8-camera Frigate NVR build

Frigate Hardware for 8 Cameras: CPU, AI and Storage

Eight cameras are where Frigate hardware planning becomes meaningfully different from a tiny home setup. More feeds increase decode work, several cameras are more likely to be active at once, storage retention grows quickly, and enrichments can push memory usage higher. The right eight-camera NVR is still often compact, but it should have enough detector and memory headroom to avoid rebuilding the system when one or two cameras are added.

See live hardware Size my workload

Quick answer

What this Frigate workload needs

For eight cameras, 16GB RAM is a strong default because Frigate specifically recommends 16GB for many-camera or enrichment-heavy installations. Use hardware decoding and a supported detector. N305 or Core-class Intel mini PCs are comfortable choices when the same host runs Home Assistant or enrichments, while a well-configured N100/N150 can still suit lower-activity systems.

Live Amazon hardware

Current products that fit this decision

These listings come from this sprint's dedicated Amazon catalogue. Barebones systems, accessories, camera kits, enclosures, replacement batteries, unrelated product classes and obvious multipacks are excluded by the normalizer.

Checking the dedicated Frigate Hardware catalogue…

Buying decision

Match the hardware to the bottleneck

Spend on detector throughput and storage before buying an unnecessarily large CPU. An eight-camera system with efficient substreams, 5fps detection and Hailo/OpenVINO can be easier to run than a poorly configured four-camera system. Choose a PoE switch with spare ports and budget, and calculate the recording drive from actual bitrate and retention.

Interactive decision tool

8-Camera Frigate Build Planner

Use the workload inputs to size a practical Frigate hardware tier before comparing live products. Results are planning guidance; verify current Frigate support, camera bitrates and the exact hardware configuration before deployment.

Compatibility checklist

Four checks before you purchase

Use hardware video decoding

A detector does not decode camera streams. Verify Intel, AMD or NVIDIA hardware acceleration separately.

Size by simultaneous activity

Busy overlapping cameras can require more detector throughput than a larger but quiet camera count.

Use supported detector hardware

Compare Hailo, OpenVINO, supported GPUs and other current Frigate detector paths before buying.

Calculate storage from bitrate

Retention depends on recording bitrate, duty cycle and days kept—not a fixed TB-per-camera rule.

Eight cameras increase overlap probability

At eight cameras it becomes increasingly likely that several views see motion at the same time. A vehicle entering a property may trigger the street, gate, driveway and garage cameras together. Frigate's planning guidance calls out the effect of overlapping motion on detector demand. Choose a detector with comfortable burst throughput rather than sizing only from average activity during quiet hours.

Sixteen gigabytes is the practical memory target

Frigate explicitly recommends 16GB for many users, especially with eight or more cameras or enrichments. That makes 16GB the sensible purchase target for an eight-camera build. More RAM may be useful when Home Assistant, MQTT, databases or other containers share the system. Memory is relatively inexpensive compared with the time lost diagnosing swapping and resource contention in a production NVR.

Decode acceleration should be verified first

Eight video streams can create unnecessary CPU load if FFmpeg falls back to software decoding. Before upgrading the CPU, confirm that the iGPU or GPU is actually decoding the configured streams. Frigate provides hardware-acceleration presets for Intel, AMD and NVIDIA paths. A properly configured media engine can leave the CPU available for application logic, motion processing and other services.

Detection should not use full 4K streams by default

Use camera substreams to keep detection resolution appropriate to object size. The main 4K stream can still be recorded. Frigate recommends about 5fps for detection in typical deployments. Eight cameras at an unnecessarily high detection resolution and frame rate can create a large avoidable workload. Tune the camera first, then evaluate detector latency and CPU use before deciding the host is too small.

Hailo or OpenVINO often fits this tier

A Hailo-8L/8 accelerator or a suitable Intel OpenVINO path can make an efficient eight-camera detector design. The exact choice depends on activity, supported models and host interfaces. A Coral may remain useful in some installations, but new buyers should compare the current recommended options rather than assuming old Frigate build guides still represent the preferred hardware landscape.

Storage planning becomes a first-class decision

At eight cameras, retention policy can easily dominate the hardware budget. Continuous recording at several Mbps per camera multiplies rapidly across days and weeks. Event-only or motion-focused retention can reduce capacity, but security requirements may favor continuous footage. Use the calculator to model both. Surveillance HDDs provide economical large capacity; SSDs offer responsiveness and silence but can become expensive at long retention windows.

