Hi. Welcome to another episode of Cold Fusion, a Cambrian explosion of creativity. That's the exact phrase Sam Altman used a few months ago when describing Sora. To him, it was an AI video tool so powerful, so unprecedented that it was going to fundamentally change how humans express themselves. Artists, filmmakers, everyday people, all of them apparently on the brink of a new renaissance. Then, almost as suddenly as Sora burst onto the scene, it was gone. The app, the API, all of it scrubbed from the face of the earth. Only a few months ago, Disney had signed a billion-dollar deal for Sora, but apparently, they were blindsided by the decision, too. According to Reuters, the Disney team was actively working with

OpenAI on a Sora-related project on a Monday evening. 30 minutes after that meeting ended, they got the call that Sora was being discontinued. A person familiar with the matter called it, and I'm paraphrasing here, a big rug pull. Disney has since pulled its investment so there's one major question that people have on their lips. Why would OpenAI kill something this big, this fast? As many of you saw in the episode, OpenAI is suddenly in trouble. Financial pressures, scandals, and competition are really eroding confidence in the company's future. But the closure of Sora is more than about just saving money. It indicates a sharp change in strategy. OpenAI is trying to ensure its

own survival, but that's not all. Later in this episode, we'll see how OpenAI somehow raised another 122 billion dollars, but strangely, their very first purchase was a podcast. They paid reportedly hundreds of millions of dollars for it, a decision that leaves everyone scratching their heads. And beyond this, more top executives are either being shuffled around or leaving for health reasons. Among the absolute chaos, OpenAI is planning a pivot into robotics, but Figure Robotics, who previously partnered with them, gives them an F grade. The Figure Robotics CEO called the OpenAI team virtually worthless. Pretty close relationship with OpenAI, correct? Yeah. I end up firing them.

And we had we just had a hard time getting them in the office. We just found that like that team we had internally, we just ran like kind of circles around them like every day. Then, as I was finishing up this episode, The New Yorker dropped a bombshell. It shone a light on what insiders of OpenAI really think of Sam Altman. I'll give you a hint, it's not good. So, there's a lot to get through in this episode. So, let's get straight into it and break it all down, beginning with Sora. You are watching Cold Fusion TV. And now to what's being touted as another giant leap in artificial intelligence, Sora is the latest tool released by the company behind the groundbreaking chatbot ChatGPT. It can create realistic-looking videos based on

a simple text prompt. OpenAI first revealed Sora in early 2024, and the reception was strong. I covered its release back then. Although Sora is nothing compared to the quality of modern video models like Seed Dance 2, visually, the demos of Sora were genuinely impressive for that time period 2 years ago. The launch of Sora sent tech companies across the US and China scrambling to expedite their own video generation models. Then, in October of 2025, OpenAI launched Sora as a standalone app. And with that, the growth followed. A million downloads in just a matter of days. It hit the top of the App Store almost immediately.

Essentially a TikTok clone, Sora let users generate and share AI videos and even create AI versions featuring themselves or their friends. But it quickly became clear that giving everyone a free and easy AI video generation tool was more akin to giving a monkey a semi-automatic weapon. Not a good time. According to reporting by The Atlantic, one journalist opened up the app one morning and scrolled through fake scenes of the Iran war, AI-generated Trump clips, and in what she described as her least favorite, a video of a man deep-frying an infant. Like a virus, AI content from Sora found its way onto other social media platforms, polluting them with AI slop to an incredible degree.

Despite an abundance of copyrighted cartoon characters in unfamily-friendly situations, somehow Disney seemed to enjoy it. Instead of suing OpenAI, in a bizarre deal, they agreed to shove over a $1 billion investment in OpenAI as part of a 3-year licensing agreement. Why Disney would pay OpenAI to use their characters is beyond me, but regardless, the end result was over 200 iconic characters available for Sora users to generate AI videos with. Disney was also set to use OpenAI's APIs across its platforms, including Disney Plus, and to deploy ChatGPT internally for employees. It kind of seemed like this was the next big shift in tech. Then 103 days later, Sora was suddenly dead. The announcement

was unceremoniously made publicly on a Tuesday. The Sora team posted on X, quote, "We're saying goodbye to the Sora app. We know this news is disappointing." End quote. Looking at the overall reaction, though, not sure I'd call it disappointing. Apparently, even some Sora staffers didn't see it coming. The announcement was made just 1 day after OpenAI published a blog post about Sora safety standards, which by the way was still up at the time of making this episode. But this all begs the most important question. Why did they do it? And why now? Later in the episode, we'll talk about the rest of the wild things going on behind the scenes at OpenAI. But for now, let's really focus in on Sora

because there's more to the story the deeper you look into it. OpenAI is planning to shut down its popular video generation app Sora. Sora is getting shut down. Yep, I saw the good news. For many, killing a big project like Sora, as well as some other side projects, means only one thing. OpenAI is under pressure. But there's more to understand about the situation the further you look into it. Here are five likely reasons Sora got the boot. Number one, the compute problem, and it's a serious one. Generating video is extraordinarily expensive compared to generating text or even images.

