Is the AI Bubble Bursting?

If you looked at the chart for Nvidia, the poster-child for this artificial intelligence bubble that is racing Apple and Microsoft to be the most valuable company in the world, you would think this title is crazy. As of this writing, it is almost challenging its previous all-time high and looks poised to push higher, as the entire market continues this astonishing bull run from the ashes of the 2022 crash.

But if you ask Goldman Sachs, yeah, absofrigginloutely. They released a scathing 31-page report last week titled “Gen AI: too much spend, too little benefit?”

“Tech giants and beyond” will spend over one trillion dollars in AI capital expenditures in the coming years “with so far little to show for it” says the famed vampire squid who only sees the world in dollars and cents. We have been sold a lie by hype merchants posing as technologists like OpenAI’s Sam Altman, and now that AI has existed for some time, their hyperbolic claims are coming into contact with reality. Altman really asked for $7 trillion dollars to make his chatbot work the way he claims it can, yet Goldman throws cold water on the notion that OpenAI could even generate that much revenue to justify the investment. That we still take this clown seriously after requesting one-third of annual U.S. GDP with a straight face is an indictment of the shallowness of most tech and business media coverage.

It’s not just the investment banking giant who is doubting what is starting to look like the most overpriced asset class since the tech bubble of the early 2000s, but yesterday Microsoft surprisingly quit OpenAI’s board now that regulators are scrutinizing a business whose entire model is dependent on taking other people’s content and reproducing a Frankenstein version of it, then branding it as “intelligence.”

Alex Haffner, competition partner at U.K. law firm Fladgate, said in a statement to Forbes that “It is hard not to conclude that Microsoft’s decision has been heavily influenced by the ongoing competition/antitrust scrutiny of its (and other major tech players) influence over emerging AI players such as Open AI.”

Reports like this from 404 Media highlight the litany of legal issues that AI creates for itself. It’s not difficult to see how even a light regulatory touch could crater parts of an industry that produces results like these:

This month, “Christina Warren” started blogging again for The Unofficial Apple Weblog (TUAW), a legendary and long-dead Apple-centric tech news blog that she worked at more than a decade ago. Warren was for years a well-known and very good tech journalist, before she went on to work for Microsoft and GitHub. The real Christina Warren hasn’t been writing these new posts on the zombie TUAW, however. The site’s new owners have stolen her identity, replaced her photo with an AI-generated one, and have been publishing what appear to be AI-generated articles under her byline.

Additionally, Microsoft is not the only tech giant backing away from this overheated and overhyped industry, as Oracle was unable to come to an agreement with big dumb manbaby Elon Musk over expanding an existing arrangement where Musk’s xAI rents Nvidia AI chips from Oracle. Elon-centric quirks influenced this deal, as it fell apart over his demand to build supercomputers faster than Oracle deemed possible according to The Information. However, Oracle refusing to give into one of the AI hype merchants is another data point that existing big players are tapping the breaks to varying degrees on this technology that still has yet to produce a real mass-market product.

Wendys introduced a “FreshAI” ordering system and fourteen percent of its orders required human intervention to actually complete, as companies around the world are beginning to really scrutinize how AI actually impacts their business. Copywriters who lost their jobs to AI are now even being paid to make AI copy sound more human. AI companies themselves are also slimming down, using less powerful models as gigantic server and energy costs mount. At the very least, we are witnessing some kind of sea change in this nascent industry, as the grand promises of ultra-intelligent everything bots have given way to specific applications more grounded in reality.

A lot of people who work in AI bristle at the hostility some of the public has shown it, but the skepticism expressed is simply proportional to the grand promises being made. That companies are now dialing back their aspirations and operations is an admission that its critics at least had a point decrying all that excess that created this bubble.

So what is influencing this change? Professional-services firm KPMG conducted a survey of 100 major corporate executives making it pretty obvious: big firms now look to revenue and not productivity as their primary gauge of AI’s usefulness. That’s not good for the present state of AI valuations!

Many AI defenders like to compare it to the early days of the internet, but Goldman Sachs’ Head of Equity Research Jim Covello rightly calls bullshit on this babybrained logic, as the cost-benefit dynamic of AI is the opposite of the early web, pointing out that “Even in its infancy, the internet was a low-cost technology solution that enabled e-commerce to replace costly incumbent solutions. AI technology is exceptionally expensive, and to justify those costs, the technology must be able to solve complex problems, which it isn’t designed to do.”

Daron Acemoglu, an economist at MIT, told Goldman that “Given the focus and architecture of generative AI technology today… truly transformative changes won’t happen quickly and few—if any—will likely occur within the next 10 years.” In light of this long-term expectation that is shared by more rational AI thinkers in the industry, the slimming down of AI companies makes all the sense in the world. They are simply reorganizing the business to reflect reality as it is and not as they have sold it to everyone including themselves.

So is the AI bubble bursting? In one sense, absolutely. In another, absolutely not. Companies are still investing huge amounts of money to develop this technology, as that ten-year-plus time horizon Acemoglu laid out is what a lot of this capex investment is targeting. However, the immediate AI bubble is another story.

The word “bubble” appears eleven times in Goldman’s report, and while everyone using it broadly agrees that current AI business valuations are vastly overpriced, what will happen in the near future is much less certain. “Bubbles take a long time to burst” appears twice in the report, and if there is one thing I took away from my finance master’s program, it’s that the general trend of the market is up. Ten percent annual S&P 500 returns is the baseline expectation and the number most people aim to beat. Everyone could spend the next six months echoing Goldman and proclaiming AI to be completely bunk while more companies follow Microsoft’s lead, yet we could end the year with Nvidia being the world’s first $4 trillion company for all we know. If you are not humble in the face of the market, the market will humble you.

That said, from where we stand today, Goldman finds it very difficult to envision how AI will improve overall market returns, as they wrote on page 3 of their report:

GS senior multi-asset strategist Christian Mueller-Glissmann finds that only the most favorable AI scenario, in which AI significantly boosts trend growth and corporate profitability without raising inflation, would result in above-average long-term S&P 500 returns, making AI’s ability to deliver on its oft-touted potential even more crucial.

Reputationally, AI definitely seems to have reached at least a local top. Google rolling out their search-AI which recommended that people eat rocks and put glue on their pizza was as high-profile a failure as this technology has seen. ChatGPT can help with repeatable tasks but if you ask it to do 4th-grade-level math it all of a sudden forgets that it’s a computer. AI is still a technology that is evolving, and what it becomes is anyone’s guess, but the people like Sam Altman who have been selling it to us as something it is manifestly not are now being exposed as the charlatans they have always been. If the bubble does burst, his lone remaining shred of public credibility will fall along with OpenAI’s valuation, as his avalanche of lies about his business’s capabilities are now becoming more apparent for all to see.