OpenAI’s chatbot ChatGPT is about to reach a historic milestone, with its global weekly active users (WAU) expected to surpass the 1 billion mark soon. The generative AI product, which launched in late 2022, will have reached this scale in less than four years, making it one of the fastest-growing applications in internet history, even though this milestone is about seven months behind the company’s internal schedule.
According to internal OpenAI information obtained by the media, the company had originally aimed to hit 1 billion weekly active users by the end of last year, but actual progress was delayed. Market sources indicate that ChatGPT’s user growth momentum has slowed, primarily due to multiple competitive pressures. Google’s Gemini AI model is accelerating its deep integration with the search engine, which itself boasts a user base of billions, posing a direct threat. Furthermore, the development of OpenAI’s next-generation model, GPT-5, was fraught with setbacks when it launched last fall, and its release sparked criticism from some users, leading to a decline in user growth at the time.
Simultaneously, Anthropic, founded by former OpenAI executives, has seen its Claude chatbot successfully capture a portion of the market. Research data shows that Claude has gradually expanded its influence while ChatGPT faced headwinds.
Although ChatGPT still holds a lead in user numbers, the views of professional bettors in prediction markets on “who will be the strongest AI by the end of 2026” diverge significantly from public perception. According to data from Kalshi, a U.S. federally regulated prediction platform, total wagers on the “Best AI by End of 2026” contract have exceeded $7.6 million (approximately NT$248 million). Bettors currently give Anthropic’s Claude a 65.2% probability of winning, far higher than ChatGPT’s 14.6% and Google Gemini’s 10.2%. Claude’s odds have also recently risen by 2.7 percentage points, indicating that the market holds high expectations for its technical prowess and commercialization prospects.
Anthropic just released its latest flagship AI model, Claude Opus 5, last week. The company emphasized that Opus 5 delivers significant leaps in software engineering, business automation, and scientific research. Particularly in software engineering, Anthropic claims that Opus 5’s performance on the Frontier-Bench benchmark test is double that of its predecessor, Opus 4.8, while the execution cost per task has been further reduced.
Facing mounting pressure from competitors, OpenAI is actively adjusting its strategic focus, turning its attention to the enterprise market. CFO Sarah Friar revealed that revenue from corporate clients currently accounts for 40% of the company’s total income and is expected to rise to 50% by the end of this year. ChatGPT itself is also evolving from a simple Q&A bot into an “AI agent” with autonomous execution capabilities. It can now connect to Gmail and user calendars to handle tasks like making reservations and scheduling meetings, and it has launched a “super app” on the desktop that integrates multiple functions in an attempt to deepen user stickiness and commercial value.
On the infrastructure front, OpenAI’s ambitions are equally staggering. Nvidia is reportedly considering backing a financing package worth $250 billion (approximately NT$8.15 trillion) to help OpenAI drive its massive AI infrastructure expansion. This funding is part of a hyperscale data center project in southern Ohio, the total cost of which could ultimately exceed $500 billion (approximately NT$16.3 trillion), demonstrating that OpenAI is willing to invest heavily in computational power to maintain its lead.
However, OpenAI CEO Sam Altman has privately admitted to employees that the company faces “tough times ahead.” Current evaluations show that Google still lags behind OpenAI and Anthropic in AI program development progress, but Anthropic, with its strong enterprise revenue capabilities and prominent cybersecurity technology, has become the competitor OpenAI fears most. While ChatGPT’s user count is about to cross the 1 billion milestone, the ultimate outcome of this AI race will still depend on technological breakthroughs, the speed of commercial deployment, and a long-term battle for market confidence.