The star player in domestic AI programming, aiXcoder accelerates the industrialization implementation.

The successful paradigm of AI entrepreneurship is not confined to large models.

In 2024, the popularity of Cursor successfully kindled the AI programming track. The valuations of new-generation AI programming companies such as Magic, Replit, and Poolside have successively exceeded $1 billion, making AI programming a global trend-setting opportunity.

In the domestic market, however, there are not many products that have been highly regarded for a long time. Silicon Core Technology (hereinafter referred to as aiXcoder) is one of the AI programming entrepreneurial teams that is currently the fastest and most favored by capital.

In the recently held second season of "Lenovo's New Business Innovation Ecosystem Roadshow", aiXcoder was selected as one of the TOP 10 projects, becoming a new star in Lenovo's ecosystem. This not only demonstrates aiXcoder's technological innovation ability but also closely aligns with Lenovo's SMB business group, which has always been committed to accompanying customers to succeed and working hand in hand with customers: by providing comprehensive, high-quality, and one-stop services, efficiently accompanying the development of growth-oriented enterprises throughout their life cycle, and growing together with innovative enterprises.

"In the field of AI programming, there are actually very few self-developed code large models." In the view of Li Lixing, COO of aiXcoder, in this vertical field of AI programming, the ability to self-develop code large models remains a hard barrier.

When the Wind of AI Programming Blows

As the foundation for enhancing productivity in the digital era, the "wind" of AI programming is strong, but starting a business in the domestic market is not effortless. There are two main reasons:

First, there is a considerable difference between the domestic and foreign development markets. In foreign countries, there are more development needs from independent developers and small and micro enterprises, and they have a stronger sense of paying for general-purpose products; in China, enterprise-level systematic development is the main focus, and there are more opportunities in the toB market.

Second, AI programming is a track that seems to liberate productivity but is actually challenging to position in the market. Whether it is to assist professional development or to popularize it to a broader audience, whether to focus on auxiliary capabilities or full-generation capabilities, and whether the model is self-developed or called, these all directly determine the ceiling of a product.

aiXcoder's ability to stand out stems from its in-depth and extreme self-developed technology and the ability to apply large models in domain-specific applications.

aiXcoder was incubated from the Institute of Software Engineering at Peking University. Under the leadership of Professor Li Ge since 2013, they began to study how to apply deep learning to code generation and code understanding, and in 2014, they published the earliest paper in this field globally. In 2017, with hardcore technological achievements in hand, they made an important decision: to bring the incubation results to the market and establish aiXcoder.

At that time, Google was still the company at the forefront of the artificial intelligence field. At that time, there was no intelligent programming tool based on deep learning on the market, and aiXcoder was the first to develop a plugin for code completion and generation, and has been continuously iterating its self-developed model following the development of the technological environment.

Li Lixing recalled that before the advent of the ChatGPT trend in 2022, there were disputes about the technical direction of AI. "After the release of ChatGPT, for our AI programming field, it is like a huge advertisement."

It is worth mentioning that aiXcoder had already developed the first large model with tens of billions of parameters in China at that time, which is used to support method-level code generation described by natural language.

It was not until 2023 that they successfully entered the commercialization stage and began to contact software development teams in pillar industries such as finance, military industry, aerospace, and technology.

"We realized that the most important demand of these enterprises for applying AI programming is that the code large model must be privately deployed." In order to ensure data privacy and the security of code assets, the efficiency tools and AI generation capabilities of the development team also need to be customized locally, making it difficult to write code using cloud-based products. The strict control of programming technology by these enterprises is a good entry point for AI programming to move towards commercialization.

Filling the Gap in the Ecosystem

In a track with high market imagination, the emergence of intelligence requires the joint promotion of more ecological forces.

In the era of AI entrepreneurship, Li Lixing clearly feels that everyone's sense of competition has become more and more benign.

Due to data differences and vertical category models, AI service providers like aiXcoder are better at focusing on capabilities. But at the same time, what enterprises need is definitely comprehensive AI. Therefore, compared with the individual efforts of entrepreneurs, a collaborative ecosystem is naturally needed. Industry giants such as Lenovo, who have recognized this, are working hard to promote the establishment of the ecosystem.