According to a Forbes estimate, Sora may have been costing OpenAI up to $15 million a day to run. At that rate, Disney's entire $1 billion investment would have been burnt through in about 2 months. Bill Pebbles, who leads the Sora team, reportedly acknowledged that the economics were, in his words, completely unsustainable. And this wasn't happening in a vacuum. Reports suggest that Sora was chewing up valuable compute resources that could have been used for other teams within OpenAI. OpenAI is burning through cash at an alarming rate. The company is on track to lose $14 billion in 2026 alone, even as revenue has crossed $20 billion annually. You can't sustain a $15

million a day loss just for an AI slop side project. With those figures of millions of dollars a day in losses in mind, guess how much revenue Sora made throughout its lifetime. According to TechCrunch, just a measly $2.1 million. Number two, weak monetization. Viral doesn't pay the bills. Sora was popular at the start, but making money turned out to be harder than expected. AI slop videos do go viral on social media, but getting users or advertisers to pay for it is another story. It's just not a solid revenue model. And even worse, interest in Sora was falling off a cliff. Comparing Sora to OpenAI's coding products, the latter has a clear value proposition. Businesses pay for tools that hypothetically make the developers

faster and more efficient, even though studies nowadays say otherwise. But still, the demand is concrete. But Sora seemed more like throwing a dart at a dartboard strategy just to see what sticks. As The Atlantic put it, OpenAI seems to be finally learning that slop is not a business strategy. Number three, a major strategic shift is underway. And here's where it gets really interesting because Sora being shut down isn't an isolated event. Fidji Simo, OpenAI's applications chief, reportedly told staff in a company-wide meeting, quote, "We cannot miss this moment because we're distracted by side quests." She went on to say that the company needs to nail, quote,

"productivity on the business front." End quote. OpenAI was reportedly spreading itself thin across too many products, too many apps, with the quality suffering across the board as a result. And that's something I talked about in my last episode about OpenAI. But the business model that they're now copying comes from, drumroll, please, Anthropic. Anthropic's Claude code has gained serious traction among developers. Its enterprise market share grew from 18% to 29% in 2025. Eight of the Fortune 10 are now customers. OpenAI, which has historically had the stronger brand, is now playing catch-up to Anthropic. OpenAI taking the Department of War deal that Anthropic rejected on moral grounds only adds to

the brand image problem. The company has even renamed its product organization to AGI development. According to some accounts, the Sora team has now been reassigned to world model research with a focus on robotics. And that's clearly where OpenAI thinks the next frontier is. Now, you may remember a few years ago, I did a couple of episodes on Figure Robotics. Back then, I talked about how Figure had partnered with OpenAI. Now, Figure CEO Brett Adcock has come out saying just how poor the relationship was. He stated in a podcast that apart from brand recognition close to a funding round, the collaboration with OpenAI was useless. Uh Sam and OpenAI led my Series B. And uh they were they joined my board, and then we ended up spending

basically a year with them working on um the goal was like try to advance AI models for human and robots together. And at the time we had like a we know we still do but we had our we had a full like AI team internally so we were basically working weekly kind of like language models for robotics. Yeah, I ended up firing them. We just found that like that team we had internally we just we just found like kind of circles around them. Like every day and we had a we just had a hard time getting them in the office. We had a hard time like basically like basically you know advancing stuff together as a team. It also we just like got no value out of the whole

relationship for a very little. I mean this it was helpful having them lead the route clearly the round. It was like there was some there was like some good brand association there but like beyond like that there wasn't much. In fact, he insinuates that they were on their way to ripping off proprietary technology. We were like also week to week like showing them how we were doing all this work. And I got a call one day saying like hey we're like you know we've been watching your progress is unbelievable. And um you know we're thinking about doing robotics work internally. And I was just like this is over.

Yeah, I just get out of here. Like this is like we're we're like teaching you how to do like robot learning. You're seeing our progress. We had like a couple of the you know Sam and couple of co-founders on site at one point where they were still at a point where they continued want to work together after this and I was like there's no way we're going to teach you how to do this stuff anymore. Open AI is now a company doing a hard pivot before an IPO. Now I cover a lot of technology stories on this channel and one thing I see time and time again is how vulnerable our data is online.

