AI Power Ranking 2025 Interpretation: Who Is Writing the Script for the AI Industry?

Industry Context

The 2025 AI Power Index released by Observer is not a simple celebrity list. Using two forms of power—"capital" and "ideas"—as its framework, the list selects 100 leaders with the ability to shape the AI industry. Behind this lies a dual characteristic never seen before in the AI industry: on one hand, technological breakthroughs are penetrating various industries at an unprecedented speed; on the other hand, the risks that come with concentrated power are intensifying in tandem.

As the list's introduction points out, AI's influence is more immediate and more widespread than that of any previous technology. From drug development to education reform, AI demonstrates enormous potential; however, data rights and ethical issues in model training also make this power deeply controversial. This contradiction makes "documenting power" necessary, but it is not the same as "endorsing power."

The list also emphasizes that AI is no longer just a technological competition—it is a geopolitical contest. Countries treat AI development as a matter of technological sovereignty, and export restrictions and trillion-dollar infrastructure investments are reshaping the global balance of power. In this context, leaders who can sustain influence over the long term are the ones truly worth watching.

Market Impact

Changes in the ranking directly affect market sentiment and capital flows. The top-ranked leaders hold the say in resource allocation.

Sam Altman (OpenAI CEO) tops the list. His influence derives not only from the pace of product iteration but also from striking commercial growth. As of July 2025, ChatGPT had 700 million weekly active users, and OpenAI's annualized revenue reached $12 billion—double what it was at the start of the year. In addition, OpenAI struck a $30 billion cloud deal with Oracle and signed an agreement with Microsoft that allows it to reorganize as a for-profit company. These moves not only change the company itself but also ripple through the entire cloud computing and data center investment market.

Jensen Huang (NVIDIA CEO) ranks second, representing the absolute dominance of AI infrastructure. NVIDIA's GPU supply determines the scale and speed of AI training, and its market position cannot be ignored by any AI company.

Other leaders on the list, such as Dario Amodei (Anthropic CEO), Elon Musk (xAI founder), and Satya Nadella (Microsoft CEO), influence the industry chain from different angles—model R&D, capital operations, and enterprise market strategy. For investors, this list effectively provides a "map" of risks and opportunities: with power concentrated in a few hands, investment decisions are more vulnerable to the strategic moves of individual leaders.

Competitive Landscape

• The ranking reveals the current competitive dynamics of the AI industry:- Model layer competition: Labs such as OpenAI, Anthropic, Google DeepMind, and Meta AI lead frontier breakthroughs. Sam Altman ranks first, the Amodei siblings from Anthropic are in the top five, and Google DeepMind's Demis Hassabis ranks 14th, showing that the battle between closed-source and open-source paths continues.
• Infrastructure layer: NVIDIA (Jensen Huang) and AMD (Lisa Su) represent compute supply, while cloud giants such as Oracle and Microsoft compete for the AI compute market through massive contracts (e.g., the OpenAI-Oracle deal).
• Application layer and enterprise services: Accenture's Julie Sweet, Salesforce, and others on the list indicate that enterprise AI applications have entered the stage of commercial deployment.
• Capital and thought factions: The list can be divided into "capital factions" (e.g., Peter Thiel, Vinod Khosla) and "thought factions" (e.g., Geoffrey Hinton, Yoshua Bengio). The two drive each other forward, but there is also tension between "rapid advancement" and "responsible development."- 12 months: Leading labs such as OpenAI and Anthropic will continue releasing more powerful models, but disputes over data copyright and safety will intensify. More quantifiable ROI cases will emerge in enterprise AI applications.
• 24 months: Investment in computing infrastructure will continue to overheat, and there may be local bubbles or consolidation. Geopolitical factors will increasingly affect chip exports and model deployment.
• 36 months: Power in the AI industry may shift from model developers to companies that own data and distribution channels. Competition between open-source and closed-source models could reshape the market structure, and regulatory frameworks (such as the EU AI Act) may become actual constraints.

For investment institutions, paying attention to the "new faces" on the list—such as areas engaged in AI safety and AI for Science—may reveal the next round of growth opportunities.