Alibaba Group chairman Joe Tsai has stated that open-source artificial intelligence offers Europe its strongest route to technological independence, urging the region to build on its manufacturing expertise and retain control over the industrial data needed to develop AI systems tailored to its own needs.
Speaking earlier this week at the Wave technology event in Italy, the Alibaba co-founder pointed to China’s open-source AI development as a model for countries seeking to strengthen their technological capabilities despite constraints on access to advanced computing resources and technologies.
Tsai said Europe had an opportunity to develop its own AI capabilities because of its deep industrial base, which he compared with China’s manufacturing strengths. Factories across the region generate valuable operational data that could be used to train and refine AI models, potentially giving European companies greater control over how the technology is deployed.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
“You don’t want to feed it into some closed-source API. You want to maintain those data and be able to take open-source models and then further train your model, and improve it for your use,” he said.
His argument centers on the relationship between data ownership, AI development and industrial competitiveness. Rather than relying exclusively on proprietary AI services controlled by foreign technology companies, European manufacturers could use open-source models as a foundation for developing systems adapted to their production processes, equipment and operational requirements.
The approach could also reduce dependence on a limited number of AI providers. However, access to open-source models alone does not guarantee technological independence. Developing competitive systems also requires computing infrastructure, technical expertise, investment and the ability to maintain and secure AI applications.
Tsai’s comments come as Alibaba intensifies its own open-source strategy, positioning the Chinese technology group to compete in a global AI market increasingly shaped by disagreements over access to advanced chips, model development and the commercial value of proprietary technology.
Tsai described Alibaba’s current corporate era, beginning in 2023, as its “AD period”, marking a shift away from what he characterized as unfocused expansion towards two principal priorities: e-commerce and full-stack AI.
The designation captures a broader restructuring of the company’s ambitions under Tsai, who became chairman in 2023 after co-founding the group with Jack Ma and other partners in 1999.
Alibaba, headquartered in Hangzhou, is using the cash generated by its e-commerce operations to finance the development of AI models and supporting infrastructure. Tsai said the business generated US$25 billion in free cash flow, which the company was using to support its AI ambitions.
He also said Alibaba had doubled its capital expenditure on computing infrastructure every year for the past three years, signaling the scale of its commitment to the hardware and systems needed to develop and deploy increasingly capable AI models.
The spending highlights the capital-intensive nature of the AI race. Companies seeking to compete at the frontier must invest not only in model research but also in computing capacity, data infrastructure and the systems required to serve customers at scale.
Alibaba’s e-commerce operations provide a source of funding for that investment, potentially allowing the company to pursue AI development without relying exclusively on external financing. But the strategy also creates an important test: the company must prove that its AI investments can generate durable commercial returns while sustaining its established businesses.
Tsai’s description of the company’s dual focus suggests Alibaba is seeking to concentrate resources rather than expand indiscriminately across multiple sectors. AI is becoming a central part of that strategy, with the company aiming to build capabilities spanning chips, computing infrastructure and models.
At its annual Apsara Conference in September, Alibaba introduced the Zhenwu V900 processor, describing it as “the most powerful AI chip in China today”. The company also outlined plans to train an AI model with up to 10 trillion parameters.
The announcements underpin the breadth of Alibaba’s ambitions. Developing its own processors could help strengthen its control over parts of the computing stack, while larger models would require substantial computing resources and engineering capabilities. The announcements, however, do not by themselves establish how the processor or planned model will perform against competing systems.
Export Controls Sharpen The Divide Between Chinese And US AI Strategies
Tsai also contrasted the open-source strategies of Chinese AI developers with the proprietary approach of leading US laboratories, arguing that restrictions imposed by Washington have helped motivate Chinese companies to pursue alternative paths to innovation.
“Chinese companies are open sourcing their innovation, they write papers,” he said. “The American closed-source labs don’t write papers any more because they don’t want to share.”
The remarks frame open-source development as both a competitive strategy and a response to the constraints facing China’s technology industry. By making models and research more widely available, Chinese developers can encourage outside researchers and businesses to adapt their work, identify improvements, and build applications without having to develop every component independently.
That process can expand the reach of a model beyond the company that originally created it. It can also allow developers with fewer resources to benefit from work conducted by better-funded organizations, potentially accelerating adoption and innovation across an ecosystem.
Tsai argued that US export controls on advanced technologies had further encouraged the development of China’s AI industry. Restrictions on access to leading-edge chips and other technologies have increased the pressure on Chinese companies to find alternative ways to improve model performance and computing efficiency.
He compared the circumstances facing US and Chinese technology companies to children born into different levels of privilege. US firms, in his analogy, resemble children from wealthy families with access to abundant resources, while Chinese companies have stronger incentives to work harder and innovate because they cannot rely on the same advantages.
The comparison reflects Tsai’s view that constraints can encourage experimentation and alternative approaches. It does not mean those restrictions have eliminated China’s dependence on advanced computing resources or removed the technical and financial challenges involved in developing competitive AI systems.
Nor does the contrast between open and closed development necessarily capture the full range of strategies across either country. Open-source models can coexist with proprietary products, while companies may release some components of their systems and retain control over others. The commercial value of each approach depends on factors including model performance, security, deployment costs, and the ability to turn technical capabilities into widely used products.
For Alibaba, however, open-source development offers a way to expand its influence beyond its own platforms. Wider adoption of its models can attract developers and enterprise customers, potentially strengthening demand for its cloud services and supporting technologies. The approach also creates competitive pressure on companies whose business models rely heavily on keeping their most capable models proprietary. If open models deliver sufficient performance at lower deployment costs, customers may have greater incentives to run or customize them rather than depend entirely on external AI services.
AI Expected To Become Embedded In Everyday Business
Tsai expects the gap between AI as a separate technology and AI as a routine part of business to fade over the coming years.
“Five years from now, we won’t be talking about AI per se,” the 62-year-old said. “AI would have been infused in every aspect of our business. For the same reason, we don’t talk about the internet today, but the internet is everywhere.”
The prediction points to a future in which AI becomes embedded in ordinary business operations rather than remaining a separate product category. In manufacturing, that could involve systems trained on industrial data to support production planning, maintenance, quality control, and other specialized tasks. In commerce, AI could become integrated into customer service, product discovery and operational decision-making.
Such adoption would make control over data and the ability to customize models important to companies seeking to develop applications for their own operations. But it would also raise questions about infrastructure costs, data protection, system reliability, and whether the productivity gains justify the investment.
For Europe, Tsai’s argument presents open-source AI as an opportunity to build on existing industrial strengths rather than depend exclusively on technology developed elsewhere. The region’s manufacturing base could provide valuable data and use cases, but translating that advantage into technological independence would require sustained investment and the technical capacity to develop, adapt, and deploy models.
Alibaba is pursuing a similar proposition in China, combining open-source model development with investment in chips and computing infrastructure. Analysts believe its ability to convert that approach into commercially successful products will help determine whether openness can become a durable advantage in the global AI market.