bottleneckScore 55/100Research
AI factory power infrastructure overprovisioning by 40% creates massive efficiency gap; DSX Max LPS targets cost recovery through dynamic power allocation
Jensen Huang· NVIDIA· AI· 2026-06-02· about NVIDIA (NVDA)
“Today's AI factories over provision power by up to 40%. DSX Max LPS lets operators safely deploy more GPUs inside the same power budget, adding billions in annual revenue.”
Why it matters
Current AI factory deployments are wasting 40% of their power budget, representing a critical infrastructure inefficiency. This gap signals that power delivery, cooling, and electrical infrastructure are becoming the binding constraint on AI factory economics, not GPU availability.
Investment implication
Companies supplying power distribution units (PDUs), liquid cooling systems, power smoothing electronics, and grid management software will see accelerated capex as operators retrofit existing facilities. NVIDIA's DSX software stack is positioned to capture margin on software-driven efficiency gains, but suppliers of physical power infrastructure components face near-term demand surge.