"Last mile" is infrastructure jargon borrowed from telecom, but it's become the accurate shorthand for the hardest part of digital access policy: it's rarely the backbone network that's missing. It's the final stretch — a household connection, a working device, someone available to fix it when it breaks — and that's where national connectivity funds live or die.

Three models, three trade-offs

We compared three currently active national programs, each built around a different core mechanism:

  • Direct subsidy model — households receive a monthly credit toward any ISP of their choice. Fast to deploy, but funds can be absorbed into price increases where ISP competition is weak.
  • Public infrastructure model — government builds and owns last-mile fiber, competitively priced. Slower to deploy, but produces durable capacity independent of any single company.
  • School-anchor model — connectivity is built out from schools as hubs, extending to nearby households. Cheapest per-household cost, but access quality drops off with distance from the anchor site.

What the early data shows

The public infrastructure model produced the most durable gains but took nearly three times longer to show measurable enrollment or completion impact in online courses. The direct subsidy model moved fastest but showed the weakest link to actual learning outcomes — several regions saw connectivity numbers rise without a corresponding rise in course completion, suggesting the subsidy alone wasn't solving the device or support gap.

The piece nobody's funding well

Across all three models, the consistent weak point was device longevity and repair — not the initial connection. Programs that paired connectivity funding with a local repair or device-refresh mechanism saw meaningfully better sustained usage over 18 months than programs that treated the first device handout as a one-time event.

We'll be tracking all three programs' three-year outcomes as that data becomes available.