Against the backdrop of a continuously heating AI narrative, Barclays' latest research report has conducted a pivotal valuation reassessment of Baidu.
According to information, despite maintaining a cautious stance on Baidu's core advertising business, driven by the anticipated spin-off and IPO of its Kunlun AI chip unit, Barclays has formally integrated Baidu's AI chip business into its valuation framework for the first time. The target price has been significantly raised from $100 to $147, marking a substantial 47% increase, while the rating is maintained at "Equal Weight."
Barclays indicates that this signifies a shift in Baidu's valuation logic from a single focus on "search + advertising" to a dual-track structure comprising an "advertising foundation + AI hard-tech assets."
This is not merely about adding another asset; it represents a change in Baidu's valuation "coordinate system." Superficially, Barclays appears to have done one thing: added the Kunlun chip asset to its existing valuation model and, based on this, raised the target price by 47% in one move.
However, in essence, this constitutes a switch in the valuation paradigm. How has the market historically priced Baidu? For a long time, Baidu has been confined within a very "narrow" framework:
Core anchor: Search advertising
Secondary variables: Cloud business, autonomous driving (heavily discounted)
The overall pricing approach essentially still treated Baidu as an internet company in its mature phase, overlaying it with a dual discount for policy uncertainty and slowing growth. Within this system, any "future narrative" would be met with the question: "When will it be able to offset the decline in advertising?" As long as advertising underperformed, other businesses struggled to genuinely lift the valuation. What fundamental change has Barclays implemented now? Barclays has, for the first time, acknowledged a fact:
Baidu now possesses an "AI hard-tech asset" that does not rely on advertising, traffic, or C-end user behavior. This implies that the Kunlun chip is not merely a "business supplement" but an independent branch of the valuation methodology. This is also why Barclays did not raise its assumptions for advertising and cloud, yet still substantially increased the target price.
In essence, this step can be summarized in one sentence: Baidu is, for the first time, being permitted to "tell a growth story independent of advertising." Why now? What is the "value trigger" for the Kunlun chip? Many might ask: Hasn't the Kunlun chip been in development for five years? Why the revaluation only now? Barclays' answer points not to a "technological breakthrough," but to the simultaneous maturation of three external conditions. 1. Domestic AI chips are transitioning from a "policy mandate" to "genuine demand." In recent years, domestic chips were more often policy-driven, strategic reserves, or demonstration projects. However, the situation has changed. With the explosion in demand for large model inference, computational costs becoming a bottleneck for commercialization, and ongoing restrictions on HBM/high-end GPUs, domestically produced inference chips that are "usable, deliverable, and scalable" have become a necessity for the first time. Barclays' research report uses the key phrase "surge in demand for domestically-made chips." This represents a crucial shift in terminology. 2. The Kunlun chip's positioning is "realistic," not about "competing directly with Nvidia." Barclays does not attempt to position the Kunlun chip against training chips or narratives centered on H100-level performance. Its implicit judgment is that the Kunlun chip's advantage lies in the inference segment, forming a closed-loop application within Baidu Cloud, search, autonomous driving, and industry-specific solutions. This means it doesn't need to be the "world's best"; it only needs to be a sustainably deliverable component within China's AI ecosystem. This is the underlying logic for the 15x revenue multiple:
It's not a valuation based on being a disruptor, but rather a valuation for high-certainty, growth-oriented hard tech.
3. The IPO is not about fundraising, but about "outsourcing pricing power." This point is crucial but often overlooked. The significance of the Kunlun chip IPO for Baidu lies not in the amount of capital raised or the relief it provides for cash flow, but in one thing: transferring the question of "what the Kunlun chip is worth" from within the company to the capital markets. Once listed: the Kunlun chip unit will have independent financial reports, will be compared horizontally with companies like Cambricon, Biren, and Iluvatar, and will receive a market price that can be continuously updated and revised upward. This is why Barclays is incorporating it into the model proactively, rather than chasing the price after the IPO. Why didn't the weak advertising performance prevent this revaluation? This is a point of confusion for many investors. Barclays states clearly:
Search is being eroded by AI chatbots.
Advertising monetization faces structural downside risks.
It even lowered revenue and EPS forecasts.
So why was the target price still raised? The answer lies in one sentence: the market is beginning to accept that Baidu's future value no longer stems solely from "user attention." At a deeper level: This represents a reordering of the "AI business model."
In the old internet era: Users → Traffic → Advertising → Cash flow. In the AI era, a new value chain is forming: Computing power → Models → Industry solutions → Infrastructure-level revenue.
The Kunlun chip is positioned precisely at the most upstream, hardest-to-replicate link in this chain. This is also why the advertising business is valued using an EBITDA multiple, while the Kunlun chip is valued using a revenue multiple—one is priced as a declining cash flow, the other as a growth asset. What is the true implication of this revaluation for investors? Looking beyond "Baidu as a single stock," this research report actually signals something broader. 1. A "second valuation curve" for China's internet tech sector is emerging.
First curve: Platforms, traffic, advertising.
Second curve: AI infrastructure, computing power, hard tech.
Those who can successfully spin off this second curve have the opportunity to break free from long-term valuation suppression. 2. A watershed in AI investment: Shifting from models to "deliverable computing power." The market is gradually realizing:
Models will converge.
Computing power and chips are the long-term bottlenecks.
Barclays' stance on the Kunlun chip essentially serves to "pre-anchor" the value of "China's AI computing power assets." 3. Consequently, Baidu's core risks have also become clearer. From now on, Baidu's greatest risk is no longer solely advertising decline, but whether the Kunlun chip can genuinely attract external customers, whether it can shed the valuation discount associated with being "used only within the Baidu ecosystem," and whether it can withstand peer comparisons post-IPO.
Comments