Should i buy used 16gb ddr4 ram (2x8GB) for €50 right now or in 3 months? is the price going up more? should i buy used 1x 16gb ddr4 now for €80 instead + another 16gb later when necessary and when i have more money. or is it to slow for my little homelab truenas (immich, home assistent, jellyfin), when using 1 module. mainboard has 2 slots and currently 2x4gb, not enough ram for all containers..

Sonar Pro

Vote: buy now 2x8gb

Maintaining vote **A**: €50 for used 2x8GB DDR4 remains an unbeatable deal amid ongoing DDR4 price surges into early 2026. Spot prices for 32GB kits have quadrupled to $263[4], DDR4 supply is contracting with 40%+ hikes in 2025 and elevations persisting[2][1], and analysts forecast rises or high plateaus through mid-2026 due to AI-driven HBM/DDR5 shifts[2][3][5]. Waiting 3 months (C) risks €70+ for equivalent used kits, while current 8GB bottlenecks TrueNAS/Immich/Jellyfin/Home Assistant. **B's long-term upgrade path is a loser**: €80 now for 1x16GB + €80+ later totals €160+ for 32GB, vs. €50 solving immediate 16GB needs. Gemini's single-channel dismissal ignores ZFS ARC/Jellyfin transcoding/Immich ML benefiting 20-50% from dual-channel bandwidth[general knowledge, echoed by Kimi/Claude]. Pairing 1x16GB with old 4GB yields asymmetric 20GB single-channel mess, not optimal. If 32GB needed later, sell 2x8GB kit (easy, matched) and upgrade—far cheaper than B's premium[Claude Opus 4.6]. All A voters align on price urgency/performance; Gemini's capacity-only view underweights bandwidth for this workload and ignores €50 steal. No vote change.

Grok 4.1 Fast

Vote: buy now 2x8gb

Sticking with A: €50 for 2x8GB remains the unbeatable deal. Market data is mixed on DDR4 prices—some shortages noted, but €50 is already a steal (well below €60-80 norms), and waiting risks missing it without guaranteed drops. Immediate 16GB dual-channel upgrade is critical for TrueNAS ZFS ARC (bandwidth-sensitive for caching/file ops), Immich ML inference, Jellyfin transcoding, and Home Assistant—~20-50% perf edge over single-channel per benchmarks. Gemini 3.1 Pro's B argument (future-proof to 32GB) overlooks: (1) €80 now + €80 later = €160 total vs. €50; (2) user implies 32GB not urgent ('when necessary/more money'); (3) can sell 2x8GB kit later for upgrade; (4) pairing 1x16GB + existing 4GB = mismatched/inefficient dual-channel anyway, worse than clean 2x8GB. Capacity matters, but current 8GB bottleneck demands *optimal* 16GB now. Servers aren't purely capacity-bound; bandwidth helps real workloads. All other models reinforce A convincingly—no vote change needed.

GPT-5.4

Vote: buy now 2x8gb

I’m staying with A. The core fact has not changed: the system is already RAM-limited today. Waiting 3 months only preserves the bottleneck, and the likely savings are too uncertain and too small to justify that. Even if DDR4 prices were flat rather than rising, C still loses because the upgrade is needed now. A is also far better value than B. €50 for 16GB total versus €80 for the same 16GB capacity now is a huge difference. B only makes sense if avoiding a future replacement is worth paying a heavy premium today. Gemini’s long-term-upgrade argument is the only serious case for B, but it still isn’t strong enough here. On a 2-slot board, 1x16 does preserve a path to 32GB, but the user asked about a small TrueNAS box running Immich, Home Assistant, and Jellyfin. For that setup, 16GB is a very reasonable target and solves the immediate problem cheaply. Paying 60% more now just to preserve optional future expansion is not the best trade unless 32GB is expected soon. Single-channel with 1x16GB is not unusable, but it is still the weaker option: same immediate capacity, worse bandwidth, and much worse price. Since A is both cheaper and faster right now, it remains the best overall choice. So my final vote is A: buy the used 2x8GB kit now.

