This summer, Lenovo AI achieved a highlight moment on the World Cup's green field.

As the official technology partner of the FIFA World Cup, Lenovo's ThinkSystem servers and AI infrastructure are operating at the Dallas International Broadcast Center and various stadiums, supporting the event broadcasts and video referee systems for billions of viewers worldwide.

Recently, Gartner updated its Global Supply Chain Top 25 list, with Lenovo Group ranking fifth, entering the global top five for the first time and achieving its highest-ever ranking while maintaining its position as the top supply chain in Asia-Pacific. Lenovo officially released an article to systematically discuss the strength of its supply chain.

These high-profile news stories outline a main theme for Lenovo's AI transformation. Whether it's the Ultimate polishing (extreme refinement) of the supply chain or the stable delivery for major events, each demonstrates that Lenovo is vigorously transforming into an AI infrastructure provider, positioning itself upstream in the AI industry chain.

Servers, storage, data networks... everything operates orderly in the domains most familiar to Yang Yuanqing, stably and predictably.

This stability forms a stark contrast to Lenovo's attitude when it first embarked on AI transformation nine years ago.

In 2017, Yang Yuanqing stood on the Tech World stage, wearing a smart ECG shirt and opening with a holographic projection doppelgänger, declaring to the world that Lenovo would "bet its entire fortune on AI."

That year, Lenovo's financial results were not impressive. Revenue from its Personal Computer and Smart Device business fell 2% year-on-year, while revenue from its Mobile business dropped 10% year-on-year. The mobile business slumped, and the PC business was surpassed by HP. Yang Yuanqing needed to tell the market a new story.

All in AI was pushed forward. To achieve this, Yang Yuanqing proposed a "Three-Wave" strategy: the first wave was to maintain the global leadership and profitability of the core PC business; the second wave was to expand into smartphones and data center businesses, hoping they would quickly become engines for growth and profit; the third wave was to bet on natural language interaction and artificial intelligence to make Lenovo's hardware devices smarter and keep customers engaged with Lenovo after purchasing its devices.

In short, Lenovo did not want to rely solely on selling computers for a living.

To prove this was not just talk, Lenovo released three AI products on the spot: the smart ECG shirt, Smart Speaker+, and the mixed reality headset device Morning Star AR. The newly appointed CTO Rui Yong took the stage to introduce that Lenovo would deploy AI across three dimensions: life, work, and society.

At that time, Yang Yuanqing's attitude toward AI transformation left little room for doubt: "Transformation is never an easy journey for any enterprise. If it's not a transformation that strips several layers of skin, it's not called a transformation."

Unfortunately, most of these products later faded into obscurity. Lenovo's AI transformation did continue, but today's Lenovo AI looks vastly different from what Yang Yuanqing declared in 2017. Over ten years, Yang Yuanqing has gradually scaled back Lenovo's ambitions for AI, embarking on a conventional yet subtly dynamic path.

I. Yang Yuanqing Retreats to the Safe House

Yang Yuanqing was not born conservative.

In 1989, Yang Yuanqing graduated with a master's degree in Computer Science from the University of Science and Technology of China and became one of the first university graduates publicly recruited by Lenovo. He originally wanted to pursue R&D, but Lenovo needed salespeople the most at the time, so he was assigned to sell Sun workstations. A master's graduate in computer science began peddling products door-to-door on a bicycle.

In 1994, Liu Chuanzhi put the 29-year-old Yang Yuanqing in charge of the Microcomputer Business Division. Liu later admitted that entrusting a young man with the heavy responsibility of saving the market was a gamble. Yang Yuanqing did not disappoint, and by 1997, Lenovo had become the top PC brand in the Chinese market.

An even bigger gamble came in 2004. Lenovo acquired IBM's PC business, which had annual revenues of $17.5 billion, for $1.75 billion. Despite significant internal controversy among decision-makers, Yang Yuanqing's persuasive arguments led to the approval of what the media called a "snake swallowing an elephant" merger at the time.

A company with annual revenues of just a few hundred million RMB daring to acquire a multinational giant with revenues exceeding $10 billion laid the foundation for Lenovo's globalization and defined Yang Yuanqing's early career image: daring, aggressive, and willing to stand against consensus.

In the AI era, Yang Yuanqing seems to have adopted a different persona.

Looking back now, the Three-Wave strategy in 2017 may have had a safety margin from the start. The first and second waves seemed like Lenovo's fallback options—even if the AI bet didn't pan out, Lenovo could still rely on PCs and servers.

