I’m turning 40 in a month or so, and at 40 years young, I’m old enough to remember as far back as December 11 2025, when Disney and OpenAI “reached an agreement” to “bring beloved characters from across Disney’s brands to Sora.” As part of the deal, Disney would “become a major customer of OpenAI,” use its API “to build new products, tools and experiences (as well as showing Sora videos in Disney+),” and “deploy ChatGPT for its employees,” as well as making a $1 billion equity investment in OpenAI.
Just one small detail: none of this appears to have actually happened.
Despite an alleged $1 billion equity investment, neither Disney’s FY2025 annual report nor its February 2, 2026 Q1 FY2026 report mention OpenAI or any kind of equity investment. Disney+ does not show any Sora videos, and searching for “Sora” brings up “So Random,” a musical comedy sketch show from 2011 with a remarkably long Wikipedia page that spun off from another show called “Sonny With A Chance” after Demi Lovato went into rehab.
It doesn’t appear that investment ever happened, likely because — as was reported earlier this week by The Information and the Wall Street Journal — OpenAI is killing Sora. Shortly after the news was reported, The Hollywood Reporter confirmed that the deal with Disney was also dead.
Per The Journal, emphasis mine:
CEO Sam Altman announced the changes to staff on Tuesday, writing that the company would wind down products that use its video models. In addition to the consumer app, OpenAI is also discontinuing a version of Sora for developers and won’t support video functionality inside ChatGPT, either.
Oh, okay! The app that CNBC said was “challenging Hollywood” and “freaking out the movie industry” and The Hollywood Report would suggest could somehow challenge Pixar and was Sam Altman successfully “playing Hollywood” and that The Ankler said was OpenAI “going to war with Hollywood” as it “shook the industry” and that Deadline said made Hollywood “sore” and that Boardroom said was in a standoff with Hollywood and that the LA Times said was “deepening a battle between Hollywood and OpenAI” and “igniting a firestorm in Hollywood” and that Puck said had “Hollywood panicking” and TechnoLlama said was “the end of copyright as we know it” and that Slate said was a case of AI "crushing Hollywood as it we’ve known it” is completely dead a little more than five months after everybody claimed it was changing everything.
It’s almost as if everybody making these proclamations was instinctually printing whatever marketing copy had been imagined by the AI labs to promote compute-intensive vaporware, and absolutely nobody is going to apologize to the people working in the entertainment industry for scaring the fuck out of them with ghost stories! Every single person who blindly repeated that Sora existed and was changing everything should be forced to apologize to their readers!
I cannot express the sheer amount of panic that spread through every single part of the entertainment industry as a result of these specious, poorly-founded mythologies spread by people that didn’t give enough of a shit to understand what was actually going on. Sora 2 was always an act of desperation — an attempt to create a marketing cycle to prop up a tool that burned as much as $15 million a day that most of the mainstream media bought into because they believe everything OpenAI says and are willing to extrapolate the destruction of an entire industry from a fucking facade.
Thanks to everyone who participated in this grotesque scare-campaign, everybody I know in the film industry has been freaking out because every third headline about Sora 2 said that it would quickly replace actors and directors. The majority of coverage of Sora 2 acted as if we were mere minutes from it replacing all entertainment and all video-based social media, even though the videos themselves were only a few seconds long and looked like shit!
Sora 2 was never “challenging Hollywood” or “a threat to actors and directors,” it was a way to barf out videos that looked very much like Sora 2’s training data, and the reason you could only generate a few seconds at a time was these models started hallucinating stuff very quickly, because that’s what Large Language Models do.
Yet this is what the AI bubble is — poorly-substantiated media-driven hype cycles that exploit a total lack of awareness or willingness to scrutinize the powerful. Sora 2 was always a dog, it always looked like shit, it never challenged Hollywood, it never actually threatened the livelihoods of actors or directors or DPs or screenwriters outside of the tiny brains of studio executives that don’t watch or care about movies. Anybody that published a scary story about the power of Sora 2 helped needlessly spread panic through the performing arts, and should feel deep, unbridled shame.
You have genuinely harmed people I know and love, and need to wise up and do your fucking job.
I know, I know, you’re going to say you were “just reporting what was happening,” and that “OpenAI seemed unstoppable,” but none of that was ever true other than in your mind and the minds of venture capitalists and AI boosters. No, Sora 2 was never actually replacing anyone, that’s just not true, you made it up or had it made up for you.
But that, my friends, is the AI bubble. Five months can pass and an app can go from The End of Hollywood that apparently raised $1 billion to “discontinued via Twitter post that reads exactly like the collapse of a failed social network from 2013” and “didn’t actually raise anything.” It doesn’t matter if stuff actually exists, because it’ll be reported as if it does as long as a company says it’ll happen.