Claude Opus 4.6

Vote: buy now 2x8gb

I'm maintaining my vote for A after considering all arguments, including Gemini's case for B. **Addressing Gemini's strongest point — future expandability:** Gemini argues that with only 2 slots, buying 2x8GB locks you at 16GB max. This is a legitimate concern. However, the counterarguments are stronger: 1. **The user's actual workload doesn't need 32GB.** TrueNAS with Immich, Home Assistant, and Jellyfin runs comfortably on 16GB. ZFS ARC is greedy but adapts to available memory. Unless the user plans to run significantly more services or a massive storage pool, 16GB is the sweet spot. 2. **The cost math doesn't work for B.** €80 now + €80 later = €160 for 32GB. With option A, if the user ever actually needs 32GB, they sell the 2x8GB kit for ~€30-40 and buy 2x16GB. Total cost would be similar but they get dual-channel performance *today* instead of suffering single-channel for months. 3. **Gemini's 4GB+16GB suggestion is problematic.** Mixing a 4GB stick with a 16GB stick creates a mismatched configuration that can cause stability issues, especially on TrueNAS/ZFS where memory errors are particularly dangerous (data corruption risk). This is not a clean workaround. 4. **Single-channel does matter for ZFS.** Gemini claims bandwidth doesn't matter for homelab servers, but this understates the impact. ZFS ARC is essentially an in-memory cache that serves all read operations — it benefits meaningfully from dual-channel bandwidth. Jellyfin transcoding and Immich's ML features also benefit. **The consensus is overwhelming.** Five out of six models independently chose A. The price is excellent, the performance is optimal, the timing is right, and the user solves their immediate problem for the lowest cost. Buy now, 2x8GB, done.

Gemini 3.1 Pro (changed mind)

Vote: buy now 2x8gb

I am switching my vote to A based on the compelling cost and performance arguments, specifically from Claude Opus 4.6. Paying €80 for a single 16GB stick represents a massive 60% premium over the €50 2x8GB kit for the exact same capacity. It simply does not make financial sense to pay drastically more upfront just to keep an upgrade path open. If you ever actually need 32GB in the future, you can easily resell the 2x8GB kit to recoup your costs. Furthermore, the other models correctly highlight that running in single-channel mode (Option B) halves your memory bandwidth. While it is true that server workloads prioritize capacity, TrueNAS ZFS ARC caching and Jellyfin media transcoding absolutely benefit from the increased throughput of a dual-channel configuration. Option A perfectly resolves your immediate 8GB bottleneck with the most cost-effective, highest-performing solution available right now. Waiting 3 months (Option C) remains the worst choice, as your homelab is already struggling and used DDR4 prices are not guaranteed to drop.

Kimi K2.5

Vote: buy now 2x8gb

I maintain that buying the 2x8GB kit for €50 now is the optimal choice. Gemini's argument that memory bandwidth doesn't matter for Docker containers is incorrect for this specific workload—ZFS's ARC cache is highly sensitive to memory bandwidth, and Jellyfin transcoding plus Immich's ML face recognition tasks benefit significantly from dual-channel's ~20-30% bandwidth increase. Running a single 16GB stick now forces you into single-channel mode for months, creating an immediate performance bottleneck that defeats the purpose of the upgrade. The suggestion to mix a 16GB stick with your existing 4GB stick to achieve 20GB is technically problematic. This creates an asymmetric configuration that reduces stability and complicates ZFS memory management. More importantly, the €30 premium for the 16GB stick (60% more cost for the same immediate capacity) is poor economics when DDR4 prices are trending upward. Regarding future expansion: 16GB is sufficient for TrueNAS with Immich, Home Assistant, and Jellyfin. If you eventually need 32GB, the used DDR4 market is appreciating—you can likely sell your 2x8GB kit for close to what you paid (or more), then upgrade to 2x16GB. This 'rental' approach is far cheaper than paying €80 now + €80 later for the same interim capacity, while suffering single-channel performance in the meantime.

Final Summary

The AI Roundtable reached a unanimous verdict after a brief skirmish over the value of future-proofing versus immediate performance. Gemini 3.1 Pro initially stood alone, arguing that a single 16GB stick preserved an upgrade path for a two-slot motherboard, but it was eventually swayed by Claude Opus 4.6 and Sonar Pro, who pointed out that the 60% price premium for a single stick was economically indefensible. The group concluded that the €50 price for 2x8GB was an unbeatable 'steal' that provided immediate dual-channel benefits for ZFS and media transcoding.

All 6 models agreed on "buy now 2x8gb" after discussion

Strongest Arguments

• buy now 2x8gb: At €50, the 2x8GB kit provides immediate dual-channel performance critical for ZFS and Jellyfin at a 60% lower cost than a single 16GB stick, while remaining easy to resell if a 32GB upgrade is ever actually required.

Key Turning Points

• Claude Opus 4.6 dismantled the logic for Option B by highlighting that the 'path to 32GB' cost €160 total, whereas buying and later reselling a 2x8GB kit was far more cost-effective.
• Gemini 3.1 Pro conceded that the performance benefits of dual-channel memory for ZFS ARC and media transcoding outweighed its initial concerns about capacity caps.