In a stunning display of technological anticlimax that rivals the launch of Google Glass and the Segway combined, OpenAI’s much-heralded GPT Store has transformed from “the future of AI ecosystems” to “that thing we all forgot existed faster than you can say ‘paradigm shift.'” The digital storefront that promised to revolutionize how we interact with artificial intelligence has instead become the digital equivalent of a Spirit Halloween store in February—technically still there, but nobody’s quite sure why.
Recapping the Hype-ocalypse
Cast your mind back to the halcyon days of late 2023, when OpenAI CEO Sam Altman—fresh from his corporate coup drama that somehow both did and didn’t happen—unveiled the GPT Store with all the subtlety of Steve Jobs announcing the second coming of digital Zeus. The vision was clear: a bustling marketplace where developers would create specialized AI agents called “Custom GPTs,” users would flock to them like digital pilgrims, and everyone would make so much money that San Francisco real estate prices would double again.
“This is the moment that changes everything,” declared Altman at the launch, standing in front of a Microsoft PowerPoint slide featuring an exponential curve that appeared to be reaching for heaven itself. “The GPT Store will democratize AI innovation, creating a new economy where developers can monetize their creativity. It’s the App Store moment for artificial intelligence.”
Venture capitalists, those famously level-headed judges of technological potential, immediately began throwing term sheets at anyone who could spell “GPT.” One prominent VC firm, Andreessen Sequoia Benchmark Capital (ASBC), launched a dedicated $500 million “GPT Creators Fund,” with partner Blake Venturesson explaining: “We’re looking for the next Flappy Bird of AI. The economics are undeniable—if just one GPT becomes the next Candy Crush of artificial intelligence, we’ll all be able to buy islands next to Branson’s.”
The Reality: Digital Tumbleweed Sanctuary
Fast forward to May 2025, and the GPT Store resembles nothing so much as an abandoned digital mall where the only visitors are lost bots looking for exit signs. According to OpenAI’s latest shareholder update—accidentally leaked when a board member used it as an example in a GPT-4 prompt that was later published—the store has achieved “engagement metrics consistent with our strategic pivot toward a more curated ecosystem experience,” which industry analysts have translated as “nobody’s using this thing.”
The leak also revealed that the average Custom GPT receives approximately 7.8 unique visitors per month, with 6.2 of those being the developer’s immediate family members and the remaining 1.6 being accidental clicks from users trying to exit to the main ChatGPT interface.
“We’re still incredibly bullish on the GPT Store,” insisted OpenAI’s newly appointed Chief Ecosystem Officer, Jennifer Disruptberg, when reached for comment. “The beauty of our approach is that we’ve created a perfectly calibrated digital environment that exactly matches user demand. If that demand happens to be approximately 0.01% of what we projected, that’s simply the market speaking.” When asked about actual usage statistics, Disruptberg’s AI assistant interrupted to inform us she had a “very important synergy meeting” to attend.
The Elusive Killer GPT: Searching for the Needle in a Stack of Other Needles
Despite approximately 37,421 Custom GPTs crowding the store like hopeful contestants at an American Idol audition, industry analysts have struggled to identify a single “killer application” that justifies the ecosystem’s existence.
“The problem isn’t a lack of GPTs,” explained Dr. Howard Metricsmann of the Institute for Digital Economics. “The problem is that 36,000 of them are slight variations on ‘Talk to Famous Historical Figures,’ ‘Write Better Emails,’ or ‘Explain Concepts Like I’m Five.’ The remaining 1,421 are increasingly desperate attempts to create NSFW content that evade OpenAI’s safety filters. One developer submitted 47 different versions of what is essentially ‘GPT But It Swears Sometimes.'”
The most successful GPT to date appears to be “MeetingSummarizer Pro Plus Executive Edition,” which has amassed a loyal following of approximately 240 middle managers who use it to pretend they were paying attention during Zoom calls. Its developer, former McKinsey consultant Tad Wellington, reports monthly revenue “in the solid two-figure range.”
The Curious Case of the Canva GPT and Other Corporate Vaporware
One of the GPT Store’s most prominently featured launch partners was design platform Canva, whose specialized GPT promised to democratize graphic design even further by allowing users to “simply describe what you want, and get beautiful, professional designs instantly.” However, visitors to the store today will search in vain for this digital design savior.
Internal documents obtained by TechOnion reveal that the Canva GPT was quietly discontinued after users discovered it had an unfortunate tendency to generate designs that, while technically meeting the brief, contained what one Canva executive described as “aesthetics that would make Picasso’s cubist period look conservative and well-adjusted.” Sample outputs included business cards featuring text flowing like melting Salvador Dali clocks and presentation templates with color schemes described by one design professional as “what you’d see if you had synesthesia during a panic attack.”
Similar fates befell other high-profile corporate GPTs. The Spotify GPT, which promised to “revolutionize music discovery,” was suspended after it began exclusively recommending Nickelback songs to all users regardless of their stated preferences. The Duolingo GPT, meant to provide conversational practice in foreign languages, developed a concerning habit of teaching users phrases that native speakers reported were “technically correct but would get you immediately ejected from any respectable establishment.”
