Starlink's laser system is beaming 42 petabytes of data per day

Starlink’s satellite network is now moving an estimated 42 petabytes per day over inter-satellite laser links, enabling 100 Gbps optical connections between spacecraft thousands of kilometers apart. Commenters dig into how such links work technically—tracking, Doppler effects, alignment, and latency versus fiber—while also probing what the aggregate capacity means in “real world” terms (household traffic, global share of internet data) and how much of that capacity is actually used. Many see the growing laser mesh as both a practical way to improve backhaul and latency-sensitive routes, and a potential strategic communications asset that could complement or back up undersea cables.

Inter-satellite laser links & tracking

  • Discussion focuses on how hard it is to keep 100 Gbps optical links aligned between fast-moving LEO satellites, especially across orbital planes.
  • In-plane links are easier; cross-plane links require tighter tracking, look-ahead pointing (aiming where the other satellite will be), and complex guidance/navigation/control.
  • Doppler shift is a serious issue for RF cross-links but considered less problematic for simple intensity-modulated lasers.
  • Acquisition and tracking methods (beam widening, raster search, closed-loop fine pointing) are described as “secret sauce” and major engineering effort.

Use of lasers vs RF and ground links

  • Current lasers are for satellite–satellite links; satellite–ground is still RF, though there are experiments with atmospheric-tolerant optical downlinks.
  • Clouds and weather are a concern for any future laser-to-ground; routing around bad links via other sats and falling back to RF are suggested.
  • Many users are still likely served via simple “ground–sat–ground” paths; the quoted 42 PB/day is only inter-satellite laser traffic and likely includes multi-hop counting.

Traffic volume and per-customer view

  • 42 PB/day ≈ 3.9 Tbps aggregate, or ~486 GB/s.
  • With 2.3M customers, back-of-envelope gives ~1.7 Mbps average per customer (550 GB/month), which commenters say is plausible for heavy video use.
  • Each laser can do 100 Gbps, so average utilization (0.4%) is low; links are up and down as geometry changes, with ~266k “acquisitions” per day and an inferred average link lifetime of ~50 minutes.

Latency and comparison to fiber/HF

  • Several note light in vacuum/air is ~50% faster than in conventional fiber, so long Starlink laser paths can beat subsea cables for some routes, though hop count and routing dominate real-world latency.
  • HF radio can still win in raw path length for certain financial-trading scenarios, but has low bandwidth and reliability issues.

Reliability and user experience

  • Anecdotal reports: 40–300 Mbps down, 40–80 ms latency, generally fine for gaming and streaming.
  • Some mention brief outages and rare multi-hour events; overall judged “good but not telco-grade,” with occasional DHCP / IP reassignment quirks when bypassing the official router.

Visibility, deorbiting, and safety

  • Many describe seeing Starlink trains and individual sats; visibility depends heavily on light pollution and timing near dusk/dawn.
  • Deorbiting Starlinks are designed to fully burn up; odds of debris reaching the ground or harming people are described as extremely low.
  • Risk of rockets intersecting laser beams is considered negligible; any transient link loss is handled by normal network protocols.

Strategic and broader implications

  • Commenters see a second, space-based backbone as a future complement/resilience layer to undersea cables, with clear military value (resilient comms, drone links).
  • Questions raised about reliance on a private operator; several note governments can commandeer or heavily direct such infrastructure in crises.