A data packet can travel from Europe to America in well under a second because it does not travel by satellite, but primarily through fibre-optic cables on the ocean floor. The signal is light guided through glass, and in optical fibre it travels at approximately 200,000 kilometres per second. This means that, although the ocean is enormous on a human scale, it can be crossed digitally in a few tens of milliseconds.
What matters, however, is what exactly “arrives” means. A message, a command in a game or a request to a website can reach an American server very quickly. An entire page, a video or a large file may take longer, because data size, connection capacity, server response and network congestion also come into play.
The real route: from your home router, through Europe, under the Atlantic and into the US network
The internet is neither a single highway nor a single cable linking two continents. It is a network of networks. When a user in Romania opens a service hosted in the United States, the information is split into small packets. They are given instructions about their destination and sent successively through network equipment.
A simplified route looks like this:
- a phone, laptop or television sends the request to the home router or mobile network;
- the internet service provider receives it and carries it over its fibre network to major nodes in Romania or elsewhere in Europe;
- interconnected networks select an available route to an undersea cable landing point;
- the packet crosses the Atlantic through the fibres in the undersea cable;
- on the American shore, the signal enters terrestrial networks and reaches a data centre or the requested server;
- the response travels back along the reverse route, sometimes by a different path.
There is no guarantee that all packets use the same route. Network operators may alter routes according to capacity, outages, costs and interconnection agreements. This distributed structure is precisely what makes the internet more resilient: if one route becomes unavailable, traffic can be redirected, although sometimes with higher latency.
The calculation that explains the speed: 33 milliseconds on a 6,600 km cable
A specific example is the MAREA undersea cable, which links Virginia Beach in the United States with Sopelana, Spain. Its operator, Telxius, gives its length as approximately 6,600 km and its capacity as up to 200 Tbps. It is not the only cable on this route, but it is a useful reference for seeing why the crossing time can be surprisingly short.
Using the approximate speed of 200,000 km/s for an optical signal in fibre gives:
| Calculation | Theoretical result | What it represents |
|---|---|---|
| 6,600 km ÷ 200,000 km/s | 0.033 seconds | about 33 ms, one way |
| 33 ms × 2 | 0.066 seconds | about 66 ms, round trip, through the undersea segment only |
| 1,000 km ÷ 200,000 km/s | 0.005 seconds | approximately 5 ms for every 1,000 km of fibre |
This is a physical-limit calculation, not a speed test that every user will automatically see. It does not include the journey from a home to the core network, terrestrial sections in Europe and the United States, processing in routers, possible detours in the route, or the time taken by the server to process the request.
Still, the calculation shows why the statement that “the internet crosses the Atlantic in under a second” is correct. Even after many tens of milliseconds are added for the rest of the route, there is an enormous margin before the one-second threshold.
Why light does not travel at 300,000 km/s in a cable
In a vacuum, light travels at nearly 300,000 km/s. Optical fibre, however, has a core made of special glass, and the properties of the material reduce the signal’s effective speed to approximately two-thirds of the speed of light in a vacuum. For explanations of undersea cables, the International Telecommunication Union uses the approximation of 200,000 km/s, or 5 milliseconds per 1,000 km.
This limit cannot be removed with a faster subscription. A 1 Gbps subscription can transfer more data per second than a 100 Mbps one, but it cannot make light travel through glass faster. That is why a user can have excellent download speeds and still notice delay in a game hosted on another continent.
Latency, speed and capacity: three different things
In discussions about connections, the term “speed” is used for different phenomena. They need to be distinguished.
- Latency is the travel time of a packet from one point to another. It is usually expressed in milliseconds.
- RTT, or round-trip time, measures how long it takes for a request to reach its destination and for the response to return. This is close to what the ping command observes.
- Bandwidth shows how much information can be carried in a given period of time; it is generally expressed in Mbps or Gbps.
- A cable’s capacity is the total capacity available on that infrastructure, shared among numerous customers and services. It is not the speed available to a single user.
The MAREA cable may have a capacity of up to 200 Tbps, but that does not mean a laptop in Romania downloads at 200 Tbps. It is the transport capacity of the system, intended for the aggregated traffic of the networks and services that use that infrastructure.