Well in order to get that information about you they get that data from data brokers. These companies store your personal information that exists online not just on social media but on sites you've probably never heard of. Stories and events like that are exactly the reason why I started using DeleteMe to protect my online presence. There's an entire industry of data brokers collecting and selling personal information. Things like your name address phone number even details about your relatives often to anyone willing to pay. And apart from surveillance this can lead to scams fishing attempts or just a general loss of privacy. DeleteMe is a hands-free subscription service that removes your personal information from

hundreds of data brokers and people search sites and importantly it keeps checking and removing it over time as it reappears. What stood out to me about it was how simple it is to use. You basically submit your details once and within about a week you start seeing privacy reports showing where your data was found and what's been removed. So it's not a one-time cleanup but ongoing protection. They've been doing this for over 15 years. Everything is handled in house and if you need something specific removed there are real privacy advisors that you can reach out to. And then it just makes sense to and then it just makes sense to have something quietly working in the background on your behalf. If you want

to take control of your personal data you can get 20% off DeleteMe customer plans by going to joindeleteme.com/codefusion and use the code codefusion at checkout. That's joindeleteme.com/codefusion. Links in the description and you can also use the QR code on screen. Thank you to DeleteMe for supporting ColdFusion. Now back to the episode. Number four, the legal and IP risk was becoming untenable. Let's get this straight. Sora was a copyright lawyer's nightmare. From the beginning questions circulated about the training data Open AI used to build the model.

Sora could generate copyright infringing material at scale. It was a huge liability. Throw in deep fakes and illegal material on a platform that was actively incentivizing users to generate as much content as possible and Sora was a content moderation nightmare. Number five, the competition pressure. Google's Veo platform now supports 4K output and spatial audio. Adobe's Firefly is integrated directly into Premiere Pro and is trained on clean licensed imagery. There's also Runway and Kling. And then there's the king and the newest player on the block SeaDance 2.0. Looking at all of this the world of AI video is suddenly getting crowded. Whatever edge Sora had last year is all but gone. In the end all Open AI had left was an

expensive legally complicated slot machine that made content moderation almost impossible. At that point every dollar and every GPU hour spent on Sora was a dollar and GPU hour not spent on the stuff that actually matters for their future. In early April 2026 Open AI announced a $122 billion funding round. That sounds good on paper until you realize that the vast majority of the money raised are from the very same stakeholders that need for-profit AI to succeed. Amazon invested up to 50 billion Nvidia 30 billion and SoftBank another 30 billion. Amazon needs AI demand to justify its AWS infrastructure buildout.

Nvidia needs Open AI compute demand to sell their GPUs and the notorious SoftBank has bet big on the AI story and they can't afford to lose. This massive investment again boils down to yet more circular financing. Outside investors are less keen. Bloomberg reports that private Open AI shares are being sold off on the secondary market and sometimes there's no buyers for these shares. Investors are much more interested in Anthropic. Next round capital founder Ken Schmit tells Bloomberg quote we literally couldn't find anyone in our pool of 100 institutional investors to take these Open AI shares. Buyers have indicated they have 2 billion of cash ready to deploy into Anthropic and quote. The

bottom line is unless there's a drastic change it's going to be a hard road ahead for Open AI financially. As the Guardian puts it in relation to the latest funding round quote the startup Open AI is nowhere near being profitable. In fact, if things stay as they are it will burn through half a trillion dollars by the end of the decade and quote. In order to have a chance to even make it Open AI absolutely has to cut the fat. And Sora was a big part of that waste. Is what I would say if things didn't keep moving so quickly. As this episode was being made a new development occurred. After so much talk of cutting the fat and getting their house in order Open AI's first purchase after the

funding round was a podcast. Although the exact dollar figure of the purchase isn't known reportedly the range is in the low hundreds of millions of dollars. Axios interviewed Sam Altman and he seemed very coy about revealing the price he paid for the podcast. The Financial Times says you paid in the low hundreds of millions is that about right? I won't comment on that actually. Um that seem high seem low? I think that was a complete sentence. The podcast is called TBP N. You're probably asking now who's that? In short it's a daily live show about tech and business that launched in late

2024. They regularly feature some of the biggest names in the technology industry. Think Mark Zuckerberg Satya Nadella Mark Cuban and so on. So their audience likely comprises of individuals who are founders CEOs and business owners. Essentially people with money who are well established in the tech world. But with 70,000 average daily viewers according to some estimates and 324,000 subscribers on X and only 70,000 on YouTube the sheer size of the acquisition does raise an eyebrow. Why would Sam Altman reportedly pay hundreds of millions of dollars for a podcast that's seemingly so small? I had to find out so I did some more digging. As it turns out in December of 2025 the New York Stock Exchange announced a

formal partnership with TBP N. Apparently a lot of investors watch the show. That's very interesting seeing that Open AI is trying for an IPO later this year. Must be a coincidence. So what else? I decided to watch the episode of their podcast where they announced the acquisition by Open AI. Besides the ecstatic tone co-founder John Curgan mentioned that he'd known Sam Altman for 13 years. We have some huge news. This is from the Open AI blog. Open AI acquires TBP N accelerating the global conversation about AI. This is real. This is a very interesting deal. I think a lot of people will be interested in this. We're very excited about this.

