Baidu’s artificial intelligence chip subsidiary Kunlun Chip is moving closer to an initial public offering in Hong Kong that could value the company at about $50 billion, a level that would make the chipmaker worth more than its parent company’s current Hong Kong market capitalization and set a major new benchmark for China’s domestic AI hardware sector.
The planned listing comes as China’s technology industry is racing to secure alternatives to foreign-made advanced processors, especially as export controls continue to restrict access to the most powerful chips used for AI training and inference. Kunlun’s offering, if completed at the targeted valuation, would place it among the most closely watched semiconductor listings in Asia and could become the largest public offering associated with Baidu founder Robin Li Yanhong.
At recent exchange rates, a $50 billion valuation would equal roughly 360 billion yuan, well above the valuation Kunlun received when it was separated from Baidu in 2021. It would also exceed Baidu’s own Hong Kong market value, which has recently remained above 300 billion Hong Kong dollars after a rebound in the company’s shares following reports of Kunlun’s listing progress.
Kunlun has already submitted a confidential A1 listing application to the Hong Kong Stock Exchange through its co-sponsors, according to market filings cited by people familiar with the process. The company is now understood to be in the final stage before listing and has begun discussions with potential cornerstone participants. That phase is often one of the last opportunities for large institutions and strategic buyers to secure shares before trading begins.
The proposed IPO is more than a financial event. It is also a test of how much public markets are willing to pay for China’s domestic AI computing supply chain at a time when demand for computing power is rising sharply across cloud services, telecommunications, finance, energy, industrial automation and government-backed digital infrastructure.
A listing shaped by chip demand
Kunlun’s listing process has drawn attention partly because of the conditions reportedly attached to cornerstone participation. The company is said to require any cornerstone participant to first commit to buying Kunlun chips worth three to seven times the value of the shares subscribed.
That condition is unusual because it favors buyers with real demand for AI hardware rather than purely financial institutions seeking exposure to a high-growth semiconductor name. In practice, it makes the IPO as much a commercial distribution exercise as a capital-raising event.
For Kunlun, the condition could serve several purposes. It may help demonstrate demand for its products before the listing, strengthen future revenue visibility and deepen ties with large enterprise customers. It could also reduce the risk that early shareholders are only short-term market participants with little strategic connection to the company’s business.
Market sources have described demand for the limited cornerstone quota as strong. The intensity is understandable. AI chips have become one of the most important building blocks in China’s technology strategy, and companies that can supply domestic alternatives to Nvidia products have become highly sought after by traders, industry funds and strategic buyers.
The structure also reflects a broader reality in the AI chip market. Access to computing hardware is no longer simply a procurement matter. For cloud platforms, telecom operators, banks, energy companies and large internet groups, it has become a strategic resource that can influence product development, data center planning and long-term competitiveness.
Why the valuation matters
A $50 billion valuation would represent a dramatic increase from Kunlun’s earlier funding level. When Baidu spun off the unit in 2021 to form Kunlun Chip (Beijing) Technology Co., Ltd., the company’s only public fundraising round valued it at about 13 billion yuan. That round was led by CPE Yuanfeng, with participation from IDG Capital, Legend Capital and Yuanhe Puhua.
Since then, Kunlun’s shareholder base has expanded significantly. The company now has 57 shareholders, including state-backed funds, industry funds and private capital groups. The widening shareholder list reflects both policy support and commercial interest in China’s AI semiconductor ecosystem.
The scale of the potential valuation also matters for Baidu. Although the exact post-listing shareholding structure will depend on the offering terms, a valuation near $50 billion would imply that Baidu’s stake could be worth more than 100 billion Hong Kong dollars. That would be meaningful for Baidu’s market value and could help traders reassess the company’s hidden assets in AI infrastructure.
Baidu’s stock rose for four consecutive trading sessions after reports of Kunlun’s IPO progress, keeping its Hong Kong market value above 300 billion Hong Kong dollars. That reaction suggests the market sees Kunlun as a potentially important source of value, even as Baidu’s core businesses continue to face questions around advertising growth, cloud competition and the commercialization of generative AI.
For Baidu, Kunlun’s listing could reshape how the market views its AI portfolio. Instead of being seen mainly as a search, cloud and autonomous driving company, Baidu would also have a visible stake in a major AI chip supplier. That could strengthen the argument that Baidu has built an integrated AI stack covering algorithms, cloud services, applications and hardware.
From internal unit to national AI supplier
Kunlun’s roots go back to 2011, when Baidu established an internal chip and architecture division. The team was built by engineers from Baidu, Qualcomm, Marvell and Tesla, and began developing large-scale AI computing chips at a time when China had few domestic alternatives for advanced AI workloads.
For much of its early history, the chip unit served Baidu’s internal needs. Baidu’s search engine, recommendation systems, cloud services and AI research required large amounts of computing power, making custom chips strategically useful. The company’s work in autonomous driving and large language models later added to that demand.
