Starlink’s U.S. median download speed reached 128 Mbps in early 2026 — yet millions of subscribers open a speed test and see something far lower. The gap is real, it has specific causes, and the causes are rarely the same for any two households. This guide separates them one by one.
🔍 Run a Diagnostic Before Anything Else
A speed test result with no context doesn’t tell you much. Run the in-app Starlink speed test first — it tests the satellite link directly and removes your home network as a variable. Then test from a device plugged in via Ethernet (requires the $25 Starlink Ethernet adapter). If the wired result is much faster, the problem is your Wi-Fi, not Starlink. If both are slow, the bottleneck is the satellite connection itself. Tap to run each test.
📌 Note: Run three tests at different times of day — morning, midday, and evening. A single result misses the pattern. The gap between your 3 AM result and your 8 PM result is your congestion profile.
The questions people ask most — answered plainly, without the runaround.
Normal — and one of the most common complaints. Starlink’s bandwidth is shared among all subscribers in your local coverage cell. Evenings between roughly 5 PM and 11 PM are peak usage hours when households simultaneously stream video, game, and browse. The result is real, measurable congestion that slows everyone in the same cell proportionally. This is not a hardware issue, not an obstruction issue, and not something a power cycle fixes. If your 3 AM speed test shows 150 Mbps and your 8 PM test shows 35 Mbps, congestion is the diagnosis. The only fixes are upgrading to Priority data (which is deprioritized last during congestion), waiting it out, or shifting high-bandwidth tasks to off-peak hours.
Open the Starlink app, go to Visibility, and wait 12 to 24 hours for the system to accumulate enough satellite-pass data to draw a real obstruction map. Red markings on the map are hard obstructions — buildings, chimneys, solid structures. Orange or amber areas are soft obstructions — trees, branches, leafy canopy. Even a seemingly small orange zone causes real problems, because it creates periodic signal interruptions as satellites pass through that part of the sky. The app also shows total outage time attributed to obstructions under Statistics. An obstruction rating above 5% typically causes noticeable performance issues; above 10% and the dish is in a genuinely poor location that mount adjustments should address.
High download speed doesn’t guarantee smooth performance for real-time applications. Three separate problems produce this pattern. First, upload speed — video calls use upstream bandwidth heavily, and Starlink’s upload is asymmetric (typically 10–25 Mbps, occasionally lower during congestion). If upload is consistently under 5 Mbps, video calls break down regardless of download speed. Second, latency spikes — intermittent obstructions cause the dish to switch satellites mid-session, creating 1–3 second gaps that feel like a freeze. Third, Wi-Fi interference inside the house can cause high packet loss that a speed test averages out but real-time applications feel acutely. Run a ping test to 8.8.8.8 during a stutter — if you see latency spikes above 200 ms, obstructions or satellite handoff is the culprit.
Light rain and overcast skies cause virtually no measurable effect. The problem weather is heavy rain — the kind that produces more than half an inch per hour. At that intensity, water droplets absorb and scatter Starlink’s Ka-band signal (a phenomenon called rain fade), causing speed drops of 30–50% and occasional brief outages during the peak downpour. The key difference from older satellite systems: Starlink recovers within minutes once the rain eases, because the dish communicates with multiple satellites simultaneously and switches to a cleaner orbital path as conditions shift. If you’re seeing slow speeds every time it rains hard — and only then — weather is the diagnosis and the only fix is waiting it out. Speeds return automatically.
This is almost always a local Wi-Fi problem — not a Starlink problem. The satellite connection delivers the same signal to your router regardless of which device you’re using. Speed differences between devices inside the house come from: the device’s Wi-Fi standard (an older laptop using 802.11n will be slower than a phone using Wi-Fi 6 even on the same router); distance and walls between the device and the router; interference from neighboring Wi-Fi networks on the same channel; or that specific device running a background update or upload that’s consuming bandwidth. Test both devices on the same 5 GHz band, from the same physical location close to the router. If they still differ significantly, the problem is the device itself — its Wi-Fi card, its drivers, or something running in the background.
Yes — by design. Starlink’s plan tiers have explicit prioritization built in. Standard Residential plan subscribers are served first. Roam and RV plan subscribers are deprioritized behind residential users — meaning that during any congestion event, Roam customers experience the speed reduction first, and it’s typically more severe. This is not a bug or a billing error — it’s disclosed in Starlink’s terms of service. The practical effect: a Roam subscriber in a cell with heavy residential usage may see 20–40 Mbps during peak hours while a Residential subscriber in the same cell sees 80–100 Mbps. Upgrading to a Residential plan at a fixed address, or purchasing Priority data, are the only plan-level fixes.
