Lenovo just closed the strongest fiscal year in its history, and the reason is no longer just PCs. AI servers, services, and liquid-cooled infrastructure are now pushing the company toward a much larger role in the data center build-out.
Lenovo has spent years being treated as a PC story with a server business attached. Its latest results make that view harder to defend. The company reported record annual revenue of $83.1 billion for fiscal 2025/26, up 20 percent from a year earlier, while adjusted net income rose 42 percent to $2 billion. The clearest signal came from AI-related revenue, which doubled for the full year and accounted for 33 percent of total group revenue.
According to Lenovo's May 22 earnings release, AI-related revenue grew 84 percent in the fourth quarter alone and represented 38 percent of group revenue. That is the kind of mix shift investors usually associate with chipmakers, not a company still best known by many consumers for ThinkPads. But AI demand is not only about GPUs. It is also about the systems, racks, cooling, storage, deployment services, and supply chains that turn expensive chips into working infrastructure.
The Infrastructure Solutions Group is where that change is most visible. ISG posted record full-year revenue of $19.2 billion, up 32 percent, and reached full-year profitability with operating profit of $73 million. In the fourth quarter, ISG revenue hit $5.6 billion, up 37 percent, with operating profit of $202 million. That matters because Lenovo had taken a $285 million restructuring charge in the previous quarter to reposition the business around AI training and inference demand. The early evidence suggests that reset is already paying off.
Lenovo is building around the AI bottleneck
The usual AI infrastructure discussion starts with Nvidia and stays there. Lenovo's results point to a different part of the market: the physical systems that let those chips run at scale. The company said its AI server pipeline now stands at $21 billion, with more than 5,800 customer AI deployments. It also shipped its first GB300 NVL72 racks last quarter and says Rubin-based platforms remain on track for the second half of 2026.
Those details are important because customers are not just experimenting with AI workloads anymore. They are trying to move them into production. That changes what they need from suppliers. A lab cluster can be assembled around availability. Production infrastructure needs cooling, service contracts, predictable deployment, and enough manufacturing capacity to scale without creating new delays.
Lenovo says its annual server manufacturing capacity now exceeds 70,000 racks across AI, compute, and storage systems, including more than 11,000 direct liquid-cooled racks built for AI workloads. Its Neptune liquid-cooling business grew 300 percent year over year. That is not a cosmetic product line. Dense AI racks generate heat that traditional air-cooled data centers increasingly struggle to manage, and cooling is becoming one of the practical limits on AI expansion.
The PC business still pays the bills
None of this means Lenovo has stopped being a PC company. The Intelligent Devices Group remains the anchor, with full-year revenue of $58.9 billion, up 17 percent. In the fourth quarter, Lenovo's global PC market share reached 24.4 percent, and the company widened its lead over the number two vendor. Premium PCs also made up 50 percent of fourth-quarter shipments, showing that the company is not only defending volume but also improving mix.
That gives Lenovo an advantage many infrastructure specialists do not have. The PC business provides scale, cash flow, supplier leverage, and a global customer base. The AI infrastructure business provides the growth story. Together, they give CEO Yuanqing Yang a credible path toward his stated ambition of turning Lenovo into a $100 billion company within two years.
The services business adds another layer. The Solutions and Services Group passed $10 billion in annual revenue for the first time, up 19 percent, with operating profit more than doubling over the past five fiscal years. In the fourth quarter, 62 percent of SSG revenue came from managed services and project and solutions work. That is where Lenovo can capture more margin than it would from hardware alone, especially as enterprise customers look for help deploying AI rather than just buying equipment.
The pressure is in supply and margins
The risk is that growth is arriving alongside cost pressure. Lenovo pointed to a complex external environment marked by supply shortages and rising component costs. That is the uncomfortable part of the AI boom. Demand may be strong, but memory, silicon, and advanced cooling systems are not free, and customers will eventually push back if infrastructure costs keep climbing.
For founders and AI startups, the takeaway is practical. Compute pricing is unlikely to fall sharply while the supply chain remains tight and large buyers keep absorbing capacity. Startups building AI applications should plan for infrastructure costs to remain a strategic variable, not a background expense. Hardware startups, meanwhile, should pay close attention to the parts of the stack Lenovo is emphasizing: inference, liquid cooling, storage, deployment tooling, and managed services.
Lenovo's next test is whether ISG can stay profitable as Dell, HPE, Supermicro, and cloud infrastructure suppliers fight for the same demand. The $21 billion AI server pipeline suggests customers are still lining up. The harder question is who can deliver enough systems, at acceptable margins, while the market moves from AI pilots into full production. That is where the next phase of infrastructure competition will be decided.
Also read: Polyend Endless puts AI into a $299 stompbox • Spotify brings ElevenLabs AI voices to author platform • Spotify Studio turns your data into personal podcasts
