Question of the Day
One question per day to look beyond the headlines.
What does DeepSeek’s “permanent” 75% API cut reveal about its compute supply and export-control constraints?
Take-away Calling the cut “permanent” is a demand-lock tactic: it pre-commits devs before Ascend 950 capacity ramps, since export curbs cap Nvidia supply and keep compute scarce.
DeepSeek's permanent 75% price cut for its V4 Pro model API reveals several strategic and market conditions related to compute supply and export-control constraints. The price reduction is a strategic move, recognized as a pre-commitment to secure developer usage before the supply of Huawei Ascend 950 processors can scale to meet demand. There are supply chain constraints as the 750,000 Huawei Ascend 950 units planned for production in 2026 are insufficient to meet both current and anticipated demand, reflecting the impact of U.S. export controls on Nvidia chips which have significantly affected the semiconductor ecosystem, pushing DeepSeek to capitalize on Huawei's domestic chip developments. This reflects ongoing capacity constraints faced by Western AI providers, which DeepSeek aims to exploit by being positioned as a low-cost yet high-usage option in the AI market [1], [2], [3].
- DeepSeek cuts V4 Pro prices by 75% and locks in permanent discount | Ukraine news - #Mezha mezha.net (opens in new tab)
- DeepSeek ignites AI price war with huge price cut | Al Bawaba albawaba.com (opens in new tab)
- 750,000 Chips, 140 Trillion Tokens: The Math Behind DeepSeek's Permanent Price Cut - DEV Community dev.to (opens in new tab)