The increased spending on AI chips by tech companies is driving a competitive 'arms race' in the industry.

• This week, executives from Meta and Google acknowledged that their companies might be investing excessively in AI infrastructure.
• However, they said the risks of spending too little are far greater.
• Mark Zuckerberg, CEO of Meta, stated on a podcast that being behind in technology can result in being out of position for the most important advancements for the next 10 to 15 years.

Mark Zuckerberg, CEO, has amassed a significant collection of chips through the allocation of billions of dollars, enabling his company to create and hone sophisticated AI models.

But even he says the AI hype may be driving too much investment.

Zuckerberg believes that many companies are currently overbuilding, and in the future, they may realize they have spent more money than necessary, according to a podcast interview with Emily Chang of Bloomberg.

He's not the only one expressing that sentiment.

During a Wednesday earnings call, CEO Sundar Pichai admitted that his company might be overspending on AI infrastructure, mainly Nvidia's GPUs. However, he believes there is no other option.

Pichai stated that the risk of underinvesting is significantly higher than the risk of overinvesting when going through a curve like this.

Nvidia is experiencing significant growth in business from companies such as Meta, Alphabet, and others that have stated their commitment to investing in AI. Nvidia's revenue has increased by over 200% in the past three quarters and is projected to double in the current period.

This week, Alphabet and Tesla disclosed their AI development expenses during earnings calls, and investors can anticipate further updates from Microsoft, Amazon, and Meta when they release their results next week.

On Tuesday, Meta unveiled its latest Llama AI model, Llama 3.1, which is available in three different versions, including the most powerful and advanced AI model from Meta yet. Despite investing heavily in the underlying infrastructure, Meta is maintaining an open-source approach, allowing outside developers to access the technology for free.

Despite the high costs, Zuckerberg stated on the podcast with Chang that companies are making rational decisions in their AI investments.

Zuckerberg stated that being behind in technology for the next 10 to 15 years is a disadvantage because it puts you out of position for the most important technology.

Pichai believes that despite Alphabet's heavy investment, the infrastructure is highly beneficial to us.

'A threat and an opportunity'

Nvidia's stock has increased by 131% in 2023, following a 239% rise in 2023. The company's market value is now approximately $3 trillion, placing it third behind Apple and Microsoft. However, Nvidia briefly surpassed both companies in market cap in June.

Nvidia generates over 40% of its revenue from Microsoft, Amazon, Google, and Oracle, all of whom heavily rely on GPUs for their cloud services. Despite being among the world's most financially robust corporations, there is growing unease among investors regarding the company's stockpiling.

Last week, David Cahn, a partner at venture firm Sequoia, stated in a blog post that the driving force behind spending is competitive and follows game theory dynamics, resulting in a "cycle of competitive escalation."

"Cloud giants view AI as a potential threat and opportunity, and they cannot afford to wait for the technology to develop," Cahn wrote. "They must act immediately."

To justify the spending on data centers and chips in the technology industry, Cahn determined that there should be an annual revenue of $600 billion from AI.

On Wednesday, Cahn stated that Zuckerberg and Pichai's remarks about limiting downside strengthened his hypothesis.

Both Google and Meta CEOs have agreed with my AI Arms Race narrative that AI CapEx is driven by game theory and FOMO rather than actual revenue or usage, as posted by Cahn on LinkedIn.

Nvidia predicts that demand for its Blackwell AI chips will remain strong, with shipments beginning later this year. However, the company is now addressing investor concerns about return on investment as growth may slow due to challenging comparisons.

Nvidia's finance chief, Colette Kress, revealed in May that the company had determined that a cloud provider can rent out an Nvidia-based server for $5 after spending $1 on it. Goldman Sachs analysts stated in a recent note that Nvidia aims to share such data points to reassure investors.

On Tesla's earnings call on Tuesday, Elon Musk stated that the demand for Nvidia hardware is so high that it's challenging to obtain the GPUs. Additionally, Tesla disclosed that its capital expenditures on AI in the quarter totaled $600 million, as the company continues to invest in autonomous driving and humanoid robots.

Tesla is concentrating on creating its own Dojo supercomputer due to Nvidia chips being expensive and inaccessible.

""The high demand for Nvidia GPUs leaves us with no choice but to raise their prices to whatever the market is willing to pay," Musk stated."

WATCH: Investors should look forward to earnings catalysts for Tesla