Alibaba’s shares surged on Wednesday after the Chinese tech giant announced its latest artificial intelligence (AI) reasoning model, QwQ-32B, which it claims can rival DeepSeek’s globally celebrated R1 model.
This announcement ignited investor enthusiasm, driving Alibaba’s Hong Kong-listed shares up by 8.39% to reach a new 52-week high. The company’s U.S.-listed stock also saw a 2.5% rise in premarket trading. So far this year, Alibaba’s shares in Hong Kong have jumped nearly 71%, underscoring the market’s confidence in its AI endeavors.
The debut of QwQ-32B is a crucial step forward in Alibaba’s AI strategy. The new model features 32 billion parameters, significantly fewer than DeepSeek’s R1, which boasts 671 billion parameters, with 37 billion actively engaged during inference—the process where an AI model applies learned data to real-world tasks.
Register for Tekedia Mini-MBA edition 17 (June 9 – Sept 6, 2025) today for early bird discounts. Do annual for access to Blucera.com.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register to become a better CEO or Director with Tekedia CEO & Director Program.
While a smaller parameter size may seem like a disadvantage, it can also signal greater efficiency, an increasingly sought-after trait as developers aim for AI models that offer high performance while consuming fewer resources. Alibaba claims that QwQ-32B has achieved “impressive results” and expressed optimism about its potential for further enhancements, particularly in math and coding applications.
The introduction of QwQ-32B is part of a broader push by Chinese technology companies to position themselves as leaders in the global AI race. Since the unexpected release of DeepSeek’s R1 earlier this year, the competition to develop more efficient and powerful AI models has intensified, with both established giants and new players vying for dominance.
Alibaba, which launched its first AI model in 2023, has steadily increased its investments in the technology, particularly through its Cloud Intelligence unit. This division was a key contributor to the company’s robust profit growth in the December quarter, and CEO Eddie Wu has predicted that AI-driven revenue growth will continue to accelerate.
However, the ambitious growth of China’s AI industry faces significant challenges, particularly from its own government. Beijing’s stringent censorship policies are a major hurdle for AI development, as models like QwQ-32B and even DeepSeek’s R1 must navigate restrictions on data and content. These constraints can stifle innovation, particularly in generative AI models that rely on vast, diverse datasets to improve their reasoning and output. The Chinese government’s tight grip on information and insistence on aligning AI outputs with state-approved narratives have often led to concerns about the authenticity and reliability of Chinese AI models.
Moreover, while Chinese companies like Alibaba push to rival or surpass their U.S. counterparts, Western markets are increasingly wary of AI technologies originating from China. This skepticism stems from fears that these companies may have deep ties with Beijing and the Chinese Communist Party (CCP).
The opaque relationship between Chinese tech firms and the government fuels concerns that AI tools developed by these companies could be leveraged for surveillance, censorship, or influence operations. Western regulators and businesses are particularly cautious about integrating Chinese AI solutions into critical infrastructure or consumer technologies, fearing potential backdoors or data-sharing requirements under China’s national security laws.
These geopolitical and regulatory challenges have led to a de facto isolation of Chinese AI companies from lucrative Western markets. U.S. lawmakers have repeatedly flagged risks associated with the adoption of Chinese technology, leading to bans and restrictions, especially in sectors related to data and AI.
The U.S. government has also imposed export controls on advanced semiconductor technology, critical for AI development, in an effort to curtail China’s technological advancement. These restrictions not only limit Chinese companies’ access to cutting-edge hardware but also hinder their ability to train large-scale AI models as efficiently as their Western counterparts.
The global AI landscape is thus marked by a deepening divide, with Chinese firms like Alibaba, Baidu, and Tencent investing heavily to develop alternatives to Western models, while the U.S. and its allies focus on creating secure and transparent AI ecosystems. Despite this, Alibaba’s QwQ-32B represents a bold attempt to bridge this gap, particularly by emphasizing the model’s efficiency and potential for continuous improvement. The model’s ability to perform well with fewer parameters could appeal to markets looking for AI solutions that require less computational power and lower costs.
Industry experts are optimistic about Alibaba’s prospects. Bernstein analysts noted that advancements in AI could propel Alibaba’s stock and strengthen its earnings trajectory. The potential for AI-driven growth is seen as a critical factor in setting Alibaba on a “more upwardly-pointing trajectory,” particularly through its cloud and enterprise services. However, the broader geopolitical context could affect these gains, especially if tensions between China and the West lead to further trade or technology restrictions.
Dan Newman, CEO of Futurum Group, highlighted the rapid pace of AI innovation during an interview with CNBC’s “Squawk Box Europe.” He noted that the debut of DeepSeek’s R1 had initially prompted speculation about whether OpenAI would maintain its lead in the AI sector. With Alibaba now introducing QwQ-32B, the competitive field appears more crowded and dynamic.
“The pace of innovation is incredibly fast right now. It’s really good for the world to see this happening,” Newman said. He added that as large language models become more “commoditized,” the focus will shift to reducing costs and improving accessibility.
Newman emphasized that the future of AI lies in the inference phase—the stage where AI models are actively used in practical applications—rather than just in their initial training.
“The inference, the consumption of AI, is really the future, and this is going to exponentially increase that volume,” he explained.
His comments underline a key strategic advantage for Alibaba: if the company can make AI applications more affordable and accessible, it could open new revenue streams and solidify its standing in the global market.
The Chinese retailer Alibaba unveiled its latest AI model Thursday, boasting that its reasoning and coding capabilities outperform models from rivals OpenAI and DeepSeek while using fewer resources, per CNN. The AI agent, dubbed QwQ-32B, operates with 32 billion parameters compared to DeepSeek’s R1 model at 671 billion parameters. It arrived shortly after Chinese startup Monica released its newest AI model. Alibaba has pledged to invest $52.4 billion in AI and cloud computing.