Since the AI mania began in 2023, there have been voices suggesting that it is a dot-com-like bubble. This chatter gained traction last year, as many companies struggled to justify their massive AI capital expenditures with commensurate revenues.
DeepSeek’s low-cost AI model, released earlier this year, gave credence to the “AI bubble” narrative, as the Chinese startup claimed to have developed its AI model for a fraction of the billions of dollars that giant US tech companies spend on their products.
However, AI stocks have managed to overcome the pessimism, and AI trading leader Nvidia (NVDA) has risen to record highs and become a $4 trillion giant despite losing business in China.
www.barchart.com
Meanwhile, over the past couple of months, chatter about AI being a bubble has gained momentum. He joins the ranks of Federal Reserve Chairman Jerome Powell, Goldman Sachs (GS) CEO David Solomon, OpenAI CEO Sam Altman, Meta Platforms (META) CEO Mark Zuckerberg, and Amazon (AMZN) founder Jeff Bezos, who have warned of a potential intelligence bubble. Artificial, in one way or another.
While Powell was very astute and talked about “unusually large amounts of economic activity through building AI,” Bezos was perhaps the most outspoken and said there is an “industrial bubble” in AI.
Fears about an AI bubble are certainly not unfounded, especially in the startup space where companies are getting eye-popping valuations. For example, OpenAI was valued at $500 billion in a recent deal that provided liquidity to employees, while Elon Musk’s xAI is said to be worth $200 billion.
The virtual scramble for AI talent, where big tech companies, especially Meta, scoop up talent for eye-popping compensation, also raises concerns about a bubble.
Praetorian Capital has done some numerical analysis, which shows that the so-called hyperscalers could collectively spend $400 billion on data centers this year, the value of which would decline by about $40 billion annually. However, where things get worse is that the devaluation is double what these companies are expected to receive in revenue (not profits) from AI this year.
The rise in the value of the currency means that technology companies will feel pressure on their bottom lines. For example, Meta expects total expenses to grow 20% to 24% this year and warns of “significant upward pressure” on expenses in 2026. The company expects higher depreciation expenses on incremental capital expenditures to be the biggest driver of expense growth, followed by employee compensation.
Meanwhile, Meta is seeing the benefits of AI, and in the second quarter of 2025, its revenue grew 22% year-over-year, which beat estimates, which Zuckerberg attributed to “AI unlocking greater efficiencies and gains across our advertising ecosystem.”
www.barchart.com
Specifically, he said the AI-powered recommendation model helped increase ad conversions by approximately 5% on Instagram while the corresponding figure for Facebook was 3%. While growth may not yet justify the huge capital expenditures toward technology, there are signs that things are moving in the right direction.
Notably, companies like Meta and Google (GOOGL) have the advantage of a captive user base, for which they can offer AI services, and during the Q2 2025 earnings call, Zuckerberg said that Meta AI now has more than 1 billion monthly active users.
Another pillar of Meta’s AI strategy is hardware, and it is betting on glasses as the next computing platform. The company is not alone in this thought process, as Apple (AAPL) and Alphabet are also working on smart glasses.
Business AI is another key pillar of Meta’s AI strategy, and Zuckerberg believes companies will “soon own business AI just like they own an email address, a social media account, and a website.”
Besides AI, Meta has several other growth drivers, including monetizing WhatsApp’s user base. The company will also eventually be able to start monetizing Threads although that won’t be a significant driver in the near term.
While the froth in the unlisted AI ecosystem is quite evident, I don’t see much euphoria, other than a few pockets, in the listed space. For example, the Meta is trading at a forward price-earnings multiple of 25.84x, which is nowhere near bubble territory.
I continue to believe Meta is among AI’s top plays, and the company’s core digital advertising business continues to perform well, with AI only fueling its growth. While the stock remains vulnerable to any potential AI collapse amid bubble fears, I continue to invest and find current levels attractive enough to stimulate new purchases.
On the date of publishing, Mohit Oberoi held a position in: META, GOOG, AAPL, AMZN, NVDA. All information and data contained in this article are for informational purposes only. This article was originally published on parchart.com
