Daily field briefing
POLR AIPOLR AI

AI + AEC Today

Daily AI intelligence for the built world

Weekly digest

Weekly AI in AEC stories

← Wednesday, June 17, 2026

AI & Business · Jun 4

Low

AI inference costs fell up to 10x on Nvidia's Blackwell

New chips and open models are making AI much cheaper to run. VentureBeat reports the cost per use fell by as much as ten times on Nvidia's newest hardware. Some teams cut their AI bills by around 90 percent.

Why it matters to GCs

When AI costs less to run, the software you already use can add AI tools without a big price jump. Expect more AI in your current project and estimating apps.

VentureBeat — AI·Via POLR AI Hub