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.
Once nearly synonymous with AI, it just got surpassed in valuation by Anthropic. Now it looks like it’s also going to get beaten to the IPO starting line.
What seems like eons ago, way back in December of 2024, we wrote a handful of predictions about what OpenAI might become.
In one, we said that “OpenAI is Lyft,” an analysis that involved Sam Altman’s world-beating AI company fumbling the bag so hard that it would lose its lead in the industry. It seemed pretty unlikely: it was the market leader at the time, and “ChatGPT” was on the verge of being AI’s version of a “Kleenex” or “Band-Aid” — a product so popular that one brand name becomes the term people use for it, no matter what company makes it.
We argued that OpenAI could lose its lead. It might become the Lyft to someone else’s Uber in the AI industry. Back then, we wrote:
“OpenAI may face a similar fate. An early entrant that got beat at its own game. Not your first choice but the cheaper alternative. The one that doesn’t become a verb.”
And:
“Let’s look at OpenAI’s signature product, ChatGPT. Its chatbot hasn’t proved to be any better than others. Truly, none of the chatbots have demonstrated they have any secret sauce. If you found somebody who’s been living under a rock for the past five years and asked them to use ChatGPT, Claude, Gemini, or Perplexity, I don’t think they’d be able to tell you which one has the most resources; they’re all more similar than they are different.”
It turns out, the differentiation in AI products eventually happened — and it’s a big part of the reason OpenAI fell behind. Anthropic leaned hard into coding and other enterprise work, betting it could get companies to pay for AI, and so far it has been right. Meanwhile, OpenAI lost focus, releasing a host of products such as its short-lived, cash-draining Sora, a video app where people could create realistic videos of Altman grilling a Pikachu.
Sure enough, OpenAI has become an example of what every business school strategy professor tries to hammer into their students’ heads: having a big first-mover advantage does not equate to a durable moat for your business.
That idea is really starting to surface now. One of its main competitors, Anthropic, which OpenAI once discussed merging with, just raised a funding round that values it more than $100 billion higher than OpenAI. Anthropic also just filed its IPO paperwork confidentially with the SEC, meaning it will likely make its stock market debut before OpenAI does, too. (There’s a little gray area here — it’s possible that OpenAI may have filed its paperwork confidentially and it just hasn’t been reported yet, though this seems unlikely given the amount of reporting happening in the AI space and recent statements by Altman.)
Past that, there are also warning signs around OpenAI’s financial performance. The Wall Street Journal wrote in April that OpenAI Chief Financial Officer Sarah Friar told fellow executives she’s worried the company might not be able to pay for future computing contracts if revenue doesn’t grow fast enough. The same day, The Information reported that OpenAI had a goal of 1 billion weekly active users by the end of 2025, but as of February 2026, it was still only 90% of the way there.
The fact that the company’s CFO was vocally expressing concerns inside its four walls about revenue growth, and was doing so during the run-up to one of the most hotly anticipated IPOs in history, is concerning for the prospects of that IPO. Friar has experience taking two different companies (Square and Nextdoor) public. She has also raised boatloads of money at sky-high valuations after investors got to scrutinize OpenAI’s books.
She knows how to go through those motions — and what she was likely doing was tempering expectations ahead of OpenAI’s numbers becoming incredibly public and incredibly scrutinized sooner rather than later. If Friar has concerns about future revenue growth, then she may have concerns about valuation targets — which have been reported as high as $1 trillion — too.
Altman’s message constantly seems to be “we don’t have enough compute” — but if Friar is indicating there are revenue problems, that’s a whole different can of worms. Compute needs = spending money. Revenue problems = making money. Historically, one of those is a lot easier to do than the other.
These aren’t necessarily direct comparisons, but in May, The Wall Street Journal reported that Anthropic’s second-quarter revenue would double to $10.9 billion. OpenAI itself said at the end of March that it was “now generating $2B in revenue per month.” I’m not an expert on the whims of seasonality in AI, and perhaps there are some, but $10.9 billion a quarter is a lot higher than $2 billion a month.
If you’re unfamiliar with the startup world, revenue is usually the easy part. For more than a decade, insanely well-capitalized startups like OpenAI have followed a model of converting venture capital dollars into blockbuster revenue and user growth.
Usually, turning those users and revenue into cash flow is the harder task, one that can take much longer. And a company’s stock price — which is something OpenAI will soon care a lot more about — is typically a function of the present value of future cash flows. (Of course, that’s not always true, especially when a company picks up a cult following.)
Unfortunately for OpenAI, a cult following seems to be exactly what it doesn’t have.