The honest number: most U.S. Starlink households see 60–130 Mbps download during daytime hours and 30–80 Mbps during peak evening hours. Ookla’s independently measured U.S. median for 2025 came in at approximately 117–128 Mbps overall — but that’s a median across all hours, not a floor. About 20% of users consistently pull under 40 Mbps. Nebraska recorded the highest state median at 200 Mbps; Alaska was the only state where the median didn’t reach 100 Mbps. The FCC’s minimum broadband threshold is 100 Mbps download and 20 Mbps upload. By end of 2025, roughly 45% of Starlink users met that standard in both directions — up sharply from just 17% at the start of the year, driven by SpaceX’s V2 satellite deployment increasing per-cell capacity.
Find your pattern in the table below, then scroll to that section for the full diagnosis and fix. If you’re not sure which symptom matches, check the “When it happens” column first.
| Cause | When It Happens | Typical Speed Impact | DIY Fix Available? | Fix Time |
|---|---|---|---|---|
| Cell congestion | Every evening, 5–11 PM | 30–60% drop from daytime | Partial — shift tasks or upgrade plan | Ongoing |
| Physical obstructions | Random brief dropouts all day | Intermittent outages + slower average | Yes — reposition dish or trim | Hours–days |
| Wi-Fi interference/range | Some devices slower than others | Varies by device and location | Yes — reposition router or add mesh | Minutes |
| Heavy rain / rain fade | During downpours only | 30–50% temporary drop | No — wait it out | Self-resolves |
| Snow on dish face | Snowstorms, cold weather | Full outage possible | Yes — dish self-heats; wait | Self-resolves |
| Damaged or kinked cable | Constant, no pattern | Consistent underperformance | Yes — inspect and replace | Hours |
| Router placement/interference | All the time · whole house slow | 15–40% below full speed | Yes — reposition router | Minutes |
| Old firmware / router reboot needed | Sudden change after update window | Inconsistent · worse than before | Yes — power cycle or wait | Minutes |
| Plan deprioritization (Roam/RV) | Peak hours · congested areas | Often 20–50 Mbps vs. 80–100 for Residential | Partial — upgrade plan | Account change |
| Background device usage | Specific times · specific devices | Consumes full available bandwidth | Yes — identify and pause | Minutes |
| Hardware fault (dish or router) | Constant · no improvement from fixes | Consistently well below normal | No — contact Starlink support | Replacement needed |
Starlink doesn’t give each subscriber a dedicated lane — the bandwidth in your local satellite coverage cell is shared among everyone using it at the same time. When demand spikes, performance drops for all of them equally.
A single Starlink satellite beam covers roughly 15–60 square miles of ground. Every subscriber in that footprint shares the same bandwidth pool — and when residential internet usage peaks between 5 PM and 11 PM, that pool gets divided among more users simultaneously. The result is a predictable slowdown pattern: fast in the morning, fast overnight, noticeably slower in the evening. This is not a malfunction. It’s a structural characteristic of shared wireless bandwidth, and it affects cable providers the same way in dense neighborhoods. The diagnostic test: run a speed test at 2 AM and again at 8 PM. If the 2 AM result is dramatically faster — 150 Mbps versus 40 Mbps — congestion is confirmed. Your hardware, cable, and dish placement are all fine.
As Starlink adds subscribers in your area, congestion in your cell increases. Speeds that were excellent when you were one of a handful of subscribers in your cell can degrade over months as more households sign up. This is a documented pattern from Starlink’s early growth phase — U.S. median speeds dropped from ~90 Mbps in 2021 to ~65 Mbps in late 2022 as subscriber growth outpaced satellite capacity. SpaceX’s V2 satellite deployment reversed this by adding substantially more per-cell bandwidth, which drove the U.S. median back up past 100 Mbps through 2025. If your speed was good six months ago and has gradually declined, this is worth knowing — but the only tools available to subscribers are plan upgrades or contacting Starlink to document the degradation.
