NVIDIA announces Blackwell Ultra — 2x inference speed for enterprise AI

Next-gen GPU architecture halves inference cost. Cloud providers already placing bulk orders for Q3 deployment.

Why it matters: Cheaper inference means AI features become viable for small and mid-size businesses.