Goldman Sachs Sounds Alarm on Big Tech’s $357 Bn AI Spending: Are Hyperscalers Overextending Their Budgets?

• Goldman Sachs estimates tech firms will pour an additional $1 trillion into AI investments in the coming years.
• Firms are investing approximately 72% of operating costs into AI capex and R&D
• The AI sector lacks a “killer app” and may only provide short-term gains for investors.

Analyzing the spending patterns of “hyper scalers” such as Alphabet, Microsoft, Meta, and Amazon, Goldman Sachs has warned of an AI bubble similar to that from the dot-com era of the 1990s.

In a quarterly report, big tech is revealed to have poured $357 billion into capital expenditure (capex) and research and development (R&D) over the past year, with a “significant portion” directed at AI.

Picks and Shovels Phase

According to the Goldman Sachs report  titled “Gen AI: Too Much Spend, Too Little Benefit?”, there is skepticism that generative AI is barely beginning to move past the “picks and shovels” phase as tech firms pour billions into the sector.

https://twitter.com/Carnage4Life/status/1809590088768405922

The report notes that generative artificial intelligence (AI) has a limited economic upside, no killer app, and cautions that the AI bubble could take a while to burst. It highlights that AI isn’t designed to solve “[…]complex problems that would justify the costs,” adding that there was “little to show” for spending so far.

Looking beyond the picks and shovels phase, Goldman Sachs’ Senior Equity Research Analyst, Eric Sheridan, optimistically noted :

“The handful of large tech companies developing the foundational models for generative AI are clearly well positioned. Semiconductor companies and the hyperscalers in cloud computing also look well-positioned to capture gains during the build phase”

Sheridan believes that after the introduction of AI models like ChatGPT from OpenAI, consumer Internet companies are moving into the “build phase”, in which they’ll develop foundational models, some being for specific enterprises/industries.

Following this, Sheridan notes that these solutions will make their way to market and be deployed, which is when we’ll find out if they work well enough to scale and garner mass adoption.

High Costs, Short-Term Gains

As the report explains, bubbles are a complicated market phenomenon often driven by “momentum and self-fulfilling” market dynamics. Perhaps most important, however, is the lack of a “killer ” AI application.

Goldman Sachs researcher, Jim Covello, has been critical of these hefty expenditures on AI, writing in a Goldman Sachs newsletter:

“The crucial question is: what $1tn problem will AI solve? Replacing low wage jobs with tremendously costly technology is basically the polar opposite of the prior technology transitions I’ve witnessed in my 30 years of closely following the tech industry,”

The report was critical of the cost of hardware and infrastructure required to run AI technologies, though it is for this very reason that stocks such as Nvidia continue to press upward .

“Nvidia’s substantial upward revision to revenue guidance in its Q1 earnings report, has triggered a surge in investor interest in generative AI technology.”

Ultimately, Goldman Sachs estimates $1 trillion in AI capital expenditure to flow over the coming years, and experts within the report have cast doubts over AI’s revolutionary impact in the short term.

AI Bubble

Despite a lower spending ratio compared to the dot-com era, Goldman Sachs’ lead strategist, Ryan Hammon, warned  that there’s still plenty of risk within the AI market. According to Hammond, the hyperscaler’s of today will at some point be required to prove that their revenues will be generated from these investments.

Hammond notes that whilst the AI spending frenzy “pales in comparison” to levels seen during the dot-com boom and crash, the Goldman Sachs report posits that it bears greater similarities to the tech bubble of the early 2000s.

“Adjusting for profits of these companies, the AI capex cycle still pales in comparison to the tech bubble,”

More specifically, during the 2000s tech bubble, tech, media, and telecom companies were pouring over 100% of their operating cash flows into capex and R&D. Today that figure stands at around 72%.

Hammond concludes that as demonstrated in previous bubbles, investors will be closely watching sales revenues as the key indicator to assess the integrity and durability of a given investment.