In 2026, the FIFA World Cup will bring 48 national teams together on the global stage. Concurrently, a rare collaborative spectacle has emerged within China's AI industry.

Coinciding with the tournament's opening, LENOVO GROUP (00992), leveraging its Tianxi AI Agent platform, assembled 11 of China's top AI model developers—including DeepSeek, Kimi, Tencent Hunyuan, Qwen, Baidu ERNIE, Zhipu, MiniMax, StepFun, iFlytek Spark, China Mobile Jiutian, and SenseTime's Little Raccoon—to predict the outcomes of all 104 matches.

They faced off against hundreds of millions of human football fans. By the 13th match day, over 20.18 million people had participated in the predictions. On the program "Man vs. Machine: Who is the World Cup Prophet," which premiered on June 24, six human representatives from various fields, including football commentators Yan Qiang and Han Qiaosheng and former national team player Sun Jihai, went head-to-head with this "AI supergroup."

The entity organizing this contest—the company that convinced the fiercest rivals within China's AI sector to appear on a single platform and jointly participate in predicting 104 matches—was not a regulator, a venture capital firm, or a technical standards body. It was LENOVO GROUP.

For the average consumer, this is a technological spectacle for predicting match results. For analysts and institutional investors, however, it more closely resembles a public statement of LENOVO GROUP's positioning in the AI era after four decades in business.

From acquiring IBM's global PC division to consistently ranking first in Asia Pacific on Gartner's Top 25 Global Supply Chain list for multiple years, and now bringing the leading AI giants onto the same stage at its Tech World event, LENOVO GROUP has never aimed for a winner-takes-all strategy on any single technology point of an era. It excels at something else: forming a stable, inclusive system of social productivity and becoming the indispensable connector for all parties.

In the era of PCs and enterprise IT, LENOVO GROUP connected technology, brands, and channels. In the AI era, it connects computing power, devices, and services. Amidst the current intense competition among large models, it extends this capability to bridge models and real-world scenarios—placing 12 AI prediction teams into a single set of rules, a single tournament, and a single user interface through the World Cup "Man vs. Machine" challenge.

This is the simplest way to understand LENOVO GROUP's role in the AI era: it does not build its own large language model. Instead, it positions itself as an AI device gateway and infrastructure platform where more models can be integrated, more users can be reached, and more industry scenarios can be utilized.

Assembling the Domestic AI Models

This marks the first collective appearance of 12 major Chinese AI models within a unified scenario.

On the "Man vs. Machine" program, a collaboration between LENOVO GROUP and Migu Video, China's twelve most powerful AI systems—companies fiercely competing for talent, computing power, and enterprise contracts—simultaneously appeared on a single platform before an audience of tens of millions.

They agreed to participate because LENOVO GROUP offered something none could build alone: a channel to the world's largest sports audience, mediated by a company with no incentive to favor any particular model.

After 44 matches, the latest rankings from the LENOVO GROUP-powered challenge show China Mobile Jiutian in first place with 28 correct predictions, a 63.6% success rate. LENOVO GROUP's Tianxi AI, Qwen, Tencent Hunyuan, MiniMax, and SenseTime's Little Raccoon each had 27 correct predictions, a 61.4% rate. DeepSeek, Baidu ERNIE, and Zhipu each had 26 correct, a 59.1% rate. iFlytek Spark had 24 correct (54.5%), Kimi had 23 (52.3%), and StepFun had 15 (34.1%).

The AI models had their highlights: Qwen successfully predicted the 2-0 score for the Mexico vs. South Africa opener and the 2-1 score for South Korea vs. Czech Republic. In the surprise 1-0 win for Côte d'Ivoire over Ecuador, Baidu ERNIE was the only large model to predict the correct score. StepFun was the sole winner to correctly predict Belgium's draw.

However, the AI also experienced collective failures. For Brazil vs. Morocco, all 12 AIs favored a Brazil win, but the final result was a 1-1 draw. On June 16, for Iran vs. New Zealand, all 12 AIs predicted an Iran victory, yet the final score was 2-2, with none of the 12 AIs getting it right.

Calculated across the 12 AIs and 44 matches, the AI camp collectively made 303 correct predictions, with an overall success rate of approximately 57.4%, higher than the human average of 55.3%. Data shows, however, that humans revised their predictions an average of 1.04 times per match, while AIs revised theirs 1.6 times on average. One individual AI model averaged 3.3 revisions, with a cumulative 220 revisions, and its early accuracy was only 21.9%. The AI's "indecisiveness" may stem from rapid responses to new information or from unstable model weighting.

