Missing The Mark?

Is the AI Hype Ending?

Jul 07, 2026

Postponed IPOs, lack lust stock performance, and companies getting cold feet suggests one thing to me: the honeymoon phase is over for AI.

The last few weeks have been crazy, and because of personal and work reasons I couldn’t talk about it at the time. Looking back though, I am kind of glad that I waited till now, because the AI drama keeps getting juicer with each passing day it seems, and it all seems to be barreling towards one, unfortunate conclusion for AI. The postponed IPOs, lack lust stock performance, and companies getting cold feet suggests one thing to me: the honeymoon phase is over for AI. I know that sounds confusing, but if you’ll stick around and read, I’ll explain exactly what I mean.

The End of The Honeymoon

If you’ve had a strong healthy romantic relationship before (or if your still in one) then you will understand what I’m talking about. You fall head over heels for someone, you feel safe and understood, everything seems rosy. That deep sense of happiness, excitement, and attraction lasts for a month, two months, or longer, and it seems like nothing can go wrong. Then reality hits, all of the irritating things you ignored for the whole honeymoon phase still bother you and need to be addressed. In my experience, many relationships do not survive that sudden return to reality.

Just like a romantic relationship, the average investor, business owner, and consumer have been in an extended honeymoon phase with AI. From LLMs and hyperscalers to semiconductors and space companies, wave after wave of investors have chased the same story: AI is going to change the world and create a utopia, so any and all costs are worth it. However, the news over the last few weeks suggest that is changing. That shift, in my opinion, started with a little town in the middle of nowhere Utah.

Showing Your True Colors

In my opinion, in order to keep the hype going around something, you can’t let any facts that weaken the hype see the light of day. If you do, people start to doubt, and they stop giving all their energy to keep the hype moving forward. However, at a certain point, ignoring criticisms weakens your momentum, because it looks like your lying. Boxelder county and Kevin O’Leary experienced that, about a month ago (see my post about it here), when county commissioners walked away from a crowd of critics, they’re constituents, to approve the controversial Stratos Project Data center by a video vote.

That one action told the people of Utah (and thanks to mass media, the people of the world) exactly what they needed to know. The crowd of angry Utahn’s came with reasonable concerns around water usage, land usage, power, and carbon outputs, and instead of taking a moment to try and calm people’s fears, the county commission’s actions basically said:

We don’t care about your concerns, this data center will give us a big fat property tax check, which means they matter more than you.

That not only signed the death warrants for those commissioners’ careers (in my opinion) but in conjunction with stories like a data centers illegal water usage in Georgia, this action told people exactly what they should believe: the AI companies do not care about you and cannot be trusted (We’re going to circle back to this later). A belief that the next big news story solidified even further, OpenAI’s leaked financials.

OpenAI's Financials Leaked

A vindicating moment occurred for staunch AI critics in the middle of last month when the financials of the company who started the AI hype train, OpenAI, were leaked to the public. The numbers didn’t look good. According to the Motley Fool, OpenAI did 2.5x its revenue from 2024 to 2025, rising to $13.07 billion, however:

Their cost of revenue rose 183%
Their R&D cost rose roughly 146%
Their sales and marketing rose a whopping 416% percent from $1.11 billion to $5.73 billion

Wrapping those together with a 73% increase in general and administrative costs from 2024 to 2025, and you get to an operating loss (after a $41.55 non-cash charge from their non-profit-to-for-profit transition) a 6.57x increase from their 2024 loss of $5.09 billion (after a $3.74 billion removal of non-controlling member capital losses). Overall, the picture didn’t look great. The emperor of the AI trade is finally being forced to show their clothes and so far, it looks like OpenAI might be walking around naked.

Unfortunately, it seems like all the AI companies are in that same scary position. The AI companies so far have only shown that they’re extremely good at burning cash…and not much else. I don't know every single reason why that’s case, but one new update that has come to the spotlight: SpaceX’s missing profit story.

Space Fairing Data Centers?

There may still be money for the second big AI company, but not for the third - one the common themes from the recent Prof G Markets episode, OpenAI Hits Pause On It’s IPO

Space Exploration Company, or SpaceX shocked the world in many ways this last month, with one of the largest IPOs in history, and financing concerns directly out of the IPO gate. SpaceX jumped 40% in its first full week of trading, riding high on the expectation of some magical space AI future. However, since then, the markets seem to have woken up. SpaceX stock has fallen 22% from that peak, and market commentators are now wondering if everything was just hype.

