Goldman Sachs Raises Lenovo's Target Price to HKD 27: From PC Giant to AI Full Stack Company, the Valuation Logic is Switching

Goldman Sachs Asia, in its latest research report dated May 25th, significantly raised its 12-month target price for Lenovo Group (0992.HK) from 12.53 Hong Kong dollars to 27 Hong Kong dollars, implying a 71.4% upside potential over the current stock price of 15.75 Hong Kong dollars, while maintaining a buy rating.

Behind this substantial target price increase lies a deeper restructure of the valuation framework. Goldman Sachs no longer prices Lenovo using the lower valuation multiples of traditional consumer electronics companies but instead incorporates it into the pricing system for AI infrastructure companies, with the target P/E jumping from 7.7 times to 15.7 times.

Quarterly Report Validation: Lower Than Expected Memory Cost Pressure

The direct catalyst for this increase is Lenovo's recently announced performance for the quarter ending in March.

PC shipments grew by 8.6% year-over-year, significantly exceeding the global market growth of 2.5%, with the global market share further increasing to 25.2%. Despite the continuous rise in memory prices, the operating profit margin for the IDG business (mainly including PCs and smartphones) remained stable at 6.9%, almost on par with the same period last year. Goldman Sachs believes this confirms Lenovo's clear advantage as the world's leading player in PC market share, in terms of supply chain bargaining and cost transfer, and that concerns about the impact of memory price increases are overestimated.

The performance of the server business was even more impressive. The ISG division (servers and storage) saw a year-over-year revenue increase of 37%, with the operating profit margin turning from losses in the previous three consecutive quarters to a positive 3.6%, sending a clear signal of profitability repair.

AI Server Acceleration, High-end ASP Soars

The core argument supporting Goldman Sachs' valuation increase is Lenovo's rapid advancement in the product structure of AI servers.

The average unit price of high-end AI servers (selling for over $250,000) reached $1.1 million and $1.4 million in the third and fourth natural quarters of this year, respectively, almost doubling from $576,000 in the second quarter, clearly reflecting Lenovo's accelerated shift towards the high-value-added segment of AI training servers.

On the production capacity front, Lenovo has built an annual server rack production capacity of over 70,000 units, with more than 11,000 of them being direct liquid-cooled racks. The company also announced that it will start shipping the next-generation AI server rack, Rubin, in the second half of this year. Goldman Sachs forecasts that Lenovo's overall server revenue will achieve a 37% compound annual growth rate between the natural years of 2026 and 2028, contributing 19% of total revenue by 2028; the storage business is expected to expand at a compound annual growth rate of 25%, with the ISG overall revenue share increasing from 23% in 2025 to 35% by that time.

It is worth mentioning that Lenovo completed the acquisition of high-end enterprise storage vendor Infinidat in April 2026, whose architecture is specifically designed for AI inference and large-scale data analysis scenarios, further strengthening Lenovo's layout in the high-end storage market.

AI Agent Ecosystem: The Imagination of the Software Layer

In addition to hardware, Goldman Sachs' bullish logic on Lenovo also extends to software and the Agent ecosystem.

On the consumer side, Lenovo launched the personal AI super assistant Qira at this year's CES, supporting seamless cross-device collaboration—tasks started on a ThinkPad can be seamlessly switched to a Motorola Razr to continue execution, with the system proactively providing calendar management, message reply, hotel booking, and other scenario suggestions based on user data. On the enterprise side, the xIQ platform offers full lifecycle management capabilities for AI Agents, with over 60 industry-specific Agents pre-installed, allowing enterprises to complete deployment within a week.

The hardware side's update attempt comes from the AI Host released in mid-May—a 2999 yuan priced edge computing node for homes and individuals, capable of running a large model with up to 122 billion parameters locally, supporting cloud-edge collaborative inference while ensuring that private data does not leave the local area. Goldman Sachs believes that the penetration of personal AI Agents will simultaneously drive the demand for device ends (AI PCs, AI phones, AI glasses) and infrastructure ends (AI Host, edge servers), making Lenovo's product matrix spanning both ends a rare beneficiary.

Valuation Soars Significantly, Significant Di