The Bigger They Are, the Harder AI's Losers Could Fall
The buildout of the internet was comparatively dirt cheap to what is taking place during the AI infrastructure boom across the nation.
You've reached your free article limit
You've read 0 of 1 free Pro articles.
There were some big moves in the market last week. Friday, gold dropped nearly 9% in trading, coinciding with the announcement of the new nominee to chair the Fed. Silver prices plunged by more than 25% on the day. It was the biggest daily move in "poor man’s gold" since the Hunt brothers failed to corner the silver market more than 45 years ago. Microsoft (MSFT) lost over $350 billion of its market capitalization in just one day following its quarterly results and updated guidance. To put this in perspective, Microsoft was the most valuable stock in the market during the Internet Boom, topping out at just under $600 billion market capitalization just before the Internet bust.
The market remains resilient given some of the sharp moves that investors are enduring of late in equities as well as precious metals and cryptocurrencies. With market valuations by many metrics in territory not seen since the end of the Internet Boom, I am convinced the overall market is in an AI driven bubble. It is just a matter of when investors’ perceptions change, and equities enter a bear market in my opinion.
The challenges of the AI revolution are becoming more apparent by the week. In retrospect, the birth of the internet was relatively an "asset lite" model in comparison to the trillions of dollars being currently spent on the buildout of AI infrastructure. Yes, the previous technology paradigm birthed names like Webvan, Pets.com and TheGlobe.com that came public to much fanfare and would never make money before exiting to the dustbin of history. But they were never linchpins to the narrative driving the markets. Webvan reached a market cap of just under $5 billion at its height in 1999. That is less than half of what OpenAI reportedly lost in the third quarter of this year. And these names never entered into commitments to purchase over $1.4 trillion in future computing power either.
OpenAI is currently trying to raise huge additional amounts of capital even as ChatGPT continues is losing market share at an accelerating rate since the recent debut of Gemini 3. Rumors abounded last week that Amazon (AMZN) might throw in up to $50 billion into the pot. This followed speculation Microsoft might ante up to an additional $60 billion. Nvidia (NVDA) might put in another chunk of cash but likely not the $100 billion investment that had been rumored.
Until ink hits the paper, all of this is nothing but speculation. My view is OpenAI will not go public, as has been much talked about, at least not in 2026. An initial public offering would require a huge amount of disclosure. I believe the company would like to avoid this for now given its losses. Futures were down prior to the market's open today, partly on news Oracle (ORCL) is trying to raise up to $50 billion via equity and debt to finance its AI and cloud infrastructure buildouts.
The buildout of the internet was comparatively dirt cheap to what is taking place during the AI infrastructure boom across the nation. Meta (META) is building a data center complex in Louisiana said to be the size of 70 football fields and likely cost more than $10 billion by the time it is completed. And it is hardly the only data center Meta Platforms is constructing. The company’s capital expenditure guidance for fiscal 2026 is between $115 billion to $135 billion, nearly twice that of the previous fiscal year. Much of the financing for this massive buildout is coming via off balance sheet financing via special purchase vehicles or SPVs.
To put this in perspective, Facebook paid $1 billion to acquire Instagram in 2012 and just under $22 billion to purchase WhatsApp in 2014. And these businesses did not require more than the required electrical generation for New Orleans on a hot summer day to power one huge data center complex. The biggest question for investors around the AI Revolution is will revenues ramp up in a trajectory that even begins to support the hundreds of billions of dollars being devoted to AI infrastructure annually.
I believe at some point in the coming year, the market reply to that inquiry will shift to a, "No." This obviously will have substantial impact on equities, given most of the gains in the market over the past three years have been powered by the AI Narrative. Therefore, I continue to be positioned quite cautiously within my own portfolio in anticipation of lower entry points in the coming quarters.
At the time of publication, Jensen was long AMZN.
