PC Gaming Internet Connection Requirements: Bandwidth and Latency

Network performance shapes the playability of online PC games more directly than most hardware specifications. Bandwidth determines how much data moves between a player's machine and game servers per unit of time, while latency determines how quickly that data arrives — and the two metrics impose distinct failure modes on different game genres. This page covers the technical definitions of both metrics, how they interact with game server architecture, the thresholds that separate functional from dysfunctional connections across common gaming scenarios, and how to reason about connection type trade-offs. For a foundational understanding of how PC hardware and software systems interact as a whole, see How PC Gaming Works: Conceptual Overview.

Definition and scope

Bandwidth refers to the maximum data transfer rate of a network connection, typically measured in megabits per second (Mbps). In the context of PC gaming, bandwidth governs how quickly game clients can download patch data, stream texture assets, and exchange state information with multiplayer servers during active sessions.

Latency is the round-trip time (RTT) for a data packet to travel from the player's machine to a remote server and back, measured in milliseconds (ms). In-game displays typically express this as "ping." High latency introduces a lag between player input and server acknowledgment — a failure mode that is operationally distinct from low-bandwidth problems.

Jitter is the variance in latency across successive packets. A connection with 40 ms average latency but ±30 ms jitter performs worse for competitive play than a connection with 60 ms average latency and ±2 ms jitter, because inconsistent packet timing makes hit registration and movement interpolation unreliable.

The Federal Communications Commission (FCC) defines broadband as a minimum 25 Mbps download and 3 Mbps upload — thresholds that are adequate for general internet use but were not established with low-latency gaming traffic as the primary design criterion.

The scope of this reference covers wired and wireless residential broadband connections used for PC gaming within the United States. Cellular (4G/5G), satellite, and fiber connections each impose structurally different latency floors and are addressed in the scenario breakdowns below.

How it works

Online PC games communicate through a client-server model or, in some architectures, a peer-to-peer model. In the dominant client-server model, a dedicated server maintains the authoritative game state. Each connected client sends input packets (keypresses, mouse movements, controller data) to the server and receives state update packets in return.

The data volume per session is relatively low by modern internet standards. A multiplayer first-person shooter typically consumes between 40 and 150 Mbps during active play — but that figure is less critical than the consistency of delivery (Riot Games Engineering, published technical documentation on their VALORANT network architecture). Game engines use techniques including:

1. Delta compression — transmitting only the changes in game state rather than the full state each tick
2. Client-side prediction — the local machine simulates probable outcomes of player input before server confirmation arrives
3. Lag compensation — the server rewinds its state model to adjudicate hit registration based on the latency of the shooting client
4. Interpolation — smoothing the rendered positions of remote players between received state updates

These techniques reduce the perceived impact of latency up to a threshold. Above approximately 100–150 ms RTT, lag compensation and interpolation degrade to the point where gameplay becomes noticeably inconsistent even with compensation active. Competitive titles with tick rates of 64 Hz (standard) or 128 Hz (high-fidelity servers used by titles like CS2) are more sensitive to latency variance because state updates arrive more frequently and must be processed with tighter timing tolerances.

Common scenarios

Wired Ethernet (cable/fiber, <20 ms to regional server): The reference baseline for competitive PC gaming. Wired connections eliminate the radio contention and interference that introduce jitter in Wi-Fi environments. Fiber-to-the-premises (FTTP) services in major US metros can achieve sub-10 ms latency to regional game server infrastructure. For online multiplayer PC gaming, wired Ethernet is the professional and semi-professional standard.

Wi-Fi 6 (802.11ax) on 5 GHz band: Latency overhead compared to wired is typically 2–8 ms under ideal conditions, but interference from neighboring networks and physical obstructions can push jitter into ranges that affect competitive play. Adequate for casual and cooperative genres; not optimal for sub-100 ms reaction-dependent titles.

Cable broadband (DOCSIS 3.1): Widely available in US urban and suburban markets. Latency to regional servers typically runs 15–40 ms. Bandwidth is asymmetric — download capacity significantly exceeds upload — which matters for streamers simultaneously uploading encoded video (discussed in PC Gaming Content Creation and Streaming).

Satellite (geostationary, e.g., legacy HughesNet/Viasat): Minimum latency of approximately 600 ms due to the ~35,786 km orbital altitude. Functionally incompatible with real-time multiplayer gaming. Satellite services using low Earth orbit (LEO) constellations achieve 20–60 ms latency per published service specifications from providers operating at altitudes of 340–550 km, making them viable for casual multiplayer in low-density rural areas.

4G LTE mobile broadband: Median latency in the 30–70 ms range per FCC Measuring Broadband America reports (FCC MBA Program). Adequate for turn-based and casual genres; jitter under congestion makes it unreliable for competitive real-time titles.

Decision boundaries

The following thresholds represent the functional boundaries commonly applied by game developers and network engineers when classifying connection quality for real-time multiplayer:

Bandwidth vs. latency as competing constraints: High bandwidth does not compensate for high latency. A 1 Gbps fiber connection routed through a congested or geographically distant server path will produce higher ping than a 50 Mbps cable connection routed through a nearby regional server. Server geographic proximity is the primary determinant of latency, not connection tier within normal residential bandwidth ranges.

Upload bandwidth as a limiting constraint for streaming: A player running PC gaming content creation and streaming workflows concurrently with active multiplayer sessions requires stable upload capacity. Encoding at 6,000 kbps (a standard Twitch recommended setting per Twitch Broadcasting Guidelines) demands sustained upload headroom above game traffic.

Network setup optimization: Physical layer decisions — cable grade, router placement, QoS (Quality of Service) configuration — directly affect realized latency and jitter independent of subscribed service tier. A structured approach to these decisions is documented at PC Gaming Network Setup for Gamers.

Connection type vs. hardware bottlenecks: Network performance constraints are separate from the hardware performance constraints documented across the pcgamingauthority.com reference set. A connection producing acceptable latency still requires the CPU, GPU, and RAM to sustain frame rates sufficient to render server-delivered state updates at the engine's tick rate.

References

• Federal Communications Commission — Broadband Speed Guide
• FCC Measuring Broadband America Program
• Riot Games Engineering — VALORANT Networking 101
• Twitch Broadcasting Guidelines
• Entertainment Software Association — Essential Facts About the US Video Game Industry
• Internet Engineering Task Force (IETF) — RFC 793, Transmission Control Protocol
• Internet Engineering Task Force (IETF) — RFC 768, User Datagram Protocol