In the same context, see also Orionids 2026: why “shooting stars” are not stars and when it is worth watching the sky from Romania, which presents additional relevant information.
On the other hand, a small request—such as pressing a button in an app—can cross oceans very quickly even if it does not download a large amount of data. For video calls, online gaming, remote control and transactions between systems, latency is often more important than maximum download speed.
Why a Europe–US ping is not 66 ms
The approximately 66 ms calculated for MAREA represents only the round trip through the 6,600 km undersea segment, in an idealised scenario. A test run from home to an American server measures much more than that.
The total delay can include:
- the connection between the device and the router or mobile antenna;
- the internet provider’s access segment;
- travel through European terrestrial networks to the landing point;
- equipment that selects and switches the route;
- the Atlantic crossing;
- the route through the US network to the data centre;
- temporary queues in equipment when traffic is heavy;
- any time taken by the application or server to prepare the response.
There is another essential difference: a map shows straight-line distance, but fibre must follow a practical and safe route. Cables bypass certain difficult areas, reach existing landing points and then connect onwards to data centres. As a result, the distance travelled by the signal is generally greater than the geometric distance between two cities.
Why satellites are not usually the main route across the ocean
It is tempting to imagine data jumping from Europe to a satellite and then descending into America. In practice, undersea cables are the essential solution for most data traffic between continents. They offer enormous capacity and, on transatlantic routes, generally have a latency advantage over communications that must travel up to and down from satellites far above Earth.
Satellites have important roles, especially for areas that are difficult to connect by cable, mobile connections, aviation, ships or situations where terrestrial infrastructure is absent. But a typical online service used from Romania to access an American data centre is most likely based on a combination of terrestrial and undersea fibre.
What CDNs do: bringing America closer without moving the continent
Many websites and video platforms do not send every image, file or video clip directly from their original server in the United States. They use a content delivery network, known as a CDN. It keeps copies of frequently requested content in data centres closer to users.
As a result, a user in Romania may receive a website’s images and static files from a European node, even if the company that owns the service is American. The Atlantic crossing may happen in the background, between the original server and the CDN infrastructure, but not for every request from every visitor.
This is one explanation for why an American website or platform can feel fast from Europe. It does not mean the oceans have become smaller; it means that the most frequently requested data has been moved closer to the user. CDN networks reduce distance, and reducing distance lowers latency.
Why the same app can sometimes feel fast and other times slow
A very fast cable alone cannot guarantee a perfect experience. If the server is overloaded, the local network is congested, Wi-Fi has problems or the selected route is longer than usual, the app may respond more slowly. In addition, a modern web page does not send a single request: it may simultaneously request text, images, fonts, scripts, advertisements, analytics elements and data from multiple services.
Therefore, “under a second” accurately describes a signal’s ability to cross oceans, but it is not a promise that every page will load in one second. The time perceived by the user is the result of the entire chain: device, local network, provider, international route, server and application.
The essential idea
The internet between Europe and America works quickly not because the distance is short, but because information is converted into light and transported through a global fibre-optic infrastructure. Over a crossing of about 6,600 km, physics imposes approximately 33 ms in one direction. That is why, even after the real network stages are added, a message can travel across the Atlantic and back many times before one second passes.
Frequently asked questions
Does internet traffic between Europe and America go via satellite?
In most situations, the main traffic between continents travels through undersea fibre-optic cables. Satellites are useful in certain contexts, but they are not the usual route for most transatlantic traffic.
How long does it theoretically take to cross the Atlantic by optical fibre?
At approximately 200,000 km/s in fibre, a 6,600 km route has a propagation delay of about 33 milliseconds in one direction. The real time is greater because it also includes terrestrial networks, equipment and the server response.
Why does an American website sometimes load quickly from Romania?
Large services often use CDNs, which keep copies of content in European data centres. As a result, many resources no longer need to cross the Atlantic for every visit.
Does a connection with more Mbps automatically reduce ping to the US?
Not necessarily. Mbps or Gbps mainly describe the volume of data that can be transferred. Ping depends chiefly on distance, route, equipment and congestion.
