Peter Berezin, Chief Global Strategist at BCA Research, said in a post on X late Thursday, that major cloud and AI infrastructure providers, also known as hyperscalers, could be carrying more than $2.5 trillion in AI assets by 2030.
Based on a typical 20% depreciation rate, Berezin estimates these firms, which include Big Tech names like Microsoft (NASDAQ:MSFT), Google-parent Alphabet (NASDAQ:GOOG) and Amazon (NASDAQ:AMZN), may face around $500 billion per year in depreciation expenses later this decade.
Depreciation Could Outpace Profits
Berezin noted that such a level of depreciation would exceed the companies' combined projected profits for 2025, raising questions about the sustainability of the current pace of AI investment.
His projections came after a wave of volatility swept Wall Street on Thursday, dragging tech stocks lower.
Investor optimism over Nvidia’s (NASDAQ:NVDA) strong earnings was tempered by renewed concerns about the lofty valuations of artificial intelligence companies and the rapid pace of investment by major tech firms such as Amazon, Meta (NASDAQ:META), and Oracle (NYSE:ORCL) in the data centers needed to support generative AI.
The Nasdaq Composite ended the day down 2.2% after an early surge, while the S&P 500 fell 1.6%. The weakness extended to Asia on Friday, where major indexes also retreated.
If the Music Plays, You Have To Dance
Still, when one user asked why companies would keep spending if the numbers look this unfavorable, Berezin replied, "If the music is playing, you have to keep dancing."
Other users were divided on his outlook. Some argued that depreciation could squeeze margins if AI revenue growth slows, while others noted that hyperscalers can extend equipment lifecycles or repurpose older hardware for less compute-intensive tasks to mitigate the financial impact.
Berezin's projection adds a new data point to concerns that the AI arms race could carry a much heavier accounting burden than investors expect.
READ NEXT:
Image via Shutterstock