In 2021, Baidu spun out the unit as Kunlun Chip to accelerate commercialization outside the parent company. That separation allowed the chipmaker to raise external capital, broaden its customer base and pursue contracts across industries that might not have been available under a purely internal structure.
Since the spin-off, Kunlun has moved well beyond supplying Baidu. Its external revenue now exceeds internal supply to the parent company, according to the provided company information. Major customers include China Mobile, Southern Power Grid and China Merchants Bank.
That customer mix is important. It shows that Kunlun is not only selling into internet data centers but also into telecom networks, energy systems and financial institutions. These are sectors with large computing needs, long procurement cycles and strong policy incentives to adopt domestic technology where possible.
In August 2023, Kunlun secured multi-billion-yuan orders in a national procurement program led by China Mobile. The company reportedly ranked first in all three supplier categories it entered. That result gave Kunlun a stronger public reference point and helped validate its position in China’s AI accelerator market.
The P800 and the product roadmap
Kunlun’s main product is the P800, launched in 2024 and manufactured on Samsung’s 7-nanometer process. The chip is positioned to compete with Nvidia’s A800 GPU, a processor that became important in China after tighter U.S. export controls limited access to Nvidia’s most advanced AI chips.
The P800 is central to Kunlun’s near-term commercial story. To justify a premium valuation, the company must show that it can deliver chips at scale, support real enterprise workloads and build a software ecosystem that makes its hardware practical for customers.
AI accelerators are not judged only by theoretical performance. Buyers also care about power efficiency, reliability, compatibility with common AI frameworks, ease of deployment, cluster performance and after-sales technical support. Nvidia’s strength has long rested not only on chip design but also on its CUDA software ecosystem and deep developer support. Domestic rivals such as Kunlun must close that gap if they want broader adoption.
Kunlun’s roadmap includes two follow-up chips: the M100 and M300, scheduled for release in 2026 and 2027 respectively. The M100 has reportedly already been delivered to key clients for testing ahead of its formal 2026 launch. That testing phase will be closely watched because customer feedback could influence how traders judge Kunlun’s ability to move beyond a single flagship product.
The M300, expected in 2027, could become even more important if China’s AI industry continues to scale up model training, enterprise inference and industrial AI applications. By then, domestic chipmakers will likely face higher expectations for performance, energy efficiency and software maturity.
According to IDC data cited in the source material, Kunlun and Cambricon are projected to rank third among domestic suppliers in China’s AI accelerator server market in 2025, with each expected to ship about 116,000 cards. That forecast places Kunlun among the leading local suppliers, although competition remains intense.
A market driven by policy and restriction
Kunlun’s IPO is unfolding against a geopolitical backdrop that has changed the economics of AI hardware. U.S. export controls on advanced semiconductors and chipmaking equipment have limited China’s ability to buy the most powerful foreign AI processors. Those restrictions have accelerated China’s push for self-reliance in high-performance computing.
The immediate result has been a reordering of the supply chain. Chinese cloud groups, telecom operators and state-linked enterprises have increased efforts to source domestic chips. At the same time, domestic chip designers have received stronger policy support, more customer attention and higher valuation expectations.
Hong Kong has become an important part of this system. The city continues to act as a major logistics and financial hub for semiconductor flows into China. In the first five months of 2026, Hong Kong handled more than half of China’s reported $239 billion in chip imports, according to the source material. That record share highlights Hong Kong’s role not only as a capital market but also as a vital node in the regional semiconductor supply chain.
For Kunlun, a Hong Kong listing offers several advantages. It gives the company access to global capital, raises its visibility with international traders and allows strategic customers to participate in its growth. It also provides Baidu with a clearer market valuation for one of its most important AI infrastructure assets.
At the same time, a Hong Kong IPO carries risks. The city’s equity market has seen uneven sentiment toward technology listings in recent years. Valuations can move sharply depending on global rates, China policy signals, U.S.-China tensions and earnings visibility. For a company seeking a valuation near $50 billion, timing will matter.
Competition among domestic AI chipmakers
Kunlun is not alone in trying to capture China’s AI hardware opportunity. Cambricon, Moore Threads and other domestic chip developers are also moving through capital market processes or preparing for future listings. Some are focused on AI accelerators, while others target GPUs, data center chips or broader computing architectures.
Cambricon is one of the most established domestic AI chip names and has long been viewed as an important competitor. Moore Threads has drawn attention for its GPU ambitions. Other companies are working across training chips, inference chips, edge AI processors and specialized accelerators.
This competition is healthy for China’s AI ecosystem, but it also creates pressure. Customers want domestic alternatives, yet they do not want fragmented hardware environments that are difficult to integrate. The winners are likely to be companies that combine strong chips, stable supply, reliable software tools and large-scale customer support.