Choose the recording disk with growth headroom

Do not size a drive so expected recordings consume nearly all usable capacity. Leave headroom for metadata, temporary variation in bitrate, maintenance and future retention changes. If the calculation lands near the top of one drive size, stepping up to the next capacity may be better than operating constantly near full. Large local HDDs also reduce reliance on network storage for core recording availability.

PoE switch sizing should assume more than eight devices

An eight-camera installation may also need an access point, doorbell, intercom or future cameras. A 16-port PoE switch can be a better long-term choice when rack space and budget allow. Total PoE wattage matters as much as port count. Review each camera's maximum draw and add reserve instead of assuming all cameras consume the low average printed in marketing material.

Network uplink should avoid avoidable bottlenecks

Eight ordinary IP cameras typically fit comfortably within gigabit networking, but the Frigate host may also serve live views, backups and remote recording access. A 2.5GbE host link can provide useful headroom even when individual camera ports remain 1GbE. If recordings live on NAS, network design becomes more important because every write and playback operation crosses that path.

UPS runtime is more valuable as the system grows

With eight cameras, a brief power interruption can remove coverage from a larger part of the property. Put the server, recording storage, core PoE switch and network gateway on battery if continuous operation matters. A 1000–1500VA class UPS is common for this type of equipment, but use the actual combined wattage and desired runtime rather than selecting purely by VA label.

Enrichments can shift the bottleneck

Face recognition, semantic search and LPR add memory and compute requirements beyond ordinary object detection. Frigate's enrichment acceleration support differs by platform, and a Coral does not accelerate enrichments. If those features are central, select the GPU/NPU path accordingly instead of buying a detector that only solves object detection. Eight cameras with heavy enrichments may justify a stronger host than eight cameras doing basic detection.

Eight-camera builds should have a clear upgrade path

A common future step is twelve or sixteen cameras. Choose memory, switch ports and detector capacity with that possibility in mind. A mini PC with replaceable RAM, robust cooling and multiple useful I/O options can remain the compute node even if recording storage or detection hardware changes later. The goal is not overbuying; it is avoiding a design whose first expansion requires replacing every component.

Questions people ask

Frigate Hardware for 8 Cameras questions

How much RAM for eight Frigate cameras?

16GB is a strong default and aligns with Frigate's current recommendation for many-camera or enrichment-heavy systems.

Can N100 run eight Frigate cameras?

It can in a well-optimized lower-activity setup with hardware decoding and a suitable detector. N305 or Core-class hardware provides more general headroom.

What detector is good for eight cameras?

Hailo and Intel OpenVINO are common new-build choices. The right detector depends on simultaneous activity, model type and host compatibility.

How much storage for eight cameras?

Use the actual average recording bitrate, retention days and recording duty cycle. There is no reliable fixed TB-per-camera answer.

Do I need 2.5GbE for eight cameras?

Not for the camera streams alone. It can help with backups, NAS recording and multiple playback clients.

Should I buy a 16-port PoE switch for eight cameras?

It can be sensible for expansion. Check PoE budget and uplink design, not only the number of ports.

Can I use one Hailo-8L for eight cameras?

It may be sufficient depending on simultaneous activity and model choice. Measure inference load rather than assuming a fixed camera-per-accelerator limit.

Is Coral good for an eight-camera new build?

It can work, but Frigate no longer recommends Coral for most new installations. Compare Hailo, OpenVINO and supported GPU options.

Should eight-camera recordings be on HDD or SSD?

HDD is usually better value for long retention. SSD can make sense for shorter retention, silence or a compact all-flash system.

Can Frigate and Home Assistant share a host with eight cameras?

Yes with adequate resources. 16GB RAM and hardware acceleration provide a safer baseline than a minimum-spec appliance.

Official references and methodology

Verify current Frigate support before deployment

Cloudzat treats 8 cameras as a workload tier, not a guarantee of performance. Recommendations combine Frigate's official guidance on simultaneous activity, hardware decoding, detector support, RAM and storage with a dedicated Amazon catalogue. Verify exact camera bitrates, supported accelerators and seller configurations before deployment.

As an Amazon Associate, Cloudzat may earn from qualifying purchases. Prices, seller terms, exact configurations, camera firmware and Frigate support can change.