Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.
There’s a ‘who’s the fairest of them all’ debate going on in this AI Tech Wave on who’s winning the AI chatbot wars right now. Especially given how fast OpenAI is building on it industry lead with the ChatGPT chatbot. Now in its fourth year out in the wild globally. With both consumers and enterprises small and large.
Especially on the latter vs Anthropic’s Claude and Claude Code/Cowork products vs OpenAi’s Codex and other AI offerings.
The Information provides the beginning context in “OpenAI’s GPT-5 Dip; Why Agents Are Hard to Evaluate”:
“Before we get into today’s column, Friday’s story from Sri and me confirmed what many had been hypothesizing since OpenAI’s release of GPT-5 in August: ChatGPT’s growth hasn’t been as strong as OpenAI hoped.”
“The story is more nuanced, obviously. The chatbot has hit new highs in terms of weekly active users—910 million, to be exact as of earlier this month. But OpenAI leaders had hoped to hit 1 billion WAUs by the end of 2025.”
“A slowdown in growth began with the rocky release of GPT-5 last summer, in which a number of users complained that the model had a smaller-than-expected jump in capabilities and a colder personality than its predecessor GPT-4o.”
But there is some good news for the OpenAI supporters of late:
“OpenAI leaders told investors last week, however, that ChatGPT growth has picked back up again this year, thanks to the GPT-5.1 and GPT-5.2 updates, which gave the model a warmer personality and better performance on workplace-related tasks. They expect WAUs to hit 2.75 billion by 2030, meaning that approximately a third of the world will be using ChatGPT on a weekly basis by then.”
“To get to that ambitious goal, OpenAI is going to have to make ChatGPT a whole lot more useful to a whole lot more people. Maybe it’ll integrate ChatGPT into its family of devices, which could propel the chatbot into stardom if those devices become as ubiquitous as the iPhone. (For more on what’s coming in that family of devices, check out my story from Friday here.) Or maybe it’ll ramp up its expansion into new geographic regions with its cheaper ChatGPT Go subscription. Whatever their plan is, OpenAI leaders are definitely setting the bar high for themselves.”
The VC Firm Andreessen Horowitz, of ‘Software Eats the World’ fame, made the pro OpenAI/ChatGPT growth case for its portfolio company on x/Twitter:
“Chatbots are sticky.”
”Charts of the week:“
They elaborated in a separate piece:
“How much do people love their LLMs? Better and better with time, apparently.”
“Retention curves do not usually look like this:”
“ChatGPT’s average retention rates keep getting better from week 2 onwards, and even more impressively, inflect upwards at week 23;”
“Gemini, for its part, has a “smiling” retention curve, whereby rates dip initially (as per usual), but then pick up again at week 10 (and continue to climb).”
“A smiling retention curve is generally considered very impressive. It’s common for users to drop off over time, but it’s special when they start coming back over even more time.”
“But a retention curve that just keeps going up, and doesn’t even have a chance to smile? That’s basically unheard of, especially at this scale.”
“The other thing that’s pretty impressive about LLM usage is that active users are becoming more active:”
Tech/AI growth analyst Aakash Gupta says not so fast:
“ChatGPT’s retention is climbing to 70%+ at 24 months. Looks dominant. But here’s what a16z isn’t showing you: ChatGPT’s mobile app market share fell from 69.1% to 45.3% in the last 12 months, according to Apptopia. Gemini went from 14.7% to 25.2%. Grok went from 1.6% to 15.2%.
”This chart is measuring whether people who already chose ChatGPT keep using ChatGPT. The answer is yes. The question it doesn’t answer is whether new users are still choosing ChatGPT. The answer is increasingly no.”
He then makes the crux of his argument on OpenAI’s position thus far:
”The retention curve is a lagging indicator disguised as a leading one. High retention on a shrinking share of new adopters means you’re keeping your existing base while losing the growth battle. Netflix had 90%+ retention in 2019 while Disney+, HBO Max, and Peacock were capturing every marginal subscriber.”
”The total AI chatbot market grew 152% year over year. ChatGPT’s absolute visits went from 3.8 billion to 5.7 billion, a 50% increase. Gemini went from 267 million to 2 billion, a 647% increase. When your market triples and you grow 50%, your competitor grew 7x.”
”One in five AI users now uses multiple apps. Claude leads average time spent per daily user at 34.7 minutes. The market is fragmenting by use case, and the retention chart measures loyalty to one tool while users are building multi-app stacks.”
”This is the classic incumbent trap: celebrating retention while the acquisition funnel quietly redirects elsewhere. The cohorts getting stickier are the ones who signed up 18 months ago. The cohorts signing up today have more options, and they’re taking them.”
My take on this pro-con debate is more in Aakash’s camp for now. Adding context with previous tech waves over three decades. As I explained on x/Twitter:
“It’s AOL vs Yahoo!/Google et al all over again with ChatGPT vs Google Search/Gemini et al today.”
Explained clearly by Aakash Gupta:”
“This is the classic incumbent trap: celebrating retention while the acquisition funnel quietly redirects elsewhere.”
“The cohorts getting stickier are the ones who signed up 18 months ago. The cohorts signing up today have more options, and they’re taking them.”
”As Goldman Sachs Internet Analyst in 1990s, I saw similar dynamic for AOL vs Yahoo!, Google et al Internet alternatives.
Google won back then too. “
My arguments build on what I summarized in five key points I made in an earlier post #970, “AI: OpenAI’s ‘AOL Moment’”, last month.
The crux being that OpenAI has a lot of wood to chop, especially on the distribution front, given the Google + Apple realities to shortly unfold across the world’s multi-billion smartphones.
Also in my view, we need to wait counting the ‘mission accomplished’ winners at least a year further. Especially as Apple launches its AI powered Siri n the first half of this year, leveraging Google’s Gemini AI in proprietary forms, and providing massive global distribution to the same.
We’ll know where all the LLM AI companies fare further this AI Tech Wave soon enough.
For now, all the participants are winners. Including OpenAI, Google, and others are winning in an AI market just getting started.
But that doesn’t mean that observers won’t keep score every lap. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)
No posts