Alibaba has emerged as a significant player in artificial intelligence (A.I.), notably through its open-source models, but the company faces challenges in turning this popularity into sustainable profits. The Chinese technology giant introduced its Qwen family of A.I. models in 2023, quickly releasing them as open source. This approach allowed developers worldwide to freely use and modify the technology, contributing to Qwen becoming the most downloaded open-source A.I. system by early 2024, with an estimated one million downloads per day.

Alibaba’s entry into A.I. builds on its extensive data and computing infrastructure developed through its global e-commerce and cloud computing operations. Analysts note that the company’s early investments in data centers and cloud services have positioned it well for A.I. development. Yet, despite widespread use of its open-source models, the financial returns have so far been limited. In the first quarter of 2024, Alibaba reported $1.3 billion in revenue from A.I.-related products, representing less than 4 percent of its total revenue. This figure contrasts sharply with the company’s planned investment of over $55 billion in A.I. infrastructure by the end of 2025.

Internal tensions have surfaced within Alibaba’s A.I. division, particularly around the commercialization strategy of its open-source models. Several key engineers, including Lin Junyang, the lead engineer on the Qwen project, left the company earlier this year amid disagreements over whether to maintain an open-source approach or shift toward proprietary, pay-to-use models. While Alibaba has historically kept its most advanced systems proprietary, it now appears to be emphasizing closed models more heavily, releasing three such systems in April alone.

The move toward proprietary models reflects broader industry dynamics and the high costs of competing in A.I., including expensive data centers and limited access to advanced computer chips due to U.S. export controls. Chinese companies, including Alibaba, face significant challenges acquiring cutting-edge semiconductors, which many industry insiders say hampers their competitiveness.

Alibaba’s difficulties are further exacerbated by geopolitical considerations. The U.S. Department of Defense recently added Alibaba to a blacklist of companies alleged to have ties to the Chinese military, a designation Alibaba denies and says has already affected its business operations. Additionally, rival A.I. developers, including Anthropic and OpenAI, have accused Alibaba and other Chinese firms of improperly obtaining data to accelerate their own model development. Anthropic, for instance, alleged the use of thousands of fraudulent accounts to copy its technology in a letter to U.S. senators, charges Alibaba has declined to comment on.

Alibaba’s A.I. endeavors reflect a wider challenge in the global A.I. sector, where many companies have gained technological acclaim but struggle to establish sustainable revenue models. Experts observe that the current industry resembles the early dot-com era, marked by rapid growth but unclear paths to long-term profitability. As Alibaba recalibrates its strategy amid internal and external pressures, its next moves will be closely watched as part of China’s broader ambition to lead in A.I. development.