In 2025, OpenAI harbors hidden worries beneath its peak performance. Despite ChatGPT launching a high-end reasoning model that claimed victories in international math Olympiads and top programming competitions, with annualized revenue surging to $19 billion (approaching the $20 billion target), user growth has slowed significantly. Its weekly active users have failed to meet the 1 billion expectation, as the core product is trapped in a competitive quagmire by issues such as diversified expansion and strategic divisions.
The core contradiction of OpenAI lies in the severe disconnect between cutting-edge research and public demand. A research team of over 1,000 people is focused on “Artificial General Intelligence (AGI)“. While the developed reasoning models can tackle complex problems, they come with high computing costs and slow response speeds. Most users, however, only need to ask simple questions such as movie ratings and daily consultations. This “performance surplus” results in underwhelming performance when advanced models are transformed into products. Even when embedded as “thinking mode”, only a handful of users utilize them frequently.
The multi-front battle promoted by CEO Sam Altman has further exacerbated the crisis. Beyond ChatGPT, OpenAI is simultaneously advancing numerous ambitious projects including Sora video generation, AI hardware, and robots, which continuously divert core resources and weaken investment in ChatGPT’s optimization. The internal culture of “research first” dominates decision-making, neglecting product experience and leaving OpenAI vulnerable to Google Gemini’s fierce counterattack. Leveraging its massive ecosystem including Gmail and Chrome, Google seamlessly integrates AI capabilities into users’ existing workflows. Gemini’s monthly active users have reached 650 million, with average visit duration surpassing ChatGPT’s since September. OpenAI’s wavering strategy on image generation functions has exposed problems of delayed decision-making, a reflection of the challenges brought by multi-front operations.
The paradox between commercialization and user growth is equally prominent. OpenAI has achieved explosive revenue growth through subscription services, aiming for a valuation of $750 billion, but its strategy of focusing on monetizing existing users has hindered the expansion of its user base. In the corporate customer market, its share has dropped to 27%, lagging behind Anthropic’s 40%. Meanwhile, the company burns billions of dollars annually on computing costs, and its planned $1.4 trillion infrastructure investment urgently requires stable cash flow support.Charging for ChatGPT
To address the crisis, OpenAI has issued an eight-week “red code” alert, refocusing resources on ChatGPT’s core functions, postponing short-term profit-driven projects, and launching countermeasures such as GPT-5.2 and a new image generation model. However, the hasty rollback of the model routing system—implemented for just four months—proves once again that pure technological upgrades cannot compensate for shortcomings in user experience. This system had increased the proportion of free users accessing reasoning models from less than 1% to 7%, significantly raising costs while negatively impacting daily active user metrics due to slow response times. More critically, Altman’s layout for an AI hardware strategy may trigger direct competition with Apple in the future, a development closely watched in the latest AI news.
OpenAI stands at a crossroads of fate: should it adhere to its original aspiration as a laboratory dedicated to AGI research, or transform into a product-focused company that deeply caters to user needs? This crucial decision on strategic positioning will determine whether it can break through the encirclement of tech giants and maintain its first-mover advantage in the AI survival battle.