Schedule bandwidth-heavy tasks for off-peak hours — large file downloads, cloud backups, operating system updates, and game downloads. Set these to run after midnight when your cell has spare capacity. For work-from-home users who need consistent speeds during business hours, Starlink’s Priority plan reserves dedicated throughput that gets served first during congestion events. The Priority plan costs more monthly but delivers more consistent speeds when the cell is loaded. For RV or Roam subscribers, upgrading to a Residential plan at a fixed address eliminates the deprioritization layer entirely. No amount of rebooting, mount adjustment, or router changes resolves a pure congestion problem — the bottleneck is the satellite cell, not anything in your home.
Trees are the number one hardware installation problem in Starlink deployments. A single branch clipping the edge of the dish’s sky view creates brief but repeated signal interruptions that speed tests average into a lower number and real-time applications feel as freezes.
Starlink’s dish needs to track satellites across a broad arc of sky — not just straight overhead. Anything that blocks a portion of that arc causes the dish to lose that satellite contact for the duration of the pass. With multiple obstructions, the dish experiences repeated brief interruptions throughout the day that don’t appear as a single outage but accumulate as lower average speed and occasional freezing during video calls. Trees are the trickiest obstacle because they’re soft obstructions — the signal passes through partially, creating intermittent rather than complete signal loss, and the effect worsens when leaves are wet. Seasonal changes matter: a tree that caused minimal obstruction in winter may create a significant problem once leaves fill in during spring. Check your obstruction map in both seasons.
The obstruction fixes, from easiest to most involved. First: use the Starlink app’s obstruction checker from different locations on your property to find a cleaner position. Tie your phone to an extended pole to simulate rooftop height before committing to a mount location. Second: trim branches that are entering the red or orange zones on your obstruction map — even trimming one or two offending limbs can eliminate the problem. Third: raise the mount — a higher pole or roof peak position clears canopy that a wall mount cannot. Fourth: use a standalone pole mount in an open area of the yard when no roof position can achieve adequate sky view. The guiding rule: spend money on height before anything else. A 10% obstruction score causes more actual performance damage than any router limitation.
This is the most commonly misdiagnosed slow-Starlink complaint. The satellite connection is delivering full speed to the router — but the Wi-Fi signal between the router and your device is the bottleneck. You blame Starlink, but the problem is 15 feet away on a shelf.
The Starlink Gen 3 router broadcasts on both 2.4 GHz and 5 GHz bands. The 5 GHz band is faster but shorter range — it loses signal quickly through walls, floors, and furniture. The 2.4 GHz band reaches further but has less bandwidth. A router placed in a corner, inside a cabinet, on the floor, or near a microwave oven is not serving your home well. The fix costs nothing: move the router to a central, elevated position — ideally on a shelf at waist height or higher in the room you use most. If you’re testing from a bedroom two rooms away and seeing slow speeds, connect your phone or laptop right next to the router and run the test again. If speeds jump significantly, the router’s placement or your home’s layout is the limit — not Starlink.
The Starlink Gen 3 router covers a typical single-story home reasonably well in open layouts. In larger homes, multi-story houses, or homes with thick concrete, brick, or plaster walls, the signal degrades significantly in far rooms. Starlink offers a mesh node ($130, or included free with the Residential plan) that connects wirelessly to extend coverage — place it roughly halfway between the router and the dead zone, still within a clear line of sight to the main router. For the best possible performance, connecting the mesh node via Ethernet cable to the main router (rather than wirelessly) doubles the backhaul capacity between them. If your slow device is in a room the router signal barely reaches, the mesh node usually resolves the problem completely.
A single device running a large background update — a game console downloading a multi-gigabyte patch, a PC updating Windows, a security camera system continuously uploading video to cloud storage, a phone syncing photos — can consume most of your available bandwidth without any obvious indication. The household speed you experience is the total connection minus whatever background processes are actively running. The fastest diagnostic: open your router’s device list in the Starlink app and look for any device showing high data usage. Pause or schedule large downloads to overnight hours. If a security camera system is your culprit, consider setting it to record locally rather than streaming continuously to the cloud during the day.
Starlink handles weather dramatically better than legacy geostationary satellite internet — because LEO satellites orbit at 340 miles rather than 22,000 miles, the signal travels through far less atmosphere. Most weather has no meaningful effect. The exceptions are specific and worth knowing.
