By: Takunda Basvi

The rapidly expanding AI infrastructure boom, driven by companies like Oracle and CoreWeave, resembles the Dot-Com bubble, creating an unstable market fueled by hype, monopolistic structures, and investor money that risks an eventual economic crash.

As the AI Infrastructure bubble continues to be rapidly fueled by investors and companies, the risk of economic troubles such as bankruptcy and a potential stock market crash is increasing. Similar to the Dot Com crash of the early 2000’s, the reliance on promise of profitability rather than practicality has emerged from excitement by investors of AI novelties. Signs akin to the Dot Com Crash have begun to show, especially regarding the latest major players of Ai Infrastructure: Oracle, and Coreweave. 

Oracle was founded in 1977 as a Silicon Valley startup, went public in 1986, and became the world’s largest database management company in 1987. They launched their first autonomous database in 2018 and developed it to include an AI feature in 2024. On the other hand, Coreweave was founded in 2017 to further computer capabilities for intensive tasks. They targeted a gap in high-performance computing resources, wanting to offer modern alternatives to traditional cloud services better suited for AI. Both companies are currently dominating the AI infrastructure space in competition with popularized companies such as Microsoft Azure, Google, and Amazon Cloud services. The existence of these companies is working three-fold: Saturating the AI market, Setting up a Monopoly, and misguiding others, factoring in the projected economic troubles from the AI bubble. 

The AI infrastructure space has become too saturated, dissuading others from the growing monopoly structure while fueling the bubble with empty and inflated investments. Although each new AI startup markets itself as having unique and untouched ideas, its interfaces all revolve around OpenAI’s API. This is best explained by AI expert and podcast host, Srinivas Rao, who writes: “The labels have changed, but the logic hasn’t. “AI-powered” is the new “.com.” Startups pitch wrappers. But this time, many don’t even pretend to own the tech they’re built on.”  In other words, there is no substance to the AI cloud space, each startup is an accessory to OpenAI’s behemoth of loyal customers: Perplexity, Notion AI, Puzzle, e.t.c. Each of the mentioned AI wrappers depends on OpenAI for the base of their programs. Moreover, the industry has formed a hierarchy in which the bottom startups fuel the higher ones (OpenAI, Microsoft, NVIDIA) with exaggerated investments, while taking up space for other potential startups. 

Although OpenAI is a major player among AI wrappers, it is dependent on Microsoft’s software (with a decrease in relations as they partner with other cloud companies). Unsurprisingly, Microsoft depends on the invaluable NVIDIA chips to run its own software and so on. Due to this hierarchy, the AI market has become unstable, with empty promises. It only has directions and space for the top-selling and advanced companies— like the dot.com crash, where established companies such as Amazon and eBay survived, while the others went bankrupt. 

With about 150 million startups worldwide (1.14 million in the US alone), each one thirsting for the promise of profitability, making it nearly impossible to succeed in that industry. On top of this, statistics are also against you as 95% of startups fail while a mere 5% generate rapid revenue. This is mostly attributed to lack of product demand. A large market is bound to have everything you can think of, in turn shutting down most ideas. The average 40,000-dollar cost to start a business is another factor, as most individuals do not easily have this amount of money. Most companies also get successful after being acquired, and who knows when their startup will be acquired or invested in by major companies? However, if the market continues to be as “exclusive” as it is, this could positively slow the AI boom but make the GDP growth gradual and suspend the creation of new jobs. The main issue is not casual investment into AI infrastructure but the speed at which it is occurring.

Although it is easy to mainly blame investors, we must also hold the executives who are fueling this bubble accountable. Recently, Sam Altman (Chief Executive Officer of OpenAI) said in an interview, “I know it’s tempting to write the bubble story….[with OpenAI] there’s something real happening here,” referring to the AI space. On the other hand, the Bank of England, the International Monetary Fund, and Jamie Dimon, JP Morgan’s boss, have warned against the AI bubble and encouraged speculation. With all of these well-established individuals and companies vying for and against the AI bubble, it has confused and pitted investors against each other.

Interestingly, JP Morgan has recently started heading towards becoming the world’s first ”AI mega-powered bank.” Currently, it is the world’s biggest bank by market capitalization and with the integration of AI, it is set to reach new heights. Combining Dimon’s view on the AI bubble and his new initiatives, it could be inferred that JP Morgan is trying to back others out of the game to pave the way for themselves. On top of this, a CNBC article with Derek Waldron, Chief Analytics Officer at JP Morgan on this issue, states “If JPMorgan can beat other banks to the punch on incorporating AI, it will enjoy a period of higher margins before the rest of the industry catches up.“ JP Morgan seems to be playing a fast game motivated by profit and not the good of others. On the other hand, Altman’s company is gaining exponentially off the Ai bubble–specifically off small startups that use its interface. The more AI wrappers equals more customers and eventually more money. 

Additionally, Avi Gesser, who advises companies on AI, states “Despite market concerns that the AI trade is a brewing bubble, corporate clients are actually more worried now that if they don’t start adopting it soon, they’ll fall behind and lose share.”Gesser is basically outing the incentive of all this ruckus: profit. Even if AI seems a little bit profitable, companies will gnaw at it. Gesser’s comment can also be physiologically supported. In economics, the ventromedial prefrontal cortex (VMPFC) located in the brain is infamous for “money illusions,” in which investors ignore blatant information and facts regarding the inflation of a product and irrationally invest into it. This is exactly the issue at hand–the VMPFC seems to be working over time, causing this AI frenzy, which is bound to fail. 

As we have seen time and time again, the AI bubble is a suffocating and toxic one. It is very blinding in its promises, and creates a false sense of security: South Sea, Dot Com, Tulip mania, e.t.c. The aforementioned incidents share a similar downfall. The Dot Com bubble ballooned in the late 1990’s, peaked in 2000 then burst later that year, leading to a devastating recession. Economic bubbles have always been an issue, even dating back to Tulip mania in the 1630’s, where tulips were fashionable causing a skyrocket in their market value and then a tragic collapse. Unfortunately, there are many examples of the bittersweet repetitiveness of history, proving that rapid investments usually do not end well–which makes the formation of this AI bubble a scary one. Moreover, this bubble does not only affect products, investors and businesses,  it also affects normal everyday people. It affects groceries, housing prices, and other basic needs. When the bubble does eventually pop, everyone will be affected and the companies profiting from this will be left standing–high and mighty, leaving the rest of us to ruin, which is why AI infrastructure should slow down and grow steadily.  

Leave a Reply

Trending