Meta uses idle inference capacity to improve ads recommendations
Overview
Meta says it has built a system that uses idle inference capacity across its fleet to improve ads recommendations without ordering new capacity.
The system has two elasticity levers: Rank More Ads (RMA), which evaluates more candidate ads when capacity is available, and Serve Better Models (SBM), which routes a share of requests to a more powerful model.
Meta says the approach protects baseline latency and reliability when the opportunistic capacity is reclaimed. Meta also says it helps people see more relevant ads and improves advertiser outcomes. These claims come from Meta's own Engineering blog post, and no independent figures on the gains are reported.
Written by AI from the articles below · updated Oct 12, 1:52 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- Engineering at MetaOfficialMeta turns idle inference capacity into better ads recommendations
AIMeta says it has built a system that uses idle inference capacity across its fleet to improve ads recommendations without new capacity orders. The system relies on two elasticity levers, Rank More Ads (RMA), which evaluates more candidate ads when capacity is available, and Serve Better Models (SBM), which routes a share of requests to a more powerful model. Meta says the approach protects baseline latency and reliability when opportunistic capacity is reclaimed.
Heat trend
Not enough continuous observations to show a trend yet.