Starlink’s Ka-band radio frequencies are partially absorbed by water droplets in the air — a well-documented physics phenomenon called rain fade. The severity depends on rainfall intensity, not just whether it’s raining. Light rain (under 0.1 inches per hour) has no noticeable effect. Moderate rain causes minor reductions. Heavy downpours above 0.5 inches per hour can drop speeds by 30–50% and cause brief 10–30 second outages during the most intense periods. Starlink’s LEO design partially compensates because the dish tracks multiple satellites simultaneously — when one satellite path is degraded, the phased array switches to a cleaner angle through a different satellite. The practical result: Starlink recovers from rain events much faster than older satellite services. Full performance typically returns within minutes of the heavy rain easing.
The Starlink dish has a built-in electric heater that activates automatically when the surface temperature drops to near freezing, designed to melt snow and ice off the face of the dish. For light, dry, fluffy snow with low moisture content — the kind common in cold interior climates — the heater handles accumulation quickly and performance is largely unaffected. For heavy, wet, dense snow — the kind that falls near freezing — accumulation can temporarily overwhelm the heater and cause outages until the snow either melts or is cleared. Do not pour hot water on the dish to clear snow — the thermal shock can damage the radome surface. A soft brush or waiting for the heater to work are the correct approaches. The heater draws extra power during cold weather, which is why Starlink power consumption is higher in winter.
Overcast skies, fog, mist, and cloud cover cause virtually no effect on Starlink — independent measurements in persistently cloudy climates like the Pacific Northwest and New England show under 3% signal degradation during extended overcast periods with no precipitation. Wind itself doesn’t interfere with the radio signal either — only a physical dish displacement from extreme wind (above the dish’s 95 mph rated threshold) would cause a problem. Cold temperatures, frost, and ice are handled automatically by the dish heater. If your connection is slow on a cloudy day with no rain, weather is not the cause — look at congestion, obstructions, or your home network instead.
Hardware problems are less common than congestion or obstruction issues, but they produce a specific pattern: consistent underperformance that doesn’t vary by time of day, doesn’t improve after a power cycle, and doesn’t change with weather.
The proprietary cable that runs from the dish to the router carries both power and data through a single run. A kink, a sharp bend, a crush point from where it was routed under a door edge, animal chewing, or UV degradation over time can partially compromise the cable’s signal integrity — enough to reduce performance consistently without causing a complete outage. This is deceptive because it shows up as slow speeds rather than a clear disconnection, which makes it easy to mistake for congestion or obstruction. Walk the entire cable run from dish to router and inspect every inch — look specifically at points where the cable bends around corners, where it enters the building, and anywhere it may have been compressed or stepped on. Any bend tighter than an 8-inch radius is a suspect point. Replacement cables for the Gen 3 system cost approximately $35 from Starlink.
Starlink pushes firmware updates to the dish and router automatically, typically between 2 AM and 4 AM local time with an automatic reboot afterward. Occasionally a firmware update causes a performance regression that clears with a full cold boot — unplugging the router from the wall for 30 full seconds rather than using the software reboot option in the app. The difference matters: a soft reboot flushes some system state but leaves others in place; a cold boot resets the routing buffers and forces the phased array to acquire a fresh satellite track. If your speed suddenly degraded after a reboot window and others in your area report the same thing on Starlink community forums, wait — SpaceX typically pushes a hotfix within a few days for firmware regressions, and it auto-installs overnight.
If you have cleared obstructions, confirmed the cable is undamaged, power cycled the system, verified the issue is not congestion by testing at 3 AM, and speeds are still consistently below 25 Mbps at all hours — the dish or router may have a hardware defect. Starlink has no phone support. All hardware warranty claims and support requests go through the Starlink app’s ticket system. Submit a ticket with your speed test logs attached — include results from multiple times of day and the app’s Statistics export. Response times range from hours to days depending on ticket volume. Equipment replaced under warranty ships directly from SpaceX at no charge for defects within the warranty period.
📱 Contact: Starlink app → Settings → Support → Submit a Request
Starlink’s plan tiers prioritize bandwidth allocation during congestion in this order: Priority plan subscribers are served first; Standard Residential subscribers are served second; Roam and RV subscribers are served last. This is disclosed in Starlink’s terms of service and is intentional — Roam plans are priced lower and designed for flexible use, while Residential plans pay for consistent-location service. During a congested evening cell, a Roam subscriber may see 20–40 Mbps while a Residential subscriber in the same cell sees 80–100 Mbps — not because anything is malfunctioning, but because the prioritization queue is working as designed. The fix is a plan change, not a troubleshooting step. If you have a fixed address and are on a Roam plan to save money, switching to a Residential plan resolves this class of slow speed completely.
