With the arrival of the AI era, we have witnessed the rise of large-model players, the continued maturation of embodied intelligence, and the emergence of one AI agent after another. As AI begins to exert a deep and comprehensive influence on how people produce and live, it is increasingly clear that AI is not merely a single-point battle but a comprehensive, collaborative, ecosystem-level campaign.

Therefore, as AI development enters deep waters—especially as AI becomes increasingly intertwined with commercialization—we need to make one thing clear: in the AI era, what players compete over is not merely AI itself, but the ecological moat built around AI as the core. For AI players, how to build a strong ecosystem—rather than keeping their sights fixed on AI alone—may be the real key to truly capturing the dividends of AI.

We all know that when AI first began to emerge, major companies launched their own AI tools, each attempting to penetrate every corner of their existing industrial ecosystems with AI elements. Ultimately, one very important reason they did so was to better unleash AI's collaborative and cumulative functions and roles, rather than focusing their attention solely on AI itself.

As AI gradually matures—especially as it begins to exert a deep and comprehensive influence on people's lives—how exactly should we build the AI ecosystem? And how should we build AI so that it can truly unleash its cluster effect?

AI Applications Require a Diverse and Rich Ecosystem

As AI development moves from the capital-driven stage to the commercial-driven stage, we see that more and more players are beginning to focus on the implementation and application of AI. This approach is actually in line with the laws of AI development. However, we need to recognize that if AI applications remain confined to one or a few specific scenarios, AI's functions cannot be maximized.

Expanding AI applications as much as possible—especially breaking the practice of keeping AI confined to a limited number of scenarios, truly enabling AI to be implemented in as many scenarios as possible, and realizing collaboration among these application scenarios—is the real key to maximizing AI's functions and roles. From this perspective, when we examine the issue of AI applications, we should not merely focus on the applications themselves but should focus on diverse and rich applications.

Only when we truly build a large ecosystem with diverse and rich AI applications and realize collaboration and coordination within such an ecosystem can we truly maximize AI's functions and roles. For vertical AI players, what they need to do is open-source themselves into as many scenarios as possible and establish collaboration and coordination between themselves and external application scenarios. Only by truly implementing AI applications in more scenarios and establishing collaboration with more scenarios through open-source approaches—thereby forming an organic ecosystem—can AI's functions and roles be truly maximized.

AI Infrastructure Requires Continuous Improvement

As AI begins to be implemented and applied in more and more scenarios, we should not merely see AI itself but should truly recognize the important role played by the infrastructure supporting AI. If we regard AI as the "facade," then AI infrastructure is the steel bars and cement supporting these "facades." Therefore, when we look at AI, we should not merely keep AI at the level of AI itself but should also see the related infrastructure that truly supports AI.

In a sense, the more urgent the application of AI is, the more continuous refinement is needed in building and creating AI infrastructure. Only by truly building strong infrastructure for AI and using this infrastructure to continuously provide energy for AI can we truly bring AI development into a brand-new stage and truly maximize AI's functions and roles.

Regarding AI infrastructure, we see many aspects such as computing power, data collection, hardware facilities, and talent reserves. For AI—especially for AI that has already entered a brand-new stage of development—only by continuously building and improving its own infrastructure can AI development truly be brought into a brand-new stage. From this, we may boldly predict that in the future, the players who can truly make a difference in the AI era will certainly be those who have built strong infrastructure and can truly better empower AI applications through the ecosystem.

AI Capabilities Require Continuous Evolution

What currently impresses us more is the strong iterative capabilities of different players in their own foundation models. Whether for vertical large-model players or for large giant players, they are always continuously realizing the evolution of their own AI capabilities. AI large models updated every so often and AI capabilities iterated every so often are the most direct manifestation of this phenomenon.

From this perspective, in an era of rapid AI changes, there is no true sense of being ahead; only continuous evolution—only continuous iteration and updating—can truly win in the fierce competition of the AI era. Therefore, in the AI era, the real winners are those players who can continuously improve their own capabilities, and the real winners belong to those players who truly build leading capabilities.

Only by truly focusing on continuously iterating AI capabilities and using them to drive AI development can we truly bring AI development into a brand-new stage and truly win in such a brand-new competition. In other words, in the AI era, there are no winners in the strict sense; the real winners are those players who can continuously iterate and update themselves.

Conclusion

As the certainty of AI grows greater—especially as AI begins to move from the laboratory to more commercial scenarios—we increasingly see that the power of AI is not merely AI itself but an ecosystem built around AI. At this point, what AI needs to do is continuously enrich and improve its own applications, continuously build and improve its own infrastructure, and continuously realize the evolution of its own capabilities. When AI is no longer limited to AI itself but achieves collaboration and open-source sharing, its functions and roles can be maximized, and true AI inclusiveness can be realized.