TradingView chart — Created with TradingView

The news has come rolling in, the Motley Fool says the worst may be yet to come for SpaceX. The Cryptopolitan brought up concerns around the recent bond issue SpaceX used to pay off a previous bridge loan made to XAi. 2/47 WallSt’s Rich Duprey mentioned in a article from June 23rd that SpaceX’s all-stock acquisition of Anysphere (Cursor) raises dilution concerns. Forbes furthered that criticism, saying the acquisition spooked investors. Futurism reminded us that investors are still finding the billions in losses unpalatable. None of the articles touched what my biggest criticism was, the absurd TAM, but we’ve talked about that already on the newsletter. The point is, the SpaceX IPO has rattled AI investors, forcing them to ask the question: is this what the AI industry is?

The remaining two AI companies haven’t managed to fix the problems that are currently spooking investors. Anthropic says they’ve turned a profit this quarter, but a non-GAAP profit according to AIToolsRecap. In my opinion, this means very little until we see the GAAP financials. As for OpenAI, I think they know investors are nervous, combine that with the leaked financials, and we get the next blow to the honeymoon phase: the postponed IPO.

No IPO for OpenAI

Despite the meteoric SpaceX IPO, which should have given both OpenAI and Anthropic greens light for similar high-priced IPOs. However, it seems like OpenAI has gotten cold feet. I haven’t found a concrete reason for it, but the overall sentiment seems to be this:

The banks won’t pay $1 Trillion for OpenAI, because they can’t prove it’s worth that price.

For those who don’t know how IPOs work, when a company goes public, after they do their own due diligence, meet with an investment bank, determine a price, and then travel around the country trying to get people to buy at that price.

If you can’t get the price you want, you have to either negotiate or just take a lower price. For OpenAI though, that’s not an option. At a current post-money valuation of $852 billion, the father of the AI trade is priced for perfection. If OpenAI can’t get a $1 Trillion valuation, that could do some serious damage to the hype that is keeping this whole trade afloat. AI can’t afford more bad news, especially with the public and government is turning on data centers.

Data Center Drama

One of the most recent developments I’ve seen that has caught me off guard was the recent announcement that Maine banned data center construction in the state. What caught me even more off guard, was just how many states are following suit:

As of last week, according to this Newsweek piece and data from the National Conference of State Legislatures, 14 states have either proposed or are proposing data center bans and moratoriums. A couple of key news pieces stick out here:

South Carolina has put a block on local approvals for data centers until a statewide approval framework can be developed.
Pennsylvania wants a three-year moratorium on data centers, with the added burden of impact studies
After a letter from the Utah Senate President, Kevin O’Leary has agreed to reduce the Stratos campus to half it’s original size.

On top of everything, Governor Greg Abbot of Texas, a proponent of AI development, has “turned on” the data center build out, after Texans have been calling for “no more data centers” for months.

Much to the chagrin of the hyperscalers, VCs, AI labs, and the Trump administration, all of the actual consumers living where these monolithic data centers are set to be built are saying no, and no amount of fear mongering will change that.

So…What Happened?

AI was all anyone could talk about three months ago. The AI “jobpocalypse” drove lab valuations higher, sent the SpaceX stock “to the moon”, and punished the software stocks. Now it seems like all of that is unwinding, why?

I don’t know for sure, but my guess is people are waking up, from the consumer to the investor. Now that S-1s have been filed, and we have the chance to look at what’s really going on in the AI labs, we’re left with one question: so when are we getting our money back? When will this AI Utopia actually show up? AI lab CEOs, Hyperscaler CEOs, and the most vocal of the venture capital sect have been pounding the pulpit for years now saying that the world was about to change, a world where we would be “freed” from labor was “just around the corner”. In short:

Don’t pay attention to the man behind the curtain

However, the curtain has been drawn back. AI companies couldn't hide behind the private company excuse forever, burning investor funds while telling us how much demand they had. In order to return capital to the VCs who have been supporting them for year now, Anthropic and OpenAI had to go public, which meant showing their books to the world. What happened instead:

OpenAI postpones is IPO
OpenAI financials get leaked
The Anthropic CEO testifies that cheaper, opensource models are dangerous and get’s reamed for it across the social verse.

Overall, it seems like the AI companies have overpromised and underperformed, and there’s no evidence to suggest, to me, that this will change. Which is funny to me, because it looks like to me that AI, as a technology, will stick around.

As I am finishing this, a new nail in the AI coffin has appeared, with Meta saying that they will be renting out their AI compute, something they said they wouldn’t do unless they no longer had a use case for it. I don’t know what AI future will look like, but one thing I would say is, watch the news cycle with this one, the next six to twelve months will be wild.

Thank you for reading the article!

If you found this article valuable, then subscribe to the newsletter on LinkedIn or Substack. I am on a mission to make economics, finance, and business easier to understand for the average person, and each subscription helps!

WealthNWisdom

Discussion about this post

Ready for more?