The Bubble Bursts Credit:insider.com
According to the 2025 Stanford AI Index Report which has just been published, The 2025 AI Index Report | Stanford HAI, the report does not forecast an AI-bubble which might burst soon. Published by the university’s Institute for Human-Centred AI (HAI), the report indicates that AI is poised to be the most transformative technology of the 21st century. However, its benefits will not be evenly distributed unless its development is guided thoughtfully. The Index offers one of the most comprehensive data-driven views of artificial intelligence (AI). Recognized as a trusted resource by global media, governments and leading companies, the index report provides policymakers, business leaders and the public with rigorous, objective insights into AI’s technical progress, economic influence and societal impact.
The Report’s Key Messages
Some of the key messages can be summarized as follows:
AI-Performance: The report finds that the performance of AI systems has improved rapidly in the past year. “There are very few task categories where human ability surpasses AI,” the authors write, and “the performance gap between AI and humans is shrinking rapidly.” But there are debates as to whether the benchmarks being used are actually the best way to measure human intellect.
AI-Access: AI has become more accessible in the past year as small models have become more capable and Large Language Models (LLMs) cheaper to run. The cost of running a GPT 3.5-level model has dropped 280-fold since 2022, while hardware costs have dropped, and energy efficiency has improved.
More Scientific Discovery: Some members of the AI community question why AI is not yielding more new scientific discoveries. To counter this argument the authors make the point that AI researchers have started to engage more in science, with Stanford adding a ‘Science and Medicine’ chapter. Moreover, this year’s report also highlights advancements in biology, materials science and fire monitoring via satellite.
The Myth of AGI
The idea of creating an artificial mind that can rival or exceed human intelligence persisted for a long time. Some of the earliest examples can be found in ancient myths and legends. In the last few decades, the concept of Artificial General Intelligence (AGI) has been promoted by thought-leaders and writers like Alan Turing, John von Neumann, Isaac Asimov, Ray Kurzweil, Nick Bostrom and many others. AI has contributed heavily to solving specific problems, but we still are far away from AGI, which is considered the ‘Holy Grail of AI’. Defining AGI is not easy, but there are several characteristics that AGI must provide, such as common sense, transfer learning, abstraction and causality. In recent years, deep learning models have heavily contributed to advance computer vision, speech recognition and natural language processing. However, these applications, as useful as they might be, still lack the fundamental human capacity to include meaning and causality for solving a problem. A huge language model might be able to generate a coherent text or translate a paragraph from French to English, but it does not understand the meaning of the words and sentences it creates. What it is basically doing is predicting the next word in a sequence based on statistics it has derived from millions of text documents. There is no consensus among researchers on whether AGI is possible or desirable and when it might be achieved or how it might behave. There are many open questions that need to be addressed before AGI can become a reality. Some of these are: How do we define and measure intelligence? Is there a universal standard or metric for comparing different forms and levels of intelligence? How do we ensure that intelligent systems share our goals and values? How do we prevent unwanted or harmful behaviour? How do we monitor and control the actions and outcomes?
These questions are not only technical but also ethical, social and philosophical. They require interdisciplinary collaboration and public engagement to find satisfactory answers. They also require constant reflection and revision to adapt to the rapidly changing environment of AI-technology.
First Signs of an AI-Bubble
Individuals not willing to pay the cost of an AI model to answer a specific question signal the first warning sign of an AI-bubble being formed. Hence, the more users an AI platform has, the greater it costs for the company running it. According to last year’s financial report, OpenAI’s users cost the company USD 56 billion. To cover that with subscriptions, each of an estimated 96 million users would have to pay USD 583.- a year. Compared to last year that would mean a tripling of the cost to use their AI model. These prices would allow OpenAI to break even with no profit for the company and no return on the billions already invested by third-parties, including banks, venture capital firms, governments and large corporations. This raises the question as to what investors expect to achieve, after having spent USD 110 billion on AI technology in 2024 alone. Large models are based on learning, but the content presented by the internet is finite and is already used. Adding AI-generated content to the internet reduces the usefulness of AIs and makes them more prone to hallucination. As a result the costs of trying to achieve the impossible implies that the AI bubble will inevitably burst in a downfall of its own mediocrity. AI has several real-world use-cases where it has proved effective, and the technology will likely continue to develop in those niches. But the direction mainstream AI is following implies that throwing billions at an inherently flawed concept limited by the laws of physics, will end in a massive wake-up call with investors painfully realizing that the usefulness of AI has been totally overhyped.
Consequences of an AI-Bubble Burst
According to an article published by Business Insider in late 2024, The AI-Fuelled Stock Market Bubble Will Crash in 2026: Capital Economics – Markets Insider , the research company Capital Economics predicts that an AI-fuelled Stock Market bubble will burst in 2026. Driven by investor excitement the S&P 500, led by technology stocks, could be driven to a record high by the end of 2025. But starting in 2026, those stock market gains should unwind precipitously as higher interest rates, and an elevated inflation rate start to weigh down equity valuations. “Ultimately, we anticipate that returns from equities over the next decade will be poorer than over the previous one while the long-running outperformance of the US stock market may come to an end,” Reilly of Capital Economics states. Their bearish stock market call is somewhat counter-intuitive, as researchers expect that the growing adoption of AI will spark a boost in economic growth driven by increases in productivity. That economic boost could result in higher inflation than most expect and as a result increase interest rates. Higher interest rates and inflation are ultimately bad news for stock prices. “We suspect that the bubble will ultimately burst by the end of next year, causing a correction in valuations. After all, this dynamic played out during the dot-com bubble of the late 1990s and early 2000s as well as the Great Crash of 1929,” Reilly said. The expected burst of the stock market bubble should lead to a decade of investment returns that favour bonds over stocks. Capital Economics forecasts that between now and the end of 2033, US stocks will deliver average annual returns of just 4.3%, which is well below the long-term average return of about 7% after inflation. “When and how the AI-fuelled equity bubble bursts is a key risk to our forecast. In particular, one downside risk is that the aftermath of the bursting of the bubble lasts longer than one year as was the case following the burst of the dot com bubble,” Reilly said.
Conclusion
With the Trump Administration taking control – apart from many other controversial issues – their confusing global trade war, leveraging taxes on goods and services, has reached a new level of intensity and concern among many reputable economists. Many recognized decision makers on economic issues including Jerome Powell, head of the Federal Reserve, are warning that these actions will not strengthen but rather weaken the US as a global leader. Within the next 6 months we might get an idea of the long-range impact on the value of US technology stocks. Chances are that the AI-bubble is likely to burst already this year.