Tomasz Tunguz: AI Infrastructure Bottlenecks Drive Up Costs

Venture capitalist Tomasz Tunguz points out that AI infrastructure is facing a long tail effect, with bottlenecks cascading from GPUs to memory, CPUs, and storage. This has caused each stage to freeze the supply chain of the next stage for years, locking in higher baseline costs. At the beginning of 2023, the GPU shock caused H100 rental prices to exceed $9 per hour, leading to a 22% decline in server shipments. Manufacturers subsequently shifted capacity to HBM, resulting in an 80% quarter-over-quarter increase in enterprise SSD prices and over a 60% rise in DRAM prices. By the end of 2025, the workload of intelligent agents will push the CPU to GPU ratio to about 1:1, with the average price of server CPUs rising 27% year-on-year; by 2026, nearly all annual capacity for nearline HDDs will be sold out. The construction costs for data centers have risen to about $20 billion per gigawatt, with orders for long-lead-time equipment such as transformers and turbines scheduled out to 2029 to 2031. Tunguz refers to this as the long tail effect in the hardware sector, where manufacturing delays amplify downstream demand shocks, and pressure will be delayed in transmitting to the next stage after a bottleneck is alleviated. The expansion of transformers, NAND wafer fabs, and turbine production lines scheduled for delivery in 2027 to 2028 may face the risk of overcapacity.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.