OpenAI's Phantom Deals Are Collapsing and the People Who Called It Are Feeling Vindicated

Three days. That was the gap between a Bluesky commentator saying OpenAI's uncommitted megadeals would "ultimately be revealed and collapse" and RAM prices beginning to drop because OpenAI couldn't follow through on promises to purchase roughly 40 percent of the world's supply. The post, which drew nearly 500 likes, wasn't a prediction so much as a countdown — and when it landed, the response wasn't shock. It was vindication. "AI is about to die," read one reply with 146 likes. "How loud are you cheering?"

The details behind OpenAI's phantom deals matter less than what the reaction reveals: a community that spent months absorbing bullish announcements has quietly reorganized itself around skepticism. The same week RAM prices moved, Sora was shut down, a billion-dollar partnership with Disney collapsed, and the company's erotic chatbot project was quietly shelved. OpenAI is losing products, partnerships, and the benefit of the doubt all at once, and on Bluesky the three events are being processed not as separate news items but as a single verdict. The mood curdled fast. Posts that would have read as cautious skepticism a month ago now read as prophecy.

What's striking about this moment isn't the negativity — it's how quickly the financial logic caught up with the critics. The business model skeptics on Bluesky weren't wrong to doubt whether you could build a company by reserving commodity hardware you couldn't actually pay for. OpenAI had reserved 40 percent of the world's RAM and nobody stopped them — and now the downstream effects are visible in component prices. Two separate Bluesky voices made the same point independently: the companies pressing hardest for AI adoption might be doing so precisely because they don't have a sustainable model yet, betting that lock-in will arrive before the money runs out. That argument has circulated in tech criticism for years. In the last week, it started getting upvotes.

Meanwhile, the job displacement stories kept arriving with their own momentum. A Kingdom Come Deliverance 2 translator posted that he'd been fired and replaced by AI — "I want you to know that the growing use of AI greatly affects people in the games industry and many others" — and the post spread well beyond gaming circles. It landed in the same week that a Resume.org survey found 59 percent of hiring managers admit they invoke AI when explaining layoffs because it plays better with shareholders. The translator's story and the survey data are about different industries and different mechanisms, but they're being read together as the same story: companies are using AI as cover for cost-cutting they intended to do anyway, and real people are losing real jobs while the companies report the savings as innovation.

Across the conversation right now, the question that keeps surfacing isn't whether AI will transform industries — that argument is largely over — but whether the companies leading the transformation can survive long enough to collect on it. OpenAI's Codex plugin launches, its disaster response partnerships in Asia, its new Chief People Officer appointment: all of it is happening against a backdrop of financial credibility questions that its own deal collapses created. The agent race — Anthropic, Google, and OpenAI all shipping agent frameworks in a six-week window — reads differently when one of the three competitors just proved it was promising infrastructure it couldn't afford to build. The companies will keep shipping. The question is who's still solvent when the consolidation comes.