Lenovo (00992.HK) leads the formation of L-MAN super team: the world's strongest AI computing power alliance assembles

This year’s Lenovo Tech World 2023 brings together AI leaders from all walks of life and brings a variety of breakthrough AI technologies.

But the most eye-catching thing is undoubtedly the cooperation between giants such as Lenovo, Microsoft, AMD, and NVIDIA to build the strongest alliance for AI computing power in the world - L-MAN ( Lenovo-Microsoft-AMD-NVIDIA).

As a core member of this alliance, Lenovo is actively participating in this unprecedented global AI computing cooperation and setting off a wave of AI computing called "AI for All".

L-MAN: The new power of global AI

It is undeniable that in the field of science and technology, cooperation is the engine of innovation.

Just like the members of the X-MAN team are all superheroes, the members of the L-MAN Alliance are all "superheroes" with different extraordinary abilities in the field of AI computing, and their respective characteristics and strengths complement each other. As we all know, Lenovo has outstanding expertise in the field of equipment and infrastructure, Microsoft in the field of AI applications and operating systems, AMD in the field of processor technology, and NVIDIA in the field of GPU technology. These global technology giants have pooled their respective strengths to collaborate to address opportunities and challenges in the AI ​​field. This should be the world's top AI computing power team.

At Tech World, members of the L-MAN team appeared one after another, not only demonstrating how the world's top AI computing power teams work closely together, but also announcing a number of important strategic cooperation agreements.

L: Lenovo provides comprehensive AI computing power infrastructure based on the "intelligent-device-cloud-edge-network" technical architecture. Lenovo is gradually becoming the world's largest end-to-end infrastructure provider. In the PC field, Lenovo is undoubtedly the world's number one player, and demonstrated AI PC and related technologies for the first time; in the server and storage fields, Lenovo has also achieved great success, with its market share ranking among the top in the world.

M: As Microsoft Chairman and CEO Satya Nadella said, Microsoft not only provides powerful operating systems and office software, but also cooperates with Lenovo to develop AI PCs to make the integration of hardware and software closer. The powerful combination will provide users with a smarter and more efficient computing experience and help them better embrace the new era of AI.

A: AMD CEO and Chairman Su Zifeng pointed out that AI requires huge computing power. AMD continues to invest in high-performance GPU computing and has in-depth cooperation with Lenovo in the fields of smart devices, smart infrastructure, and artificial intelligence solutions. From ThinkSystem in data centers to ThinkStation workstations and ThinkPad laptops, AMD has achieved major breakthroughs. Recently, Lenovo released the ThinkEdge SE455 V3 server using AMD's latest EPYC 8004 series processors. AMD EPYC processors combined with Lenovo’s engineering create the most AI-powered and energy-efficient edge servers on the market.

N: At the meeting, Lenovo Group Chairman Yang Yuanqing and NVIDIA CEO Jensen Huang jointly announced that Lenovo Group will cooperate with NVIDIA to open up the digital era for enterprise users and provide new enterprise-level AI solutions based on NVIDIA MGX architecture and A series of Lenovo hybrid artificial intelligence services. Whether on edge devices or in the cloud, enterprise users will be able to increase flexibility and efficiency and easily deploy, manage and optimize generative AI applications. Huang Renxun also said that following the cooperation in the past 25 years, Nvidia and Lenovo will work together to create the next 25 years.

In short, Lenovo, Microsoft, AMD and NVIDIA share resources, knowledge and experience to promote continuous breakthroughs in AI technology. On this basis, L-MAN, the AI ​​computing power alliance, has brought together the resources and expertise of major technology giants to build a global super-powerful technology ecological service platform around AI computing power, and provide Lenovo's "AI for All" Vision lays a solid foundation.

AI for all

How to understand Lenovo’s vision of “AI for all”? Literally, this means making AI computing more powerful and more accessible. But it's much more than that.

Judging from the Tech World display, the "AI for all" pursued by Lenovo actually makes AI "all the time, everywhere, and available to everyone." Although it sounds like empty marketing slogans, in practical applications and industries, this is concretely feasible, albeit with huge challenges.

The most popular public large model is the "universal" large model. Although the general large model has generalization far beyond the mainstream AI technology of the previous generation, the goal of "for all" has not yet been achieved. Mainstream public large models can only be run in the cloud based on large GPUs and require a transmission network, so they cannot achieve “AI all the time”; “public” means that it is difficult to apply in scenarios with high privacy and security requirements, and cannot achieve “AI everywhere”. AI” vision. High generality is often accompanied by low personalization, and the vision of everyone having their own AI is difficult to realize.

However, Lenovo has proposed a new hybrid AI framework based on three core basic models, which can promote the application of artificial intelligence in multiple fields and meet diverse needs. This framework covers three core foundation models: public foundation model, private foundation model and personal foundation model.

The public (Public) basic model is the most familiar model. It is trained through large-scale public domain data and deployed on the public cloud for everyone to access and use. Private models are usually tailored to a specific enterprise and can handle not only public tasks, but also tasks related to the enterprise's specific business. Private base models typically do not transfer data into the public domain to better ensure data security and privacy. The personal (Personal) basic model usually runs on the user's device to ensure privacy and data security. It can not only understand and perform personal tasks, but also provide personalized services.