I've known Sam for maybe 13 years. He invested in my first company in 2013. Once we actually started growing TBP N he was one of the first people that I texted to you know say hey do you want to come on the show and he was the first lab lead to come on the show and we're The bottom line here is that Open AI is buying a media company connected to the New York Stock Exchange to more effectively market AI to the public. That last part isn't my words. That's what Sam Altman said in the same Axios interview. They are genius marketers and someone said to me recently that if AI were a political candidate it would be the least popular political candidate in history and given the amazing things AI can do I think there's got to be better marketing for

AI. So with this podcast purchase I'm not exactly sure what's going on here but I'll leave you to make up your own mind. In other things behind the scenes Open AI CFO had doubts about the 600 billion in spending commitments. She also doesn't think the company is ready for an IPO this year. She admits that Open AI didn't have enough compute and she spent her time trying to quote find any last minute compute available. This could be because the data center buildout is going terribly wrong with buildouts either being scaled back or canceled. For her efforts Sam Altman reportedly kicked her out of attending any financial meetings with investors.

The investors said that her absence was quote noticeable and awkward and quote. Then on the 6th of April the massive bombshell from the New Yorker dropped. It was an exposé of private emails notes and interviews indicating that Sam Altman was ultimately untrustworthy. Some inside sources went as far as to say that the main problem with the company was Sam. Okay, there's new tech drama. So Ronan Farrow the investigative journalist that broke Harvey Weinstein just dropped an 18-month long investigation into Sam Altman. His own board member called him a sociopath. There are internal memos

saying that he's a consistent pattern of lying. His own mentor, Paul Graham, the founder of Y Combinator, the guy that made Altman the president of YC, told colleagues that Sam has been lying to them the entire time. So, yeah, that's the guy that runs the most powerful AI company in the world. A sociopath with a pattern of lying. Now, let's get back to Sora and the AI video generation space. Even though Nvidia has promised more money, CEO Jensen Huang reportedly raised concerns about, let's say, OpenAI's lack of discipline in its business approach. Here's the nuance, though. This is not the end of AI video as a technology. When and if compute

costs come down, other companies will step in. In fact, The Hollywood Reporter thinks that Google will win IP contracts in the end because of its scale. And, of course, Elon Musk has already said that xAI will double down on generative video in Sora's wake. They'll be upgrading Grok's video capabilities. Make of that what you will, given Grok's complicated track record in the visual space. But, we have to stop and ask ourselves, is Big Tech still misreading the room? There's a real cultural reckoning happening that goes way beyond OpenAI. AI-generated content is everywhere now. The copyright frameworks haven't caught up. The creative industries are pulling their hair out. And deals like the Disney one are going to be scrutinized for years.

For OpenAI, though, the Sora shutdown, the AGI development rebrand, the push into enterprise, Sam Altman's stated focus on, quote, "raising capital, supply chains, and building data centers at unprecedented scale." It all points to one direction. OpenAI is positioning itself for an IPO, possibly later this year. And for that, they need a story that goes beyond, "We let you make AI slop videos." Sora simply had to go. To be fair here, Sam Altman did say in his blog, quote, "The majority of users should feel that their life is better for using Sora than would have been if they hadn't. If that's not the case, we will make significant changes. And if we can't fix it, we would discontinue offering the service." End

quote. In summary, Sora failed for a variety of reasons, but most of the reason was just a bad strategic decision. It was the wrong product at the wrong time for a company struggling financially. OpenAI is trying to cut the bloat and get their act together. It's the same situation with their adult chatbot, and perhaps a few other projects are on the chopping block, too. OpenAI built the most famous AI brand in the world. They have 900 million weekly users. They've raised more money than most countries' GDP and burned through cash like no tomorrow. And yet, they're one of the least trusted companies and also have an identity crisis. The AI industry in 2026 is going through a genuine identity crisis.

There's a question of what is AI for? How does it make money? Who is responsible when things go wrong? All of these questions are still very much open. So, as always, what do you guys think? Will OpenAI be able to overcome the majority of their challenges? Or, will the company continue to deteriorate? With an IPO coming up, the clock is surely ticking. Anyway, that's about it for me. My name is Dagogo, and you're watching Cold Fusion, and I'll catch you again soon for the next episode. Cheers, guys. Have a good one.

Cold Fusion. It's new thinking.