What if the biggest constraint to OpenAI’s growth is actually that demand is fading for its products? An executive whose job revolves around AI every day recently told me, in unrepeatable terms, that OpenAI is far from the preferred AI option today. Data providers like Similarweb and Apptopia seem to back that up, showing OpenAI losing market share month after month.
Don’t take this too literally, of course. OpenAI is orders of magnitude larger than Lyft, a $5 billion ride-hailing company. OpenAI is still the market leader in terms of users, even if its market share is being gradually shaved off and it’s not generating as much revenue from its users. It will no doubt still be a huge IPO, and Sam Altman just this week told CNBC that he wasn’t in a race with Anthropic to go public, but rather that they are in “a race to deliver the best technology, build the best business.” And that’s probably the best way to look at it! Hey, maybe we should be looking past the swoon in market share and asking different questions altogether.
Enterprise seems to be the spot to win in AI, and OpenAI seems to finally realize that, having added more tools for Corporate America recently. But at least for now, OpenAI, which started the AI revolution by releasing ChatGPT just after Thanksgiving in 2022, has let its first-mover advantage erode.
And who has it ceded market share to? Oh, just a group that includes Google, Meta, Microsoft, several other well-capitalized companies, China, and to a lesser extent, Amazon and Apple. Not an easy playing field on which to stage a comeback.
Come to think of it, maybe we’re also a little off when comparing OpenAI to Lyft — because Lyft really only had one gigantic rival trying to eat its lunch.
It took Facebook and Instagram around eight years; it took YouTube just over six; even TikTok, which at the time felt like it was a global sensation almost as soon as it arrived, took more than half a decade.
Now, though, the mobile version of ChatGPT has positively left the biggest platforms (and all of your other favorite apps) in the dust, hitting 1 billion monthly active users in just three years, per new data from market intelligence firm Sensor Tower, as more users turn to OpenAI’s chatbot each month.
While rival Anthropic might be pulling ahead in terms of annualized recurring revenue, enterprise customer adoption, and valuation, the app version of Claude, a market-leading chatbot on several counts, has clocked only 56 million monthly active users in the quarter to date.
In fact, according to Abe Yousef, a senior insights analyst at Sensor Tower, ChatGPT’s monthly active user count for the quarter to date outweighs the figures for Claude, Gemini (472 million), Doubao (106 million), Dola (78 million), DeepSeek (68 million), Meta AI (61 million), Grok (50 million), Perplexity (44 million), and Copilot (31 million)… combined.
ChatGPT made a pretty big splash in the tech world when it landed toward the end of 2022, but there’s no question that the mobile versions — which launched on iOS in May 2023, then on Android a couple months later — helped to catapult the chatbot into the mainstream proper.
There are fewer humans working in Amazon warehouses these days, but those that are still there can at least talk to robots.
At its Delivering the Future event in London, the e-commerce giant unveiled the next generation of Proteus, its autonomous warehouse robot. Instead of requiring complex coding, workers can now give the machine verbal instructions in plain language, like telling it to haul a heavy cart across the floor.
While Amazon’s older generations of warehouse robots were restricted to fenced-off loading docks, Proteus is a fully untethered model that uses AI to safely navigate the entire fulfillment floor alongside human staff. The new robot is the centerpiece of a massive €10 billion ($11.6 billion) investment to modernize Amazon’s European logistics network and is currently being piloted in company labs before a planned rollout in early 2027.
While Amazon’s older generations of warehouse robots were restricted to fenced-off loading docks, Proteus is a fully untethered model that uses AI to safely navigate the entire fulfillment floor alongside human staff. The new robot is the centerpiece of a massive €10 billion ($11.6 billion) investment to modernize Amazon’s European logistics network and is currently being piloted in company labs before a planned rollout in early 2027.
Meta has repeatedly delayed the release of developer access to Muse Spark, its newest AI model, according to The Wall Street Journal. While the model launched in April and powers Meta’s AI products, developers outside the company have been kept waiting for access to the API.
That’s a glaring bottleneck for a company spending up to $145 billion on AI infrastructure this year: without an API, Meta can’t easily sell access to the model, ceding a lucrative monetization engine to rivals like OpenAI and Anthropic.
Meta told The Wall Street Journal that API access would be available this month.
That’s a glaring bottleneck for a company spending up to $145 billion on AI infrastructure this year: without an API, Meta can’t easily sell access to the model, ceding a lucrative monetization engine to rivals like OpenAI and Anthropic.
Meta told The Wall Street Journal that API access would be available this month.
The semiconductor company reported second-quarter earnings after the bell Wednesday.
