The video highlights the growing AI hype, focusing on OpenAI’s costly and error-prone video generation technology, unsustainable financial losses, and a circular investment relationship with Nvidia that inflates valuations, signaling an AI investment bubble. Despite AI’s revolutionary potential, the current market exuberance and technological limitations suggest a looming correction similar to past tech bubbles.
In this video, Sasha discusses the recent hype surrounding AI, particularly focusing on OpenAI’s release of Sora 2, a video generation tool. Despite the excitement on social media, the technology still produces many unusable clips with significant errors, and there is no current way to edit generated videos easily. The quality of videos tends to be better when the AI can rely on pre-existing similar videos in its training data, raising concerns about copyright infringement. This has sparked debates about the legality and ethics of using copyrighted content without compensation, with lawsuits already underway but progressing slowly, partly due to political lobbying and financial interests.
Sasha then delves into OpenAI’s financials, revealing that while the company generated $4.3 billion in revenue in the first half of 2025—a 130% increase from 2024—it is still losing massive amounts of money. OpenAI spent $6.7 billion on research and development, $2.5 billion on inference, and $2 billion on sales and marketing, resulting in a net loss of nearly $7 billion before accounting for capital expenditures. The cost of running AI models is rising sharply because newer models require exponentially more compute power, and economies of scale are not reducing these costs. This financial imbalance highlights the unsustainable nature of current AI business models.
A significant part of the video focuses on the relationship between OpenAI and Nvidia. Nvidia recently announced a $100 billion investment in OpenAI to build massive AI data centers requiring millions of GPUs, nearly double Nvidia’s annual production. Sasha explains this as a circular financial game where OpenAI raises funds, which it then spends on Nvidia’s GPUs, and Nvidia, in turn, invests back into OpenAI. This arrangement inflates OpenAI’s valuation and benefits Nvidia immensely, as it profits both from selling GPUs and from the hype-driven increase in OpenAI’s market value. This cycle exemplifies the speculative nature of the AI investment bubble.
Sasha also shares a provocative prediction from Jeff Bezos about building gigawatt-scale data centers in space to leverage constant solar power, a notion Sasha finds impractical and indicative of peak hype. This, combined with the enthusiastic reactions from AI fans and tech enthusiasts, signals to Sasha that the AI industry is entering a bubble phase. She draws parallels to past tech bubbles, emphasizing that while AI technology is revolutionary and will bring long-term benefits, the current valuations and hype far exceed the actual technological and economic realities.
Finally, Sasha reflects on the broader implications of AI, comparing it to the evolution of chess AI, which, despite being highly advanced, is not widely used for human play because people prefer the unpredictability of human opponents. She acknowledges AI’s potential to boost productivity in fields like software development and CGI but cautions that the market’s current exuberance is reminiscent of previous bubbles. With the stock market reaching all-time highs and valuation metrics at historic peaks, Sasha warns investors to be cautious and stay tuned for future updates as the AI hype unfolds and potentially corrects.