OpenAI's $1.4 Trillion Commitments Could Be Key to AI Bubble Bursting

Stocks ended November with a memorable rally. In fact, it was the best Thanksgiving week of trading in more than a decade. 

The S&P 500 rose nearly 4% during the holiday-shortened trading week and the Dow was up just over 3%. That was just barely enough to push each index’s monthly winning streak to seven. The NASDAQ was up better than 4% for the week but still was down almost 2% for the month of November.

December is historically one of the strongest months of the years for stocks, with equities advancing during the month roughly three quarters of the time. Therefore, it would not surprise me at all to see some sort of traditional Santa Claus rally, even as the overall market is trading at extreme valuation levels based on many historical metrics. However, once we get into 2026, all bets are off.

Crypto has recently entered a bear market. At some point in 2026, I expect the AI narrative to fade substantially. Given that the AI revolution has powered the majority of the gains in equities for three years now, that could trigger a significant correction in the overall market. There are many potential things that could eventually pop the AI bubble. 

The space is being powered in a large part by the vendor financing like what was seen late in the internet boom, but on steroids. OpenAI should book $14 billion in revenues this year and is bleeding cash. Despite this, the company has made some $1.4 trillion in commitments for future computing power. Everyone that is someone in AI, from NVIDIA Corporation (NVDA)  to Oracle (ORCL)  and Microsoft (MSFT) , is depending on OpenAI meeting those promises.

There are also legitimate concerns whether AI revenues will deliver enough growth over the next few years to begin to justify the massive investment being made building out AI infrastructure. There are also questions around whether the electrical grid can be expanded fast enough to meet the demand from these massive data centers. I also have concerns about more tech firms starting to depreciate GPU chips on a five-to-six-year depreciation schedule, when these chips are likely to have a two-to-three-year useful life.

How adversely further integration of AI into operational processes will impact the labor markets is also an open question. There were more than 153,000 jobs purged in layoff announcements in October, the largest number of jobs displaced during the month of October in over two decades. Hewlett Packard Enterprise Company (HPE)  announced last week that it plans to lay off up to 10% of its workforce.

I expect both residential and commercial real estate markets to continue to deteriorate in 2026. The ending of COVID loss mitigation programs will likely trigger a huge surge of FHA mortgage foreclosures starting in the second quarter of next year. Student loan delinquencies continue to soar and subprime auto loans that are 60 days or more have hit record highs.

Given these factors, I expect consumer sentiment to remain in the dumpster throughout 2026. This negative outlook for the economy and markets in 2026, presents a dilemma for me as I have a ton of covered-call positions that will expire in the money come December 19. This will boost my cash allocation in my portfolio to approximately 45%.

This is substantially too high, even with my pessimistic view around most of the market. In Wednesday’s column, I will highlight a few stocks that still sport reasonable values in an overbought market. They will get some of that additional dry powder when it soon becomes available.

At the time of publication, Jensen had no positions in any securities mentioned.