{
  "id": 17762543,
  "title": "I asked ChatGPT, Gemini, Claude and Perplexity if there really is an AI bubble and when it’s going to burst — and one compared it to the dot-com crash",
  "url": "https://urgent.news/2026/11/02/i-asked-chatgpt-gemini-claude-and-perplexity-if-there-really-is-an-ai",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-02T06:30:00.000Z",
  "source": {
    "name": "Tom's Guide",
    "slug": "tom-s-guide",
    "url": "https://www.tomsguide.com/ai/i-asked-chatgpt-gemini-claude-and-perplexity-if-there-really-is-an-ai-bubble-and-when-its-going-to-burst-and-one-compared-it-to-the-dot-com-crash"
  },
  "original_language": "en",
  "account": "In the ongoing \"RAMageddon,\" the rapid development of data centers and the negative perception of \"AI slop\" have led skeptics to hope for an AI industry market crash. After seeing a news story about China's AI sector suffering a massive market crash that wiped out $480 billion, the author began to ponder the possibility of a similar occurrence in the United States. Fueled by curiosity, the author asked several artificial intelligence chatbots about the existence of an AI bubble and what might happen if it were to burst.\n\nChatGPT, an OpenAI chatbot, suggested that there are signs of an AI bubble, but it is not guaranteed that a crash is imminent. The chatbot proposed separating two questions: whether AI is valuable and whether investors are paying too much for AI. ChatGPT likened the AI industry's situation to the dot-com boom of the late 1990s, stating that if the revenue for the AI industry fails to keep pace with infrastructure spending from private sectors, a serious reckoning could happen between 2027 and 2029. ChatGPT concluded that the central question is whether the money generated by AI can justify the money being invested in it.\n\nGemini, a Google AI, highlighted three similarities between the AI bubble and the dot-com crash: 1) Infrastructure overbuild, where GPU clusters and data centers may sit underutilized if software adoption lags, similar to telecom companies laying fiber-optic cables in the 1990s; 2) Circular revenue feedback loops, where modern AI startups spend funding on cloud credits and GPU compute leased back from their own venture sponsors, akin to dot-com companies purchasing advertising and server hardware; and 3) Narrative overreach, as both booms were based on the perception that a technological revolution was underway, even though monetization in enterprise workflows may take longer than anticipated. Gemini also noted that unlike the dot-com crash, a few tech leaders with diverse investments may not experience catastrophic losses if the AI bubble bursts.\n\nClaude, an Anthropic chatbot, explained why people are worried about the possibility of an AI bubble bursting. Spending on AI is enormous, with companies like Meta, Microsoft, Alphabet, and Amazon expected to invest at least $700 billion in 2026, which represents roughly 2.2% of GDP. Additionally, AI investment accounted for most US GDP growth in early 2026, meaning that a large portion of the economy now depends on this single bet. Prominent figures, such as Ray Dalio, have expressed concerns over the situation. While Claude acknowledged these worries, he also suggested that the timeline for the AI bubble bursting, if it were to occur, remains uncertain.",
  "summary": "After asking four chatbots about the hot topic of the existence of an AI bubble and when it could burst, I was met with some intriguing responses.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 2,
    "also_reported_by": [
      {
        "outlet": "Dev.to",
        "title": "7 real cases where ChatGPT and Claude got datasheet values wrong (and how to check them)",
        "url": "https://urgent.news/2026/11/01/7-real-cases-where-chatgpt-and-claude-got-datasheet-values-wrong-and",
        "published": "2026-11-01T23:18:25.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}