As it turned out, the second wave of Lenovo's transformation did not go smoothly. After 2017, Lenovo's smartphone business continued to incur losses. The smartphone and data center businesses of the second wave failed to become new growth engines and instead became a drag on financial results year after year.

In the 2017/18 fiscal year (April 2017 to March 2018), Lenovo's Mobile Business Group (smartphones) reported revenue of $7.241 billion, a 6% year-on-year decline, continuing to incur losses after the acquisition of Motorola and falling short of expectations. Moreover, the group's Data Center Business remained unprofitable, with an operating loss of $305 million.

By the 2018/19 fiscal year, the Mobile Business and Data Center Business showed growth but remained unprofitable.

At the time, Yang Yuanqing verbally gave the Three-Wave strategy a score of 80. However, his subsequent actions suggested he was gradually reining in his ambitions.

In the following years, Lenovo's AI narrative underwent two rebrandings.

In 2019, Lenovo established its 3S (Smart IoT, Smart Infrastructure, Smart Verticals) transformation strategy.

By 2023, Yang Yuanqing explicitly proposed Hybrid AI, advocating for the coexistence and complementary use of personal, enterprise, and public intelligence. By then, Lenovo was no longer trying to compete head-on with internet giants in public cloud AI and consumer markets but was instead pulling back to the more familiar domains of edge devices and infrastructure.

To some extent, Yang Yuanqing's background in sales and his past successes stemmed from an extreme focus on predictable businesses—low inventory, fast turnover, and maximizing supply chain efficiency. This was true in the PC era and remained so in the server era. The uncertainty of AI may inherently clash with his business DNA.

In April 2024, Yang Yuanqing proposed the goal of turning around the ISG (Infrastructure Solutions Group) business at the new fiscal year kickoff meeting.

By 2026, ISG delivered impressive results. Lenovo's 2025/26 fiscal year data showed that the group's annual revenue reached $83.1 billion, surpassing $80 billion for the first time. AI-related revenue grew 105% year-on-year. In the fourth quarter, ISG reported revenue of $5.6 billion and an operating profit of $202 million, both record highs for the business. Full-year ISG revenue reached $19.2 billion, turning from a loss of $68.5 million in the previous fiscal year to a profit of $73 million. AI server revenue achieved high double-digit growth, with a year-end project backlog of $21 billion.

At the earnings call, Yang Yuanqing said, "We've concluded the best year in Lenovo's history."

Yang Yuanqing is bringing Lenovo back to its most familiar territory—becoming a hardware supplier enhanced with AI capabilities.

At the 2025/26 fiscal year Q2 earnings call, Yang Yuanqing emphasized that the focus of AI computing is shifting from model training on public clouds to inference deployment in local data centers and edge devices. This statement positions Lenovo similarly to NVIDIA as a "seller of shovels" in the ecosystem.

In the summer of 2026, Lenovo became the official technology partner of the FIFA World Cup. From ThinkSystem servers at the Dallas International Broadcast Center to data processing nodes at various stadiums, Lenovo's AI infrastructure supported the event broadcasts and video referee systems. This was a large-scale demonstration of Lenovo's computing capabilities, with servers, storage, and liquid cooling solutions bundled into an event solution.

Meanwhile, on June 24, the opening day of MWC 26 Shanghai, signals emerged of a backlog of AI server orders. Reports indicated that Lenovo had approximately $150 billion in server orders to fulfill, with products in short supply. The boom in AI servers was expected to continue for some time.

However, as Lenovo vigorously showcased its computing capabilities, ambiguous signals emerged in the industry.

In July 2026, Meta announced plans to sell AI computing power and model access to external customers. As one of the world's largest GPU purchasers, Meta had previously planned a computing cluster of 1.3 million GPUs. The news triggered a correction in AI hardware-related stocks.

Discussions about computing overcapacity directly pressured Lenovo's ongoing AI infrastructure business.

Ablikim Ablimit, Lenovo Group's Vice President and Chief Strategy Officer for China, responded, "In the long term, there is still significant demand for AI computing power in both Chinese and overseas markets, and there is no overcapacity."

This timely response itself conveyed a signal: Lenovo cares deeply about how the market perceives its infrastructure business. Over the past decade, Yang Yuanqing has gradually narrowed Lenovo's AI strategy into a safe house—servers, storage, computing power, infrastructure—all familiar territory for Lenovo.

Within the safe house of infrastructure, Lenovo's business logic is self-consistent and smooth. Demand is predictable, and the business model is certain, allowing Lenovo to maximize the reuse of its decades of accumulated expertise. However, the cost of this safety is that walls surround it on all sides.

Yet inside Lenovo, someone is trying to broaden these walls.

