In every new relationship there comes a moment when both parties ask themselves “What have I overlooked about you because the sex was so amazing?” In that moment you start to see that the person you think you love drinks too much, has too much debt, works too hard, snores, and their families are psychotic. And then, right when you moved a box of stuff into their apartment, you have “the talk” about who you want them to be. They swear they will change. But they usually can’t. And this is how new couples break up.

The astounding hype about generative AI has caused people to overlook who Open AI, Google, Anthropic, Nvidia, and Microsoft depend on for continued growth. For this boom to continue, companies that are politically dysfunctional, bloated, risk averse, and heavily regulated will have to transform in ways that they never have and probably never will. The continued growth of AI depends on companies like JP Morgan, Goldman Sachs, Exxon Mobil, and Ford becoming someone else.

In this article I talk about the breakup between the sellers and buyers of AI that will cause this AI bubble to pop.

The headline

Every tech bubble bursts because at some point there isn’t enough demand for all the innovation supply. The tech companies who make AI need people and companies to buy it. Since 2023, for every dollar of AI revenue that Amazon, Alphabet and Amazon generate, investors added $150 to their stock price. Prior to the public debut of ChatGPT investors added about $8 to their stock price for every one dollar of revenue. Nvidia is trading at 47X revenue. Valuations are becoming detached from value. 

Investors assume that demand will grow exponentially and indefinitely. But at some point massive, corporate buyers will cancel orders. And then a headline will appear that will spark a panicked sell-off: “Open AI Revenue contracts 42% from last quarter As Citi, Exxon Mobil, Goldman Sachs Cancel Orders.”

I believe this headline is inevitable because I spent too much of my career inside non-tech companies trying and failing to integrate emerging technology into their core business. When you observe AI through the lens of the buyers, rather than the producers, it looks very different. 

Sam Altman, Meet Frank

Every huge corporation outside of the tech industry is a patchy network of tech fiefdoms who hate each other. Some of these fiefdoms are centers of technological excellence. But they are hard to find, and are often political punching bags for everyone who’s less advanced but more politically savvy.

And then there’s “Frank.” 

Frank is a bloated, angry, territorial upper middle manager who controls a $5 billion business with a filing cabinet and a fax machine. He’s been doing it this way since the Clinton Administration. There is a Frank in every bank, law firm, consulting company, insurer, oil driller, miner, and governmental department. Frank is why the user interface of your banking app is indecipherable. Frank is why it takes more time, money and expense to file an insurance claim today than it did in 1997. It’s why many banks still have job postings for COBOL developers. 

Frank is also why incredible tech talent never stays at huge banks. For every 4 AI experts that the biggest banks have hired over the past two years they’ve lost 5. Every tech innovator would rather cut their pay and leave finance, then spend the rest of their life fighting with Frank. 

Frank is why, who, and how this AI bubble bursts. 

The Signal is the Noise

I know that this AI bubble will eventually slam into Frank because bank executives talk about AI today as if they’re doing it right already. As though Frank has been a champion of progress, rather than an innovation obstructionist. The AI noise generated by bank executives is actually the signal that points to AI’s future demise.

Here’s Jane Frasier, CEO of Citi, in an article she posted on LinkedIn

“Citi’s innovation labs have been working on AI — including the kind of generative models that power ChatGPT — for the past three years.”

Those of us who worked on the inside, know that innovation labs are a negotiated peace treaty that was structured to placate Frank. The terms of that detente basically state: “We promise to never commercialize any of the technology that we use inside of the lab. We furthermore swear that we will never intervene with any of the biggest profit centers unless we receive a clear invitation and know our place.” An innovation lab on Wall Street is code for an innovation quarantine. Yes, there are exceptions. But they are rare.

Jamie Dimon, Chairman and CEO of JP Morgan, said something similar in his 2023 letter to shareholders:

“Since the firm first started using AI over a decade ago, and its first mention in my 2017 letter to shareholders, we have grown our AI organization materially. It now includes more than 2,000 AI/machine learning (ML) experts and data scientists. We continue to attract some of the best and brightest in this space and have an exceptional firmwide AI/ML and Research department with deep expertise.”

The message from both is that the massive cultural overhaul that would be necessary to empower technologists to transform the most inefficient parts of the business with AI is not a priority. Because they’ve already done it. 

There is no AI revolution if there is no corporate culture transformation. The biggest barrier to AI companies living up to their valuations is a culture of innovation complacency and obstructionism inside of the industries that are their customers. At a certain point, if these buyers can’t change, they’ll stop buying. That’s when this bubble will burst.

One More Thing… Punks & Pinstripes is hiring a community events manager. Basically, this person will take the events we’ve been producing for our members, do more of them, and make them better. It’s a hugely important job and requires a logistical badass who can balance punk flair and professional finesse. The right candidate must live in NYC. See the posting here.