Kunlun may have an advantage because of its Baidu background. Baidu provides deep experience in real AI workloads, particularly search, recommender systems, cloud AI and large models. That gives Kunlun a testing ground and technical reference that many independent chip startups lack.
However, Baidu’s connection cuts both ways. Traders will want to see that Kunlun can continue growing outside its parent company. The fact that external revenue already exceeds internal supply is therefore a key part of the listing story. It helps support the argument that Kunlun is becoming an independent AI infrastructure company rather than simply a captive supplier for Baidu.
The dual-track listing plan
Kunlun is pursuing both a Hong Kong Main Board listing and a STAR Market submission in mainland China, a dual-track “A + H” plan. This approach gives the company flexibility and could allow it to tap different pools of capital.
A Hong Kong listing may appeal to global funds, international institutions and strategic buyers interested in China’s AI hardware market. A STAR Market listing, meanwhile, could help Kunlun access mainland traders who have shown strong demand for semiconductor and advanced manufacturing names.
The dual-track strategy also reflects the strategic nature of the company’s business. AI chips sit at the intersection of national technology policy, capital market interest and industrial demand. By keeping both paths open, Kunlun can respond to market conditions and regulatory timelines.
Still, dual-track listings can be complex. They require careful coordination around disclosure, valuation, regulatory review and shareholder structure. The company will need to balance the expectations of Hong Kong market participants with those of mainland regulators and domestic technology policy priorities.
What traders will watch next
The most immediate focus will be the final IPO terms, including the size of the offering, cornerstone allocation, lock-up arrangements and the level of demand from strategic customers. The chip purchase requirement attached to cornerstone participation will be especially important because it could turn the listing into a signal of future sales.
Traders will also watch whether Kunlun can convert testing of the M100 into broad commercial adoption. Early customer trials are important, but the real test is deployment at scale. Large enterprise buyers, especially telecom and financial institutions, are cautious. They need proof that new chips can handle demanding workloads reliably and economically.
Another key issue is software. Domestic AI chips cannot rely on hardware performance alone. Customers need development tools, model support, migration assistance and compatibility with widely used AI frameworks. If Kunlun can make it easier for customers to move workloads from foreign chips to domestic alternatives, its commercial position could strengthen significantly.
Production capacity will also matter. Even if demand is strong, Kunlun must secure manufacturing, packaging and supply chain support. The P800 is built on Samsung’s 7-nanometer process, highlighting continued reliance on advanced foundry capacity outside mainland China. Any disruption in manufacturing access, export rules or supply chain logistics could affect delivery schedules.
Finally, traders will watch Baidu’s role after the listing. Baidu’s continued support could help Kunlun in technology development and customer credibility. But Kunlun will need enough independence to serve a wider market and avoid being viewed as too closely tied to one parent company’s internal needs.
A broader signal for AI infrastructure
Kunlun’s proposed valuation provides a new reference point for the market value of AI infrastructure in China. For the past two years, much attention has focused on large language models, chatbots, cloud platforms and AI applications. But the hardware layer is increasingly becoming the scarce resource that determines how fast those applications can scale.
The ability to secure high-performance computing power now affects almost every major digital business. Model developers need chips for training and inference. Cloud providers need chips to build AI services. Banks and telecom operators need chips to process data and deploy intelligent systems. Industrial companies need chips for automation, prediction and optimization.
That makes Kunlun’s listing relevant beyond the semiconductor sector. It could influence how traders value companies with direct access to computing power, long-term supply agreements or in-house chip capabilities. It could also raise questions about companies that depend heavily on restricted foreign hardware without clear domestic alternatives.
If Kunlun succeeds, it would show that public markets are willing to assign very high value to domestic AI computing platforms. That could encourage more listings from chip designers and related infrastructure companies. It could also increase competition for strategic placements in future offerings.
A turning point for Baidu and Kunlun
For Baidu, the IPO could unlock value in an asset that was built over more than a decade. The company began investing in chip architecture long before the current generative AI boom, when domestic AI chips had limited commercial visibility. That early commitment may now deliver a significant market payoff.
For Kunlun, the listing would provide capital to expand production, develop new products and strengthen customer support. It would also bring public market scrutiny, including pressure to disclose revenue growth, margins, customer concentration, research spending and supply chain risks.
The offering will not answer every question about China’s domestic AI chip industry. Kunlun still faces strong competition, difficult technology challenges and uncertainty around manufacturing access. But the listing would mark a major step in the commercialization of China’s AI hardware sector.
If market conditions remain supportive, Kunlun’s IPO could become one of the defining technology listings of the current cycle. It would give Baidu a clearer valuation for a core AI infrastructure asset, give Kunlun fresh capital for expansion and give the market a new benchmark for the price of domestic computing power in an era when chips have become central to economic and technological strategy.
Explore AI‑driven trading strategies aligned with Kunlun’s chip ambitions using our smart AI copy trading tools for crypto markets.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