This pattern is congestion, not a hardware or installation problem. The satellite cell you’re in is under higher load between 5 PM and 11 PM as neighboring Starlink subscribers stream and browse simultaneously. Nothing on your end is broken. The options available: schedule large downloads and backups to run overnight (set your PC, game console, and phone to update automatically after midnight). If your work requires consistent speeds during business hours, Starlink’s Priority plan costs more monthly but is deprioritized last during congestion — meaning your speeds hold up better when the cell is under load. Contacting Starlink support to document persistent cell congestion is also worth doing — SpaceX uses subscriber-reported data to prioritize satellite capacity additions.
A speed test showing 80 Mbps does not guarantee smooth streaming — because speed tests measure bulk download capacity, while streaming requires consistent throughput plus low latency plus low packet loss simultaneously. Three culprits produce this pattern: (1) Intermittent obstructions — even a 5% obstruction score creates brief signal interruptions that a speed test averages away but streaming buffers don’t handle gracefully. Check your Starlink app’s Statistics for obstruction-related outage time. (2) Upload saturation — if another device is uploading large files (cloud backup, security camera) while you stream, the saturated upload path degrades all traffic, including downloads. (3) Wi-Fi signal quality — a weak Wi-Fi signal between your router and your TV or streaming device causes packet loss that interrupts playback even when average speed looks acceptable. Try connecting your streaming device via Ethernet (requires the Starlink Ethernet adapter and a switch or cable). Wired connections eliminate the Wi-Fi variable entirely.
Video calls are sensitive to upload speed, latency, and packet loss in a way that casual browsing isn’t. The first check: run a speed test during a call dropout and note the upload speed and ping. If upload is under 5 Mbps or ping exceeds 100 ms, that’s your problem. Starlink’s typical upload is 10–25 Mbps, but during evening congestion it can drop below 5 Mbps — which breaks video calls regardless of download speed. Fixes in order of effort: (1) Schedule video calls before 5 PM when possible, avoiding the peak congestion window. (2) Pause any cloud backups, file syncing, or large downloads that run during working hours. (3) Connect your work computer via Ethernet rather than Wi-Fi — wired connections have lower latency and more consistent throughput. (4) If congestion is the consistent root cause, the Priority plan offers more stable upload during peak hours. Starlink has no phone support — all service issues go through the app ticket system.
RV and Roam plan subscribers are deprioritized below Standard Residential customers during any congestion event — this is the primary reason speeds feel lower than at a fixed address, especially when parked in a populated campground where multiple Starlink subscribers may share the same coverage cell. Additionally, your dish placement at a campsite may have obstructions that your home installation doesn’t — trees, neighboring RVs, or a parked position with limited sky view. Run the obstruction check from your parking spot before concluding the network is the problem. If obstructions are minimal and speeds are still significantly lower than expected, congestion in the local cell and plan deprioritization are the combined explanation. Upgrade to a Residential plan if you primarily use Starlink from a fixed location — the Roam plan’s flexibility comes at a real performance cost in busy cells.
If you have: confirmed the obstruction rating is under 5%, verified the cable has no damage, run a speed test at 3 AM and it’s still under 30 Mbps, confirmed you’re on a Residential plan, and power cycled the system — then the evidence points toward either a hardware fault or a cell-level network problem in your area. Contact Starlink support through the app with documented speed test logs (multiple tests at different times of day, including off-peak hours). Include screenshots from the Statistics tab showing your obstruction and downtime data. Starlink support responds via the app ticket system with typical response times of several hours to two days. If the issue is confirmed as a hardware defect within warranty, SpaceX ships replacement hardware directly. If the issue is cell-level congestion that SpaceX needs to address through satellite capacity, document it and monitor for improvement as new V2 satellites are added to your coverage area.
📱 Starlink app → Settings → Support → Submit a Request — attach speed test screenshots when filing.
This guide is for general informational purposes only. Starlink speed performance varies by location, plan tier, time of day, local cell congestion, and satellite constellation density in your coverage area. Speed figures reflect independent Ookla Speedtest measurements and publicly reported FCC data and are not guaranteed by SpaceX or Starlink. Plan pricing, prioritization rules, and hardware specifications change without notice — verify current terms at starlink.com before making any plan or equipment decisions. Starlink provides no phone support; all service requests and hardware warranty claims must be submitted through the Starlink app. This content is entirely original.