The combination of these three basic models constitutes Lenovo’s hybrid AI framework. This framework not only expands the scope of AI applications, but also ensures data privacy and security, and further creates personalized models to meet the diverse needs of different users.

This hybrid AI framework contributes the basic architecture to the "AI for All" vision. Take the PC field as an example: With the support of this framework, Lenovo AI PC can provide more comprehensive AI solutions for enterprises and individuals. Lenovo's personal basic model running on personal-grade PCs can provide a smarter user experience, helping users solve problems and provide personalized recommendations. In the field of enterprise-level PCs, private base models can efficiently handle enterprise-specific tasks and information. Lenovo promises that data will never be shared or sent to the public cloud unless authorized by the user to ensure personal privacy and data security.

The tip of the iceberg of AI PC

Intelligent devices with PC as the core are the final key link of "AI for All".

Yang Yuanqing pointed out that smart devices are like racing cars. They are the ultimate carrier for artificial intelligence to reach end users. The personal computers of the future will be AI PCs, the mobile phones of the future will be AI mobile phones, and the workstations of the future will be AI workstations. At Lenovo Tech World 2023, Lenovo demonstrated its "family bucket" of AI smart devices, including the first AI PC, AI Phone (Moto Edge 40 Pro), and AI workstations.

Among them, Lenovo AI PC is considered by many industry insiders to be a revolutionary product in the AI ​​era, filling the gap in artificial intelligence computing power for people's life and work scenarios.

We already know from the above that the hybrid AI framework provides a powerful framework for Lenovo’s PC products and solutions, but things are not that simple – as Soochow Securities Research Institute pointed out, AI PC is implemented on the PC user side Some AI applications are very bold innovations.

The issue of computing power is one of the most critical. Generative AI requires powerful computing power, so it could only be run on large GPUs in the cloud. How can personal computers provide services comparable to large-scale models in the cloud?

The solution proposed by Lenovo is to use large model compression technology. . The so-called large model compression technology is to reduce the size of the model and the demand for hardware resources by optimizing the model, thereby improving the inference speed of the model. This is large model compression technology. Yang Yuanqing said at the conference, "Lenovo's large model compression technology allows users' own smart terminals and devices to have the ability to run personal-level large models."

In addition, Lenovo has put AI algorithms into PC devices a year ago. In the future, Lenovo’s AI PC will further introduce NPU. NPU (Neural Processing Unit) can effectively handle large-scale parallel operations, conforms to the rules of deep learning calculations, and is the most suitable processor for AI deep learning and calculations. Lenovo integrates CPU, GPU and NPU to form a multi-computing unit computing power combination that combines the advantages of general computing and heterogeneous computing to meet complex computing needs and achieve a faster and smoother user experience while reducing costs. reduce power consumption and extend battery life.

At the scene, Lenovo also demonstrated the functional product Artificial Intelligence Twin (AI Twin). As the name suggests, this is a large model that belongs to each person and is an extension of the individual in the digital world. This product is imbued with personalized personal characteristics, just like twins. In the demonstration of personal-level AI Twin, we clearly saw that the answer generation speed of AI Twin is comparable to that of general-purpose large models. Unlike general large models, AI Twin generates personalized answers based on a personalized local knowledge base and the user's way of thinking. Of course, if used on enterprise-level PCs, it can generate personalized solutions based on the policies and information of different companies.

These are only some of the applications of AI PC. More functions still need to wait until AI PC goes on sale in September next year. Currently, the market and users are full of expectations for AI PC. Linn Huang, vice president of device and display research at IDC, said that although the case is not yet fully clear, market interest in this category (AI PC) is already strong.

Yang Yuanqing said that the initial positioning of AI PC is the high-end market, the volume will not be large, and the proportion will not exceed 10% of the total industry sales. Although AI PC may not immediately become a super hot product and the mainstream of the market in the early days, in the long run, it has huge potential.

It is worth noting that AI technology is being rapidly developed and promoted, which is likely to lead to a larger PC replacement cycle. Goldman Sachs expects AI's higher requirements for computing power and new AI features to support demand for Lenovo's terminal devices.

Currently, the global PC market is ending its downturn for several consecutive quarters. Lenovo pointed out that the current PC destocking management has basically ended. Lenovo's PC shipments have increased quarter-on-quarter in the past two quarters, especially in the just-ended third quarter, which increased by 11% quarter-on-quarter. Year-over-year growth is expected to begin in the upcoming fourth quarter.

As the PC market picks up, Lenovo Group’s stock price has experienced a wave of recovery since the fourth quarter. But more catalytic factors may remain unconsidered.

The upcoming turning point of the market cycle, coupled with the increasing AI, will stimulate the demand for machine replacement; there is also the upcoming release of AI PC, which will not only attract attention, but also drive the "volume and price" of the business to rise... ..Lenovo’s AI potential may be in the initial stage of release.

The above is the detailed content of Lenovo (00992.HK) leads the formation of L-MAN super team: the world's strongest AI computing power alliance assembles. For more information, please follow other related articles on the PHP Chinese website!