II. Liu Jun Refuses to Be Ordinary

On June 1, 2015, in the evening.

Liu Jun jokingly said on Weibo that he was going to celebrate Children's Day. A few hours later, Yang Yuanqing sent out an email announcing a personnel adjustment, and Liu Jun was relieved of his position as President of the Mobile Business Group. Liu Jun was completely unprepared.

Forty-eight days after his departure, Liu Jun posted his first update on WeChat Moments: "Refreshing!" The accompanying image showed nine different bottles of red wine, all opened and empty. This was Liu Jun's state in the early days after leaving—drowning his sorrows in alcohol. He went to the United States for an extended stay, spending the first few months sleeping long hours almost every day.

But Liu Jun's departure was not solely due to performance. The deeper reason was a significant strategic divergence between his vision for the mobile business and Yang Yuanqing's established goals.

The seeds of this divergence were sown much earlier.

Liu Jun had once brought Lenovo Mobile to its peak. In 2013, Lenovo surpassed Samsung to become the top smartphone brand in China. When there was a quarter when it was expected to overtake Samsung, Liu Jun even worried and discussed with his team whether to "hide our strength" to avoid becoming a target by reaching the top too soon.

However, after the acquisition of Motorola Mobility, strategic differences between Liu Jun and Yang Yuanqing widened. Whether Motorola should re-enter the Chinese market and how to synergize with Lenovo's own brand became increasingly contentious issues, leading to Liu Jun's sudden departure.

But just two years later, the story took a turn.

On May 16, 2017, before Yang Yuanqing declared that Lenovo would bet everything on AI, he personally persuaded Liu Jun to return to Lenovo. In an internal letter, Yang Yuanqing wrote that Liu Jun was a highly accomplished veteran of Lenovo, having joined in 1993 and achieved outstanding results in product development, sales, planning, and other roles.

However, upon his return, Liu Jun was only put in charge of Lenovo's China PC business.

He told the media that he had two missions upon returning to Lenovo: one was to help Lenovo's China PC business reach new heights, and the other was to bring Lenovo's good technologies to market as quickly as possible to help grow more new, non-PC businesses.

Two months later, at Tech World, he stood on stage introducing Morning Star AR and proposed renovating Lenovo's tens of thousands of stores.

After that, on the path of AI, Liu Jun and Yang Yuanqing headed in different directions.

In August 2017, Liu Jun led the China management team to a meeting at the Sunrise Kempinski Hotel Beijing by Yanqi Lake in Huairou. After over 100 days of research and visits, the discussed and summarized direction was named the "Sunrise Eastern" strategy.

Liu Jun had always emphasized to his team, "Just knowing how to sell PCs is a dead end." He proposed that Lenovo China undergo two transformations: one was a customer-centric transformation, and the other was a transformation toward smart products and intelligent services. The core was to break away from the fixed mindset of Lenovo only selling PCs.

Nine years later, the "Sunrise Eastern" strategy continues to operate within the overall framework of Yang Yuanqing's "Hybrid AI."

In March 2025, at Lenovo's new fiscal year kickoff meeting, Liu Jun announced an upgrade to the "Sunrise Eastern" strategy, defining a new mission of "Accelerating China's Intelligent Transformation with Hybrid AI."

He introduced the concept of an "Intelligent Agent Corps," with Lenovo's enterprise super-intelligent agent "LeXiang" serving as a unified gateway, and marketing, sales, and service intelligent agents forming a "Silicon-Based Team." In the personal AI domain, the Tianxi personal super-intelligent agent was upgraded from an "assistant" to a "teammate."

These actions structurally differ from Yang Yuanqing's ISG route—while Yang Yuanqing delivers servers and computing power to customers, Liu Jun attempts to enter customers' actual business processes through intelligent agents and solutions.

However, within Lenovo's overall structure, the China region is just one business unit of the group. Liu Jun can push the "Sunrise Eastern" strategy within China, but resource allocation and strategic direction at the group level remain under Yang Yuanqing's control.

Yang Yuanqing defined 2024 as the beginning of the "AI transformation period," but the actual execution main storyline (main theme) of this transformation seems more like ISG—selling servers, computing power, and storage.

Even though Liu Jun emphasized on multiple occasions that AI cannot just be about selling hardware but must go into industries, scenarios, and applications, given Lenovo's scale and Yang Yuanqing's management style, innovation may exist locally but is unlikely to become the dominant logic.

This reminds one of a similar moment in Lenovo's history.