Perhaps I sound a little deranged, but isn’t anybody more concerned that a billion dollars that was meant to move from one company to another simply didn’t happen? Or, for that matter, that this keeps happening, again and again and again?
I’m serious! As I discussed in last year’s Enshittifinancial Crisis, OpenAI has had multiple deals that seem to be entirely fictional:
- Its supposed $100 billion investment (that was always a “letter of intent”) from NVIDIA that went from OpenAI allegedly buying billions of GPUs from NVIDIA in October 2025 to “only a commitment” in February 2026 in a mere four months.
- A “letter of intent” between SK Hynix and Samsung to supply 900,000 wafers of RAM a month that was reported as representing 40% of the global supply of DRAM that never resulted in anybody buying or selling any fucking RAM.
- A supposed “definitive agreement” with AMD from October 2025 that would involve OpenAI using AMD’s GPUs to power its “next-generation AI infrastructure,” except AMD didn’t change guidance and does not appear to have any revenue from OpenAI, despite the first gigawatt of data center capacity being due by the end of this year. Part of the deal also involved OpenAI being able to buy 10% of AMD’s stock, but that was so stupid I can’t even bring myself to write it up.
- When asked about this on its latest earnings call, AMD CEO Lisa Su said that “the ramp is on schedule to start in the second half of the year,” repeating the deal existed while not increasing guidance to account for a gigawatt of chips, which would work out to somewhere in the region of $20 billion to $30 billion of sales as its weak guidance of $9.8 billion in the next quarter, sending the stock tumbling as a result.
- Isn’t it also weird that Meta signed a near-identical deal on February 24 2026 and nobody seemed to notice that guidance wasn’t changing and AMD was apparently also going to install a gigawatt of GPUs with Meta by the end of 2026? Is everybody drunk? What’s going on?
- A “strategic collaboration” with Broadcom “...to deploy 10 gigawatts of openAI-designed AI accelerators” by the end of 2029 that has resulted in no sales of any kind and no increase in guidance to match, with no mentions of OpenAI in its latest quarterly earnings report.
- On its most-recent earnings call, Broadcom CEO Hock Tan added that it expected OpenAI to “deploy in volume their first-generation XPU in 2027 at over 1 gigawatt of capacity,” but did not raise guidance or, when asked directly, say how it would deploy 10GW by the end of 2029.
- I’ll also add that there isn’t a chance in hell OpenAI deploys a gigawatt of these chips in that timeframe, and Broadcom has yet to show any proof that these chips are going to be made.
- On its most-recent earnings call, Broadcom CEO Hock Tan added that it expected OpenAI to “deploy in volume their first-generation XPU in 2027 at over 1 gigawatt of capacity,” but did not raise guidance or, when asked directly, say how it would deploy 10GW by the end of 2029.
That’s just the AI bubble, baby! We don’t need actual stuff to happen! Just announce it and we’ll write it up! No problem, man! It doesn’t matter that one of the largest entertainment companies in the world simply didn’t give the most-notable startup in the world one billion dollars, much as it’s not a big deal that the entire media flew like Yogi Bear lured with a delicious pie toward every single talking point about OpenAI destroying Hollywood, much like it’s not a problem that Broadcom, AMD, SK Hynix, and Samsung all have misled their investors and the media about deals that range from threadbare to theoretical.
Except it is a problem, man! As I covered in this week’s free newsletter, I estimate that only around 3GW of actual IT load (so around 3.9GW of power) came online last year, and as Sightline reported, only 5GW of data center construction is actually in progress globally at this time, despite somewhere between 190GW and 240GW supposedly being in progress. In reality, data centers take forever to build (and obtaining the power even longer than that), but nobody needs to harsh their flow by looking into what’s actually happening.
In reality, the AI industry is pumped full of theoretical deals, obfuscations of revenues, promises that never lead anywhere, and mysterious hundreds of millions or billions of dollars that never seem to appear.
Beneath the surface, very little actual economic value is being created by AI, other than the single-most-annoying conversations in history pushed by people who will believe and repeat literally anything they are told by a startup or public company.
No, really. The two largest consumers of AI compute have made — at most, and I have serious questions about OpenAI — a combined $25 billion since the beginning of the AI bubble, and beneath them lies a labyrinth of different companies trying to use annualized revenues to obfuscate their meager cashflow and brutal burn-rate.
To make matters worse, almost every single data center announcement you’ve read for the last four years is effectively theoretical, their nigh-on-conceptual “AI buildouts” laundered through major media outlets to give the appearance of activity where little actually exists.
The AI industry is grifting the finance and media industry, exploiting a global intelligence crisis where the people with some of the largest audiences and pocketbooks have fundamentally disconnected themselves from reality.
I don’t like being misled, and I don’t like seeing others get rich doing so.
It’s time to get to the bottom of this.