What 50 Mbps can mean in a household
At a sustained 50 Mbps download rate, a 5 GB file has a best-case transfer time of about 13 minutes. This is a calculation, not a prediction: 5 GB × 8,000 Mb per GB ÷ 50 Mbps, using decimal units. Real transfers take longer because of network overhead, server limits and changes in speed. A measured speed on one device also does not equal the capacity left for everyone else.
| Activity | At 50 Mbps | Other factor to check |
|---|---|---|
| One Netflix 4K stream | Meets its 15 Mbps recommendation on paper | Stable Wi-Fi and compatible device |
| Multiple 4K streams | At most 3 by simple division; leave headroom | Shared household traffic and stream quality |
| HD video calls | Download alone cannot determine suitability | Upload, latency, jitter and device |
| 5 GB download | About 13 minutes at full rate | Remote server and test route |
Room for several streams, subject to other devices. A broadband plan is usually advertised by download rate; check its separately advertised upload rate before choosing it for video calls, backup or content creation. For calls, repeat a test while other devices are uploading and compare ping and jitter, not only Mbps.
How to use this tier guide
- Run the test over Ethernet where possible, then on your normal Wi-Fi.
- Compare the result with the advertised download and upload of your current plan; note the time and device.
- Repeat during the hours when you experience trouble. Check whether upload or latency changes while others are active.
- If the wired connection repeatedly underperforms, share dated results with your provider. If only Wi-Fi underperforms, investigate coverage and device limits first.
Examples above use Netflix’s published 4K recommendation and Zoom’s HD call bandwidth guidance. Neither company promises that every household, title or call uses an identical amount. The CRTC 50/10 target describes service availability; it is not a minimum requirement for every activity.
What your result actually measures
The download figure measures traffic from the AI Speed test route to your browser, while upload measures traffic sent to it. Ping is the round-trip response time and jitter describes variation between samples. These values belong to the particular path, device and moment you tested. They are not an official city average or a ranking of the providers listed here. If you want to compare two plans, test on the same device and connection method at comparable times, and record the subscribed tier for each.
Use the CRTC 50/10 target correctly
The CRTC’s universal service objective includes access to at least 50 Mbps download and 10 Mbps upload, along with unlimited data at home and at work. It is a policy target for availability, not a guarantee that every address already qualifies or that a browser test must always return 50/10. If your own wired result repeatedly falls below the subscribed plan, first check the equipment and test conditions, then contact the provider with dated measurements. If you need to know whether 50/10 is offered at an address, consult the official broadband map and the provider’s address checker.
Why a local test can disagree with a plan
A plan speed describes the service tier; a browser result also includes the performance of the Wi-Fi link, computer, router, other household traffic and a test server. The slowest component determines the observed rate. A 1 Gbps access line connected through a 100 Mbps Ethernet port cannot show a gigabit reading on that device. An older Wi-Fi radio can also be the limiting part even when the line and router are capable of more. Start with a wired measurement when practical, then repeat from the room where you normally work.
Run several tests rather than treating one reading as a diagnosis. Try a quiet period and a busy evening, and note whether a VPN, cloud backup or large download was active. If the wired reading is stable but the wireless reading drops sharply in one room, focus on Wi-Fi placement and interference. If the wired reading also varies by time, collect a small series of results before discussing congestion or line issues with support. Different speed testing sites use different routes and methods, so equal numbers are not expected.
Availability is an address-level question
A list of companies in a provincial market does not establish coverage on your street or in your building. Fibre to a neighbourhood, fibre to a building and fibre to an individual unit are different claims. Where cable, DSL, fixed wireless or satellite also exist, their upload and latency characteristics may be different. Check the exact civic address and unit number, verify the plan’s advertised download and upload, and ask whether the installation requires new equipment. The National Broadband Map is a useful starting point; the provider’s current address checker and written offer are needed for a purchase decision.
Location and network ownership
An IP lookup can estimate a city and identify the organization associated with an address range, but it cannot establish the physical location of a subscriber or tell you which provider serves a neighbouring house. Mobile networks, VPNs and shared gateways can make the estimated city less relevant to where a user is sitting. AI Speed asks for a location lookup only when you request network details, and the map labels the coordinates as approximate. Use an address checker for coverage and an on-site speed test for the path you actually use. Treat an autonomous-system name as a network clue rather than proof of the retail brand on a bill.
Provider comparison checklist
Compare plans using the exact address, access technology, advertised upload rate, equipment, price after promotion and any data allowance. A citywide speed ranking would need a representative dataset and a disclosed method, which AI Speed does not yet have.
Match the fix to the bottleneck
For a slow wired result, restart equipment once, check the modem or optical terminal’s status, confirm the subscribed tier and contact the provider if the issue persists. For a fast wired result but poor Wi-Fi, move the router into an open central position, choose suitable bands for distance and device capability, and avoid placing it behind metal or inside a cabinet. For good throughput but high jitter, pause competing uploads and test during another period. For home-office calls, sustained upload and stable latency can matter as much as download. For gaming, low, consistent latency generally matters more than a headline gigabit figure.
How these pages are maintained
AI Speed separates a policy target, an advertised plan and a measured result. The 50/10 target comes from the CRTC. Service availability should be confirmed through Innovation, Science and Economic Development Canada’s National Broadband Map and the provider’s current address checker. This page does not fabricate local averages or suggest a provider is fastest without comparable evidence. Once AI Speed publishes dated, sourced measurements, the city and provider comparison rows can be updated with sample sizes and a clear method. Until then, the most useful number on this page is the result you measure on your own device.
Common questions
Does the CRTC 50/10 target mean every home in Canada has 50/10 service?
No. The CRTC sets a national universal service target. Availability and the subscribed plan vary by address.
Why can a speed test in Canada differ from my plan speed?
The device, Wi-Fi, local traffic, test route, and plan can each constrain a measurement. Compare wired and wireless tests at different times.
Does an IP lookup give an exact location?
No. City and coordinates represent an approximate database association for a network address, not the device’s GPS position.
Sources and method
- CRTC: Broadband Fund and universal service targets
- ISED: National Broadband Map
- AI Speed: why speed tests disagree
Editorial method: illustrative calculations from stated assumptions and linked application guidance; individual results depend on network conditions.