Around 2000, internal friction arose between Yang Yuanqing and Guo Wei over competing for PC market share. Liu Chuanzhi ultimately chose to split the company, with Yang Yuanqing taking charge of Lenovo PC and Guo Wei leading Digital China to leave. Liu Chuanzhi had hoped for "Yang Yuanqing to lead, with Guo Wei actively cooperating," but the situation ended in a split.

Today, the relationship between Yang Yuanqing and Liu Jun may not go as far as a split, but the tension between one operating within a safe house on predictable businesses and the other trying to forge a new path outside resembles the dynamics of that earlier era.

ISG's demand is predictable, and its business model is certain. Lenovo only needs to leverage its supply chain advantages from the PC era to achieve a predictable outcome. Liu Jun's attempts in the intelligent agent business are more like guerrilla warfare outside the safe house.

As of June 30, 2026, cumulative sales of Lenovo LeXiang exceeded 5 billion RMB, with 7 million monthly active users. However, this accounts for less than 1% of Lenovo Group's approximately 580 billion RMB in revenue for the 2025/26 fiscal year. At least for now, Liu Jun's guerrilla warfare has yielded some results but is unlikely to alter the overall trajectory of Lenovo's AI transformation.

III. Subtle Dynamics Outside the Safe House

Yang Yuanqing's choice is not difficult to understand from a business perspective.

Today, the global AI industry has shifted from training to inference. Yang Yuanqing has positioned Lenovo at this critical juncture. The 2025/26 fiscal year financial data shows that Lenovo's AI server revenue grew 50%, with over 140 billion RMB in order backlog providing high growth certainty.

In this track (sector), Lenovo does not need to compete with OpenAI on models or with NVIDIA on chips. What it needs is to transfer the supply chain, channels, and customer relationships accumulated during the PC era to the AI infrastructure business. After ISG turned around to profitability, its stock price rose, and analysts raised their target prices. Financially, this logic is proving effective.

In the long run, Lenovo's moat needs careful examination.

Lenovo lacks self-developed chips, relying heavily on leading manufacturers like NVIDIA and AMD for the core components of its AI servers. It also lacks self-developed large models, with the core of its "Tianxi Personal Super Intelligent Agent" relying on open-source large models and third-party technologies. Furthermore, Lenovo does not possess a cloud platform like AWS or Azure, and the essence of its ISG business remains the logic of assembling and selling hardware.

In the AI infrastructure industrial chain, Lenovo currently plays more of a role as a traditional IT integrator, leveraging its supply chain and channel advantages to integrate externally procured chips, models, and software components into server chassis for overall delivery to customers.

If this positioning solidifies, Lenovo will face dual uncertainties. It may lose the brand recognition accumulated in the consumer market during the PC era, as enterprise customers are more concerned with cost-effectiveness and stability, with limited brand loyalty. At the same time, it will fail to establish new technological barriers in the AI era.

Today, the global AI trend is rapidly evolving. The recent discussions triggered by Meta's sale of computing power reaffirm a definitive trend: the market's evaluation of computing power is shifting from scale to efficiency. If the industry transitions from a Token maxing phase to a Token minimizing phase, the business model of simply providing computing power hardware may undergo a revaluation.

Ablikim said that there is no overcapacity in computing power. From a long-term demand perspective, this judgment does have its basis. However, short-term market sentiment and industry structural adjustments will not cease due to one or two statements.

Meanwhile, Liu Jun's exploration at the application layer faces a completely different competitive landscape. There are numerous startups in the intelligent agent sector, and the AI PC market includes Apple, Microsoft, and HP. Lenovo's investment scale in this direction, compared to its competitors, resembles a limited exploration.

Therefore, Yang Yuanqing's strategic choice is nearly flawless in terms of business logic: preserve advantages, consolidate positions, and earn the most certain money amidst turbulent new waves. However, a company's perfect choice in business sometimes means it forgoes the potential for greater growth amidst uncertainties.

Does the 62-year-old Yang Yuanqing still have the courage to bet his entire fortune on AI?

In 2017, Yang Yuanqing said, 'I'll bet my life on it.' Back then, Lenovo at least acknowledged that the path of AI was fraught with uncertainties, requiring some adventurous attempts to seek breakthroughs. By 2026, Lenovo has eliminated uncertainties from its strategy. All AI narratives revolve around certainty: certain sectors, certain customers, and certain business models.

Liu Jun is still outside the safe house, attempting to prove the existence of another path. However, how long the guerrilla warfare can last remains an unanswered question.

The name 'Lenovo' originates from the 'Lenovo Chinese Character Card' invented by Ni Guangnan, which suggests a series of phrases after typing the first character. Today's Lenovo AI strategy follows a certain commercial correctness, but looking back at the company's history, the name 'Lenovo' represents the ability to connect and create.