DeepSeek V4 breaks the math on $500B in AI capex; Anthropic cuts Opus 4.7 pricing 67%

DeepSeek V4 Pro (1.6T total parameters, 49B active per token via MoE, 1M token context) launched via API on April 24 and requires only 27% of V3.2's single-token inference FLOPs and 10% of its KV cache at full context. The efficiency gains are pushing down API pricing industry-wide — Anthropic cut Claude Opus 4.7 prices 67% post-V4 launch.