Yes, integrated graphics is enough for many home media servers with mixed TVs, browsers, tablets, phones, and streaming boxes when most clients direct-play and the remaining codec conversions are supported by the iGPU's media engine. The answer changes when several clients need simultaneous 4K transcodes, HDR tone mapping, subtitle burn-in, or codecs the integrated hardware cannot accelerate. Client compatibility, not display resolution alone, sets the GPU requirement.
Build a Client Matrix Before Buying a GPU
A mixed-client household can contain devices with very different codec and container support. A modern TV may play HEVC directly while an older browser requests H.264. One tablet may accept the video but not the audio format, and another client may force subtitle rendering into the picture.
Android Central's current long-run NAS test found that most modern clients direct-played media while hardware transcoding became important for older or incompatible devices. That is the right buying order: count incompatible clients, not total clients.
Create a table with each important client, its normal location, supported video codec, audio codec, subtitle behavior, maximum resolution, and whether it usually plays locally or remotely. Then test representative files instead of relying on the marketing label printed on the client.
If nearly every device direct-plays, integrated graphics may spend most of its time idle and still be the correct choice. A dedicated GPU becomes valuable when the incompatibility matrix creates repeated, concurrent video conversion rather than occasional exceptions.
Direct Play Barely Uses the GPU; Video Transcoding Is the Flip
Direct play sends the stored media without re-encoding the video. In that path the server mostly reads data and delivers it over the network. Integrated graphics becomes a major purchasing variable only when the server must decode, transform, and encode video for the client.
A TechRadar review of ZimaBoard 2 as a low-power home server highlights the combination of Intel N150, compact power use, and PCIe expandability. For a media buyer, that means the integrated platform can be evaluated first and a discrete accelerator kept as an upgrade path rather than a mandatory starting part.
ZimaSpace's guide to building a family media server provides the broader boundary: storage, network, client capability, and playback software all influence whether transcoding happens.
Test the dashboard while playing the same file on each client. Record whether the session is direct play, remux, audio-only conversion, or video transcode. Integrated graphics is sufficient when the expensive paths are supported and the number of simultaneous conversions stays below the measured limit.
Codec Generation Matters More Than Raw iGPU Size
Integrated graphics is not one capability class. Different generations support different hardware decode and encode paths for H.264, HEVC, VP9, AV1, bit depth, and HDR workflows. A newer modest iGPU can therefore be more useful to a media server than an older integrated GPU with a larger-sounding model name.
A ServeTheHome test of a compact Intel NAS platform shows why integrated media acceleration changes the value of a low-power CPU for NAS and media duties. The relevant question is which codecs the hardware block handles, not whether the GPU can render games.
Build the library-side half of the matrix as well: common video codec, bit depth, HDR format, audio format, subtitles, and bitrate. If most files already match most clients, even modest integrated graphics can handle the remaining exceptions.
If a large portion of the library uses a codec the iGPU cannot accelerate, software transcoding will shift the load back to the CPU. That is a reason to choose a newer integrated platform or dedicated GPU, not simply to buy more CPU cores and hope the mismatch disappears.
Subtitles and HDR Can Turn One Client Into the Hardest Client
A client that otherwise supports the video may still force a full transcode when subtitles must be burned into frames or when HDR content has to be tone-mapped for an SDR display. Those exceptions are often why a household that “mostly direct-plays” still needs working hardware acceleration.
Digital Citizen's 4K media NAS testing illustrates the broader value of a platform designed to combine storage with media duties. The decision is not whether the server can open a 4K file; it is whether the complete playback path stays responsive under real client constraints.
ZimaSpace's analysis of subtitle burn-in forcing video work is a useful test case. Include the subtitle formats your family actually uses when validating the iGPU.
Do the same with HDR-to-SDR playback. If only one old client creates the hardest transcode in the home, replacing that client can sometimes be cheaper and simpler than buying a discrete GPU that will remain idle for everyone else.
Concurrency and Remote Streaming Decide Whether Integrated Graphics Still Wins
One successful transcode does not prove that an iGPU can serve a household. The real threshold appears when two or more incompatible clients request video conversion while background tasks such as library scans, downloads, or photo services also run.
TechHive's media NAS testing framework separates basic streaming from transcoding and other NAS workloads. That distinction is useful because a mixed-client server can be network-limited remotely even while the graphics engine still has capacity.
Test the maximum realistic household overlap: living-room TV, tablet, browser, and one remote user. Record which sessions transcode, GPU/video-engine utilization, CPU load, network throughput, and whether tone mapping or subtitle rendering changes the result.
Integrated graphics is enough when that combined state remains stable. A dedicated GPU is justified when supported transcodes consistently saturate the media engine, when several 4K conversions must run together, or when a required codec or processing path is outside the iGPU's acceleration support.
Start With ZimaBoard 2 for Moderate Media; Step Up Only at a Measured Threshold
ZimaBoard 2 1664 is the stronger ZimaBoard 2 choice for a mixed-client media server because its 16GB memory tier leaves more room for the media application, metadata, containers, and other household services while the Intel N150 platform handles the basic integrated media path.
Do not choose 1664 expecting more GPU capability than 832; the two variants share the same processor. The extra memory helps app concurrency, not codec support. If the problem is a hardware-transcode ceiling, memory alone is not the upgrade.
Move to ZimaCube 2 when media growth also requires more drive bays, more compute headroom, heavier multitasking, or faster networking. Reserve a dedicated-GPU configuration for workloads that actually exceed integrated graphics after the client matrix and concurrency test.
Integrated graphics is therefore enough for the majority of mixed-client households when client compatibility is reasonably modern and hardware acceleration covers the exceptions. Buy a dedicated GPU only after the playback path shows a repeatable conversion workload that the iGPU cannot carry.
Buying Guide
More to Read

How to Translate CPU, RAM, and IOPS Specs Into Plex Performance
A Buying Guide for turning Plex workload measurements into minimum CPU, RAM, storage, and network requirements without overbuying.

How to Shortlist Home Servers for Plex Using Weighted Criteria
A reproducible Plex buying matrix that separates mandatory gates from preferences and exposes uncertainty before purchase.

What Support and Upgrade Lifecycle Should a Plex Server Provide?
A pass-or-fail buying framework for Plex server support, update history, compatibility, repairability, costs, and migration readiness.

