Large-scale investments by technology giants in AI infrastructure have yet to be reflected in revenue growth from this sector, noted the resource. According to David Cahn, an analyst at Sequoia Capital, AI industry companies will need to generate around $600 billion per year to offset their AI infrastructure costs, including data centers.
Last year, Nvidia's revenue from data center equipment sales amounted to $47.5 billion. A large portion of this equipment consisted of computer accelerators for AI applications and high-performance computing.
Major IT companies such as AWS, Google, Meta, Microsoft, and many others invested significant resources into their AI infrastructure in 2023 to support applications like OpenAI ChatGPT. The question arises: "Will they be able to recoup these investments?" In David Cahn's opinion, this may indicate that we are witnessing the growth of a financial bubble.
Cahn's calculations are quite straightforward. He initially doubled Nvidia's current revenue forecast to cover the total costs of AI data centers, half of which are due to GPU expenses, while the remainder includes costs for energy, buildings, and backup generators. Then, the analyst doubles the amount again to ensure a 50% gross profit for end-users like startups or enterprises purchasing computing AI resources from companies like AWS or Microsoft Azure, which also need to turn a profit.
Cloud service providers, such as Microsoft, are investing heavily in GPU purchases. According to Nvidia, half of its data center revenue comes from large cloud service providers. Notably, Microsoft contributed about 22% of Nvidia's revenue in Q4 of the 2024 financial year.
Meanwhile, in Q1 of the 2025 financial year, Nvidia sold GPUs for data centers worth approximately $19 billion. With the release of more powerful and expensive Nvidia B100/B200 chips, the amount of investment will continue to rise.
While OpenAI, the developer of the AI chatbot ChatGPT and leveraging Microsoft Azure infrastructure, achieved a substantial revenue increase—from $1.6 billion at the end of 2023 to $3.4 billion in 2024—other AI companies have yet to reach even the $100 million revenue mark. Nonetheless, investments in AI equipment are growing, noted Tom’s Hardware.
As Kan states, even the optimistic forecasts of revenue from major tech companies for AI do not meet expectations. Assuming Google, Microsoft, Apple, and Meta will each earn $10 billion annually from AI solutions, while other companies like Oracle, ByteDance, Alibaba, Tencent, X, and Tesla could generate $5 billion each, there remains a gap of $500 billion.
The analyst believes the AI market faces a long and arduous journey, as enterprises and startups still need to invent applications that yield real profits. Kan thinks the industry should temper expectations for quick gains from advancements in AI, acknowledging the speculative nature of current investments and the need for sustainable innovation and value creation. Otherwise, a bubble worth hundreds of billions of dollars may burst and collapse the market.
Source:
Source: 3dnews.ru
