ChatGPT Paid Subscriptions Plateau in Europe Amid Conversion Challenges

ChatGPT paid subscriptions have stalled across major European markets since May 2025, according to a new Deutsche Bank report that raises questions about OpenAI’s ability to convert its massive free user base into sustainable revenue.

Consumer spending on ChatGPT in the UK, Germany, France, Italy, and Spain has remained essentially flat through September 2025, breaking a pattern of seasonal rebounds observed in previous years. Deutsche Bank analysts view the plateau as structural rather than seasonal, citing the fact that the typical post-summer recovery failed to materialize in 2025.

The slowdown comes despite OpenAI’s ChatGPT reaching 800 million weekly active users globally as of October 2025. Yet only about 5 percent of those users pay for subscriptions, leaving OpenAI with roughly 40 million paid subscribers worldwide. The gap between user growth and revenue conversion has emerged as the company’s most pressing commercialization challenge.

In Europe, ChatGPT now attracts more paid subscribers than Disney Plus, but consumer spending on the AI tool remains about half that of Spotify and one-quarter that of Netflix in the same markets. The €22–24 monthly subscription price for ChatGPT Plus in Europe—which includes VAT—matches the pricing of competing premium AI services from Google, Anthropic, and Perplexity.

Market saturation and intensifying competition appear to be driving the stagnation. Google Gemini, Anthropic’s Claude Pro, and Perplexity Pro all offer comparable features at nearly identical price points, fragmenting a market that OpenAI once dominated. Many users find ChatGPT’s free tier sufficient for their needs, limiting the perceived value of upgrading.

The plateau has significant implications for OpenAI’s business model. The company has committed to massive infrastructure spending, including billions of dollars in deals with chip manufacturers such as Broadcom and Nvidia. Deutsche Bank analysts suggest that the bottleneck in attracting new paying customers could limit OpenAI’s capacity to cover these costs and maintain its aggressive expansion.

European Surge then Slowdown

ChatGPT subscription growth in Europe surged more than 20 percent year-over-year in 2023, but that momentum has evaporated. In 2024, summer slowdowns attributed to students being less active during holidays were followed by strong recoveries in June and September. This year, both July and August remained flat, and September brought no uptick.

The UK market illustrates both the opportunity and the challenge. Weekly ChatGPT Plus spending in the UK reaches £5 million, outperforming Disney Plus and showing faster growth than Netflix or Spotify. Yet total spending on the platform still trails the established streaming giants, and the conversion rate from free to paid users remains stubbornly low.

OpenAI’s flagship ChatGPT Plus subscription offers access to GPT-4 and GPT-5 models, DALL-E image generation, advanced voice capabilities, and the GPT Store. Competitors have closed the feature gap: Google bundles Gemini Advanced with 2TB of cloud storage and tight integration with Workspace, while Claude Pro offers a larger context window preferred for handling long documents and complex analysis.

The European data aligns with broader trends in AI monetization. While ChatGPT processes between 2.5 and 2.6 billion messages daily, translating usage into recurring revenue has proven more difficult than scaling user numbers. The challenge extends beyond Europe; recent analyses show ChatGPT’s mobile app is also experiencing slowing download growth and daily use in other regions.

Deutsche Bank’s findings suggest OpenAI faces a critical test of whether widespread adoption can translate into the revenue growth needed to support its operational scale and infrastructure investments. With European spending growth at or near zero since mid-year, the coming months will indicate whether this plateau represents a temporary ceiling or a more fundamental constraint on the AI subscription model.