Beneath the entertaining shell of the "Man vs. Machine" challenge lies a public test of AI's capabilities, limitations, and application conditions. It does not portray AI as an omniscient system; instead, it exposes AI's boundaries in a public setting. For the market, it provides a window to assess AI's productization capabilities.

Creating a User Gateway

Sports prediction is a lightweight application for AI, but it features continuous interaction, high emotional engagement, and a strong feedback loop. Every user choice, revision, and review generates analyzable behavioral data. Similar mechanisms could be migrated to numerous other scenarios in the future: travel, work, shopping, health, education, and financial information services.

As AI shifts from cloud-based training to more on-device and hybrid inference, competition for device gateways is inevitable. PCs, smartphones, tablets, workstations, and edge devices are not just hardware sales targets; they may also become carriers for model distribution, permission management, local data processing, and personal agent operation.

A deeper transformation involves the restructuring of profit within the device market's supply chain. In the past PC and smartphone eras, the majority of the industry's profits were captured by chipmakers and operating system providers. The emergence of AI agents could disrupt this dynamic. The AI explosion puts pressure on traditional software companies, but for LENOVO GROUP, which holds a strong position in PCs, smartphones, and tablets, securing the device gateway and integrating multiple models means gaining leverage it hasn't seen in years.

Bringing Global AI Leaders Together

Extending the AI ecosystem from large models to computing infrastructure highlights the even more unique value of LENOVO GROUP as an integrator.

Starting with its 2024 Tech World event, leaders from NVIDIA, AMD, Intel, Meta, Microsoft, Qualcomm, and others have been part of LENOVO GROUP's AI narrative. These companies occupy critical positions in chips, operating systems, cloud, models, social platforms, and device ecosystems, both cooperating and competing with each other. Their willingness to appear at the same conference underscores LENOVO GROUP's central role in the AI industry chain: the implementer and commercializer of AI technology.

LENOVO GROUP spans both AI devices and AI infrastructure. It has PCs, phones, workstations, servers, edge devices, as well as enterprise services and a global delivery system. For chip and software companies, LENOVO GROUP is not merely a customer but a crucial market channel. It can integrate technology into its product portfolio and then deliver it to enterprise clients and industry scenarios through its channels and services.

At the GTC conference in March 2026, NVIDIA founder and CEO Jensen Huang defined AI as "the largest infrastructure buildout in human history." After his speech, he went directly to the LENOVO GROUP booth and told Chairman and CEO Yang Yuanqing, "This year belongs to you." Behind this intuitive statement was Huang's recognition that the future center of value creation belongs not only to chips and computing power themselves but also to the layer that turns computing power into systems, systems into deliveries, and deliveries into replicable capabilities.

Integrating the Supply Chain

The global supply chain is the physical-world foundation for LENOVO GROUP's role as an integrator.

During the World Cup, the authoritative Gartner Top 25 global supply chain ranking for 2026 was released, with LENOVO GROUP rising one spot to seventh place, setting a new historical best for the company.

Currently, LENOVO GROUP operates over 30 self-owned or partnered manufacturing bases in more than 10 markets globally, works with over 2,000 tier-one suppliers, and processes over 10 million order lines annually. This mature system and operational capability enable LENOVO GROUP to navigate cycles of geopolitical tension, memory price hikes, and maintain uninterrupted global manufacturing and delivery.

LENOVO GROUP's global supply chain is also evolving for the AI era: not only improving its own efficiency in risk warning and dynamic production capacity scheduling but also opening up this intelligent collaborative capability to its vast network of upstream and downstream partners.

"The opportunity for AI democratization does not belong to any single company; it is a shared opportunity for the entire ecosystem," Yang Yuanqing has stated, adding that LENOVO GROUP's new decade of AI is not a solitary journey but one of shared success for the ecosystem.

This statement now has a more concrete footnote: in the World Cup "Man vs. Machine" challenge, 12 competing large models were willing to share the stage, and global chip, software, and cloud providers were willing to participate in the same arena. This fundamentally points to LENOVO GROUP's structural position—within the AI ecosystem, it connects large models, devices, infrastructure, supply chains, and industry clients, and handles the final delivery.