Are AI’s Consumer Applications Hitting a Wall?

This week, the market erupted after OpenAI’s disappointing revenue numbers came to light in the Wall Street Journal. The company is barreling toward an (inevitable) $1 trillion+ IPO, and any hiccup in its story drives a mini-panic. After the chaos settled down though, the bigger, underlying story baked into the numbers got almost no attention: Consumer AI might be plateauing.

Generative AI’s current moment began with a consumer application in ChatGPT, but the technology is still struggling to find its true form in consumer apps, and even ChatGPT’s torrid growth seems to be ebbing. OpenAI, according to the Journal’s report, failed to hit its goal of 1 billion active ChatGPT users by the end of 2025 and may not even be there today (it reported 900 million weekly active users this February). Among chatbots as a whole, daily active user growth has flatlined, according to third-party analytics companies like Apptopia.

“If we include April, DAUs have fallen for 4 out of the past 5 months,” Adam Blacker, director of public relations at Apptopia, told me.

While enterprise generative AI apps are taking off in the legal field, medicine, software development, and elsewhere, the corresponding consumer breakout hasn’t materialized. ChatGPT’s widespread use is certainly an exception, but the technology otherwise hasn’t translated to the series of consumer hits many anticipated at the outset. There is no mainstream generative AI companion, fitness app, life coach, or fantasy adventure game, for instance.

Consumer AI’s bumpy progress compared with enterprise’s boom was put in stark contrast in this week’s big tech earnings reports. If your company owned infrastructure supporting enterprise AI, you were sitting pretty. Amazon reported 28% growth in its Amazon Web Services division, a jump from the 17-18% quarters it had turned in recently. “The last time we saw growth at this clip, AWS was roughly half the size,” Amazon CEO Andy Jassy said. Google Cloud Platform, meanwhile, grew 63% and Microsoft Azure grew 40%.

The consumer side of tech earnings was a different story. Meta told investors it would spend even more money in its pursuit of ‘personal superintelligence’ than anticipated (up to $145 billion this year) and the market punished it, cutting it 8.5% on Thursday. Apple, the notorious AI ‘failure,’ felt none of the pain after its consumer products, including the iPhone and MacBook Neo, delivered stellar results. Apple, for the moment, seems wise to have forgone a massive investment in foundational model training in service of building a little-wanted consumer application.

To date, consumer AI has surged mostly through novelty. ChatGPT’s Studio Ghibli moment created a surge of interest — turning many new users on to the power of generative images — but the excitement ultimately faded. OpenAI’s enhanced voice mode for ChatGPT also created a spike in users, but the company still has work to do for that feature to reach its potential.

This doesn’t mean the AI moment is running out of steam. New agentic apps like Codex and Claude Code are booming in adoption. Anthropic is on track to make more than $30 billion this year on the back of Claude’s coding capabilities. And OpenAI’s latest GPT-5.5 model doubled Codex’s revenue in less than a week. Perhaps these ‘super apps’ will grab the public’s attention and become as standard as a phone or a laptop. But for the time being, the consumer slowdown is curious.

Enterprise AI is moving from promise to production (sponsor)

I'm at ServiceNow’s Knowledge conference next week getting the real story from the people building it. New conversations dropping on my YouTube page starting mid-week.