In the past two years, aiXcoder has been linked to Lenovo's ecosystem through the "Star Project". Linked to the Star Project, Li Lixing frankly admitted that his impression of Lenovo has changed significantly. Before contacting Lenovo's business ecosystem team, he only knew that this enterprise is "the light of China", but he never expected that Lenovo would attach so much importance to the development of start-ups.

"The Star Project has been creating many opportunities for us, and we are also trying various substantive cooperations. So when we learned that Lenovo has such an opportunity for a roadshow, we were very willing to participate."

Li Lixing said that Lenovo's ecological support is not only reflected in investments for start-ups, business expansion opportunities, and channel resource support, but also in long-term cooperation in new productive force fields such as AIPC in the future.

As a full-stack AI product, solution, and service provider, the ecosystem established by Lenovo is of particular significance for AI programming entrepreneurship. The reasons are as follows:

First, as the leading enterprise in an industrial chain, it can deeply participate in the standardization of the industry. The PC industry, as a production environment, is naturally closely related to local software development. As the number one domestic PC manufacturer, Lenovo has a considerable penetration rate in pillar industries, which is consistent with the customer group needs of aiXcoder.

Second, although major technology companies are basically laying out their own ecosystems for AI start-ups, the self-developed products used for internal efficiency in large companies are often packaged as commercial solutions and directly compete with start-up companies. Therefore, looking at the domestic market, industry giants such as Lenovo, which are increasing their investment in AI, can be regarded as one of the few ecosystem builders that can achieve a 1 + 1 > 2 cooperation effect with start-up teams.

In the entire highly competitive AIGC track, the blooming of a diversified ecosystem can further provide solutions to the commercialization problems of AI applications.

Joining Hands to Root in the Industry

In order to enable the capabilities of AI programming to cover more industries, aiXcoder has currently released its self-developed code large model aiXcoder 7B to the open-source community. Just one week after its release, the Star count on GitHub reached more than 2,000, and it also ranked among the TOP 30 on the HuggingFace trend list. It is reported that the related paper of aiXcoder 7B has recently been accepted by the International Conference on Software Engineering (ICSE) 2025, a top international conference in the field of software engineering.

aiXcoder has been at the forefront in the key capabilities of AI programming entrepreneurship. Next, it is to work with ecological partners such as Lenovo to further realize the domain-specific application of code large models in various industries.

Li Lixing said that in fact, many large models perform very well in general programming scenarios and mainstream evaluation sets, but when deployed within enterprises, the accuracy rate will drop sharply. This is precisely because learning and mastering "domain-specific" knowledge is the key to the successful implementation of AI programming in enterprises.

Lenovo, which has been committed to implementing solutions in vertical fields, is an important partner to help aiXcoder understand and digest domain knowledge.

"Now we are also cooperating with Lenovo to consider how to integrate intelligent development tools into AIPC."

Li Lixing said that with Lenovo's brand reputation and market share, as well as the integrity of its technology and ecosystem, it indeed makes these start-up teams that are rooted in the local development environment more hopeful about the future of AIPC.

More importantly, Lenovo has already developed the Baiying Intelligent Entity to provide AI services for small and medium-sized enterprises, and has integrated Lenovo's industry large model and rich industrial ecosystem resources on the underlying construction platform Baichuang Lab, to accommodate, develop, and promote AIGC software and hardware applications in thousands of industries on this basis.

In 2025, finding segmented application scenarios is still the main theme of AIGC.

Under the siege of major companies, the road of AI entrepreneurship is by no means smooth. Therefore, helping outstanding entrepreneurs survive better and promoting the rapid development of the AI era is also the value and responsibility of every leading enterprise like Lenovo that is fully exerting its strength in AI.

Entrepreneurship is also about breaking through difficulties. In the future, with the joint efforts of aiXcoder and Lenovo's ecosystem, enterprise development efficiency will be higher, code generation will be more accurate, and the liberation of productivity will be sufficient to make thousands of industries operate vigorously.