted that next month there will be an Agent with mixed orchestration capabilities, and users' personal knowledge bases will be stored locally on the device.
This Agent, or entry point, has already been externally showcased through the "XiaoTian" application. Luca further clarified that he hopes XiaoTian can become the preferred entry point for human-computer interaction, allowing users to interact with the internet, ask questions, and complete all these operations through "XiaoTian."
Moreover, glasses are also his appealing AI terminal platform.
However, Luca stated that this platform may not mature by 2025, as there are currently many obstacles, such as battery life, thermal management, performance, and the design of the device form. These are all issues that need to be addressed.
He also believes that lessons should be learned from past experiences. For example, years ago, 3D televisions were considered a global trend, and everyone thought that in the future, everyone would buy televisions with 3D glasses. But it turned out to be one of the biggest failures in the industry at that time. The reason was that people were unwilling to wear bulky and uncomfortable devices.
If we want glasses to be very lightweight and comfortable, making people willing to wear them, the required technology is far beyond today's level.
By 2025, AI glasses will have made some progress, but the market will still be very small, just getting started. Some application scenarios may be more mature, while others will be less mature.
For example, a more mature application scenario might be productivity scenarios when performing specific tasks in factories or on assembly lines. Because in these scenarios, wearing glasses is not for entertainment, but to complete work tasks. You might only need to wear them for an hour, but they can significantly enhance your work value. So such scenarios are likely to be realized.
Another scenario targets "geek" users. The number of such users will not reach millions, but may only be hundreds of thousands, yet they are very passionate about this technology, and even if the devices are not particularly comfortable, they are willing to use them.
So if asked whether AI glasses can compete with smartphones and become a major platform for hundreds of millions of users?
Luca's answer is affirmative, but not in 2025 or 2026. That is a more long-term vision that requires technology to truly mature.
Risk Warning and Disclaimer
The market has risks, and investment should be cautious. This article does not constitute personal investment advice and does not take into account individual users' specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances Invest based on this, and you bear the responsibility