OpenAI asked Trump administration to expand 35% CHIPS Act credit to lower cost (www.bloomberg.com)

OpenAI reportedly asked the Trump administration to expand a 35% tax credit under the CHIPS Act to reduce the cost of semiconductors used for AI. The company’s push seeks broader coverage or a larger incentive so that more AI-specific chips and related production investments become cheaper to design and manufacture in the U.S. That lobbying move frames chip incentives not just as industrial policy for fabs, but as a direct subsidy to the economics of model training and deployment. For the AI/ML community this is important because cheaper, domestically produced accelerators would lower capital and operational costs for training large models and running inference at scale, potentially accelerating innovation and lowering barriers for smaller teams. Technically, expanded credits could encourage more investment in advanced nodes, custom accelerators, packaging and power-efficiency improvements—all of which influence model architecture, throughput, and cost-per-token. There are also policy implications: boosting U.S. chip capacity affects supply-chain resilience, global competition, and export-control debates around dual-use AI hardware. In short, a targeted expansion of CHIPS incentives could shift where and how AI hardware is built, who can afford it, and how quickly new compute-intensive models are developed.