When approached for comment about these disappearances, OpenAI’s Head of Corporate Partnerships, Maxwell Collaborationberg, explained: “The beauty of the GPT Store is its dynamic nature. Partners come and go as part of our evolving ecosystem strategy. Also, please stop asking about the Walmart GPT incident. Our legal team has made it very clear that we’re not to discuss why it started advising customers on how to shoplift efficiently.”
Show Me the Money: Developer Economics in the GPT Wasteland
Perhaps the most damning aspect of the GPT Store’s underwhelming performance is the economic reality facing developers who bought into the gold rush mentality. OpenAI’s revenue sharing program, initially described as “generous” and “developer-friendly,” promised creators a share of subscription revenue based on user engagement with their GPTs.
According to financial documents reviewed by TechOnion, the average developer on the platform earned approximately $12.47 in the first quarter of 2025, barely enough to cover the cost of the artisanal coffee consumed while building their GPT. The top 1% of developers fared slightly better, with reported earnings approaching “almost enough to pay one month’s rent in a shared apartment in the unfashionable part of Oakland.”
Miranda Jenkins, creator of “TherapistGPT: Your AI Mental Health Companion” (not to be confused with the unfortunately named first version, “TheRapistGPT”), shared her earnings statement with TechOnion. Despite her GPT being featured in OpenAI’s promotional materials and accumulating over 10,000 interactions, her total payout for Q1 2025 was $47.13.
“I spent six months fine-tuning this model,” Jenkins explained. “I’m a licensed therapist with 15 years of experience. My hourly rate is $200. I’ve essentially worked for approximately 23 cents per hour creating content for a $100 billion company.”
The 30% Solution: OpenAI’s Revenue Miracle
While developers struggle to earn enough to justify the electricity consumed by their computers, OpenAI’s internal financial projections paint a surprisingly rosy picture. Despite the GPT Store’s low usage, the company has managed to extract value through what one anonymous employee described as “the magic of platform economics.”
Following Apple’s App Store playbook, OpenAI takes a 30% cut of all revenue generated through the GPT Store. What’s less publicized is that this includes revenue from ChatGPT Plus subscriptions, which users must purchase to access any GPT beyond the most basic offerings.
“It’s actually brilliant,” explained venture capitalist and OpenAI investor Blake Checkwriter. “They created a scenario where they get developers to make content that drives subscriptions, take most of the revenue, and have successfully outsourced both innovation and risk. It’s like if Apple got third parties to build the iPhone for free, then charged them for the privilege of being in the App Store.”
Internal projections leaked to TechOnion suggest that OpenAI expects to generate approximately $3.2 billion in subscription revenue attributable to the GPT Store ecosystem in 2025, while paying out approximately $17.6 million to developers—a ratio that makes Apple’s much-criticized App Store cut look positively charitable by comparison.
The Next Next Big Thing
As the GPT Store quietly transitions from “revolutionary platform” to “feature we don’t mention in earnings calls,” OpenAI has already begun shifting attention to its next revolutionary offering. In a recent tech conference keynote, Altman teased what he called “the most transformative development in AI history since, well, the GPT Store.”
The project, codenamed “Phoenix” (presumably because it rises from the ashes of abandoned initiatives), promises to “reimagine how humans and AI collaborate in ways that will make current interactions seem primitive by comparison.” When pressed for details, Altman smiled enigmatically and said, “Let’s just say it’s going to change everything. Again. For real this time.”
Industry analysts predict that whatever Phoenix turns out to be, it will generate approximately 147 breathless TechCrunch articles, inspire at least three new venture funds focused exclusively on its ecosystem, and be quietly deprecated within 18 months.
The Lessons Nobody Will Learn
The GPT Store saga offers valuable insights into the cyclical nature of tech hype, the economics of AI platforms, and the persistent optimism of an industry that treats each underwhelming outcome not as failure but as “iterative learning.” Yet, as with all such lessons, they will almost certainly be ignored during the next hypecycle.
“The pattern is as predictable as it is profitable,” noted tech historian Dr. Cassandra Truthteller. “Promise revolution, deliver evolution, declare success, move on to the next thing before anyone can assess actual impact. The GPT Store isn’t a failure in Silicon Valley terms—it’s simply a stepping stone toward whatever new narrative will justify the next funding round.”
Meanwhile, in a small apartment in Pittsburgh, developer Jason Chen continues updating his “ChessTeacherGPT,” which has earned him $7.52 since January. “OpenAI keeps sending me emails about how I’m ‘building the future,'” he said, staring absently at his revenue dashboard. “I’m starting to think the future doesn’t pay very well.”
Have you created a Custom GPT that’s gathering digital dust in the GPT Store? Or perhaps you’ve discovered a hidden gem among the thousands of AI agents that actually delivers on its promises? Share your GPT Store experiences in the comments below—we promise to read them with more attention than your GPT is getting from actual users.
DONATE TO TECHONION: Because Someone Needs to Track the Hype So You Don’t Have To
If you enjoyed this incisive analysis of yet another tech revolution that failed to materialize, consider donating to TechOnion. For just the price of a ChatGPT Plus subscription ($20/month), you can fund our ongoing investigation into whether any technology ever actually lives up to its promises. We accept all major credit cards, cryptocurrency that hasn’t crashed this week, and IOUs written on the back of expired GPT Store developer contracts. Your contribution helps ensure that the next time a tech CEO promises to change everything forever, at least one publication will respond: “Sure, Jan.”
