ChatGPT hits 200 million weekly active users

If you thought the generative AI hype had peaked after nearly two years, OpenAI's latest milestone suggests the engine is actually speeding up.

On Thursday, OpenAI announced that ChatGPT now boasts 200 million weekly active users. That is double the 100 million weekly active users the company reported back in November 2023, effectively doubling its audience in less than a year.

Giving away the good stuff

The secret behind this sudden surge basically comes down to aggressive upgrades and tearing down the paywall.

During its Spring Update, OpenAI rolled out GPT-4o along with an updated Voice Mode. More importantly, the company moved several flagship perks that were previously locked behind the $20-a-month Plus tier into the free version. Free users can now upload docs and photos, browse the GPT Store, and run advanced data analysis.

Meanwhile, growth on the enterprise side is moving just as fast.

OpenAI stated that 92% of Fortune 500 companies are currently using its products. On the developer front, API usage has doubled since the lightweight GPT-4o mini launched in July.

OpenAI is not the only one riding this wave. Meta also announced on Thursday that its open-source Llama models are closing in on 350 million total downloads, a massive tenfold leap from last year.

Why it matters

For regular users, ChatGPT is now at peak utility without costing a dime. Features like heavy document analysis and multimodal inputs used to require a paid subscription. If you only toyed with ChatGPT early on when it was mostly a parlor trick, now is a great time to jump back in and let it handle real everyday workflows.

For the broader industry, hitting 200 million weekly active users cements ChatGPT as the default consumer interface for generative AI. Once a product captures that level of daily habit, rivals cannot simply win on slightly better benchmark scores. At the same time, Meta's 350 million Llama downloads prove that the open-source camp is mounting serious pressure from below.

We think OpenAI's decision to hand high-end features to free users was brilliant for user acquisition. It essentially crowded out smaller consumer wrappers while funnelling developers toward cheap API models like GPT-4o mini.

That said, the model has an obvious vulnerability. Running inference for 200 million active users every single week is wildly expensive. Unless OpenAI can rapidly convert this massive mainstream presence into enterprise software revenue, carrying this immense compute tab will keep putting heavy pressure on its balance sheet.