This is interesting. Goldman Sachs expects market growth for rack-level AI servers to slow due to factors like the effects of the product transition period, ongoing supply and demand uncertainty, and increasing production complexity.
As a result, it has revised its shipment forecasts for 2025 and 2026 downward—from 31,000 units and 66,000 units to 19,000 and 57,000 units, respectively (based on a 144-GPU equivalent).
Alongside this, Goldman Sachs has also reduced its target prices for several Taiwan-based companies in the AI server supply chain.
More:
- demand for high-performance AI Servers will not be completely replaced by rack-level forms
- there is optimism regarding the outlook for AI inference and general Servers, driven by the application of AI technology and the recovery of the upgrade cycle