The AI Bubble Signal: A Sovereign Wealth Fund's Warning and the Crypto Parallel

The CEO of the world's largest sovereign wealth fund has just done what no analyst dared to say aloud: he quantified the AI bubble. Not in vague terms, but with a number. Thirty-five percent. That is the value he expects to be erased from the fund's portfolio if the AI exuberance corrects. The fund manages $1.6 trillion of Norway's pension savings. This is not a hedge fund manager selling fear. This is a steward of public wealth, issuing a public warning. Trust no one. Verify everything. The signal is not the headline. The signal is the specificity. When a sovereign fund CEO speaks in percentages, they have already run the numbers. They have stress-tested their holdings. They are telling us the downside case is real. And they are preparing the public for the loss. Context: The fund in question is Norges Bank Investment Management, helmed by Nicolai Tangen. Its portfolio is heavily weighted toward U.S. tech giants. Microsoft, NVIDIA, Alphabet, Amazon—these are the pillars of the AI infrastructure trade. Tangen's warning is not about AI's technical failure. It is about the gap between market pricing and fundamental reality. The fund's internal models show that current AI-related valuations are priced for a future that may not arrive on schedule. The 35% figure likely assumes a reversion of tech multiples to historical means, combined with earnings disappointments from over-investment in capacity. This is not a crypto-specific story. But it is a story every crypto builder must internalize. Because the same pattern that inflated AI valuations inflated crypto values in 2021, and then again in 2024's meme-coin frenzy. The same mechanism: narrative-driven capital flows, disconnected from sustainable revenue, propped up by cheap money and FOMO. The same outcome: a correction that wipes out the weakest projects and leaves only those with real utility standing. Core: I have seen this movie before. In 2017, I audited fifteen ICO whitepapers for a financial engineering consultancy. I found that over 80% of them had no viable tokenomics—just a supply schedule and a dream. The market rewarded them anyway. Then the music stopped. The same pattern repeated in DeFi Summer 2020. I watched governance tokens soar on the promise of decentralized decision-making, only to discover that whale capture was inevitable without proper incentive alignment. Based on my experience modeling MakerDAO's governance simulation, I learned that hype can sustain a system for exactly as long as new capital enters. Once the inflow stops, the valuation collapses to the net present value of real cash flows. AI today is structurally identical. The capital expenditure by the largest tech firms on AI infrastructure has grown at over 50% per year for the past three years. Revenue growth from AI products? Roughly 10-15% for the same period. The gap is being filled by narrative. Investors are buying the story that AI will eventually transform everything, so they tolerate current losses. But the tolerance has a limit. The sovereign wealth fund's CEO is that limit made audible. Let me be specific. The 35% loss likely corresponds to a scenario where NVIDIA's stock corrects 40-50% from its peak, dragging down the entire tech sector. Why NVIDIA? Because it is the single largest beneficiary of AI capital expenditure. Its data center revenue has grown from $15 billion in fiscal 2023 to over $90 billion in fiscal 2025. That growth is priced in. If AI spending slows by even 20%, the earnings multiple will compress. The result: a 35% fund-level loss is plausible. But here is the nuance that the mainstream analysis misses. The bubble burst is not necessarily bad for the technology. In fact, it may be necessary. Just as the 2000 dot-com crash cleared the path for Amazon and Google to dominate, an AI correction will separate the narrative plays from the actual value creators. The same applies to crypto. When AI capital retreats, some of that liquidity will rotate into other risk assets. But more importantly, the collapse will force all speculative markets to reprice based on fundamentals. Crypto projects that have real revenue, real users, and real unit economics will survive. Those that rely on hype alone will not. Contrarian: The counterintuitive angle is that the AI bubble's burst could be a net positive for the crypto industry. Here is why. First, AI and crypto are competing for the same speculative capital pool. When AI loses its luster, crypto becomes the next narrative play. But that is a short-term, shallow view. The deeper impact is structural: AI commoditization. The correction will slash GPU prices, reduce cloud computing costs, and make AI inference cheap. For crypto projects that integrate AI—like decentralized training, smart contract auditing with LLMs, or autonomous agents—lower costs mean better margins. The bubble burst removes the artificial scarcity of compute and replaces it with abundance. Code becomes light. Gold becomes heavy. Second, the sovereign wealth fund's warning is a direct repudiation of the "infinite growth" thesis that both AI and crypto have been selling. Investors who believed that AI would grow exponentially forever are now forced to consider logistic curves. The same must happen in crypto. The days of "everything will go up because of adoption" are over. We must ask: which protocols have a clear path to sustainable revenue? Which DAOs have real governance participation? Which tokens actually capture value from the activity they enable? I recall organizing "Soulbound Berlin" in 2021. I tried to build a community around non-transferable tokens as identity markers. The experiment failed because 90% of participants sold their tokens at the first opportunity. The lesson: even a well-designed protocol cannot resist the gravitational pull of speculation. The AI bubble is the same. The technology is real. The benefits are real. But the market has priced in a fantasy. The correction will be painful, but it will also be cleansing. Takeaway: Noise is cheap. Signal is rare. The sovereign wealth fund CEO has given us a signal. He is not predicting the end of AI. He is predicting the end of overpriced AI. The same will happen in crypto. Builders who focus on delivering value to real users, who optimize for retention over hype, who measure success by revenue not TVL, will survive and thrive. Summer fades. Builders remain. The question is not whether the bubble will burst. It is whether you are building something that will still matter when it does. I will leave you with this: the next time you see a project with a billion-dollar valuation and no product, remember the words of a man who manages $1.6 trillion. He has already run the math. Trust no one. Verify everything.