OpenAI's compute strategy treats data centers, chips, models, and products as one integrated system, with Jalapeño, its first custom inference chip, showing measured gains in throughput per kilowatt and token latency.

The AI hardware supply chain is a classic bullwhip: each bottleneck relieves only to trigger the next, with multi-year lags and ever-higher baseline costs.

Local AI models now match frontier cloud models on 89% of everyday queries, promising major energy and cost savings while cloud remains vital for complex tasks.

Starcloud raises a $250 million Series A extension, hitting a $2.3 billion valuation, to scale orbital AI inference satellites despite tight launch options.

Micro1's gross annual run rate jumped from $100M to $500M in eight months, with net run rate between $150M and $200M amid booming AI training data demand.