One card, no splitting, long documents
The medium business build
Thirty-five people at 4K context, seventeen or eighteen at 8K, on a 70B model.
Built for
$26,000 – $28,500
Parts alone
$22,940 – $25,440
The difference is a flat $3,000 for assembly, operating system and serving-engine configuration, burn-in testing and warranty handling. Flat, not a percentage — a bigger GPU should not cost you more in labour. Component prices verified 2026-08-22.
What this one is for
- A 70B model with room for long conversations and long documents
- Retrieval over a real document estate, running beside everyday chat
- Departments that need the machine responsive at peak, not just on average
Where it stops
This is one machine. If it is down on a Tuesday, everyone is down on a Tuesday. Above roughly twenty concurrent people the honest answer stops being a bigger box and becomes two boxes, which is the multi-department build.
Every part, and what it costs
Verified against live US retailer listings on 2026-08-22. Nothing here is estimated. A component we could not price would be missing rather than guessed at.
| Part | Specification | Cost |
|---|---|---|
| GPU | NVIDIA RTX PRO 6000 Blackwell, 96 GB GDDR7 ECC Same 1,792 GB/s as an RTX 5090. It buys three times the memory and ECC, not speed. Launched at $8,565 in March 2025. | $14,500 – $17,000 |
| CPU | AMD Ryzen Threadripper 9970X, 32 cores First-party, in stock. The 24-core 9960X is $1,400 when available. | $2,000 |
| Motherboard | ASUS Pro WS TRX50-SAGE WIFI Two full PCIe 5.0 x16 slots with no M.2 tradeoff. | $851 |
| Memory | 128 GB DDR5 ECC RDIMM Registered ECC, four channels. This line has roughly doubled since January. | $4,099 |
| Storage | 4 TB PCIe Gen4 NVMe Enterprise drives with power-loss protection are $2,000–3,000 and quoted on request. | $640 |
| Power supply | Corsair HX1500i, 1500 W, ATX 3.1 Full output on a 120 V circuit. | $350 |
| Cooling | Noctua NH-U14S TR5-SP6 One of only two air coolers made for this socket. | $140 |
| Chassis | Fractal Define 7 XL Fits the CEB board. | $240 |
| Network | Intel X550-T2 dual 10 GbE Add-in card; this board has no onboard 10 Gb. | $120 |
| Parts subtotal | $22,940 – $25,440 | |
| Build, configuration, burn-in and warranty handling | $3,000 | |
| Total | $26,000 – $28,500 | |
Component prices moved sharply through 2026 — memory roughly doubled in the first quarter alone, and GPUs more than doubled over the year. A quote is good for the week it is written, and we re-price honestly rather than carrying an old number forward.
How it is built and configured
- Serving engine
- vLLM, single GPU, no parallelism to tune
- Model format
- AWQ 4-bit, 37 GiB of weights, leaving about 45 GiB for context
- Memory
- ECC throughout, on the card and on the board
- Power
- About 1,000 W under load with the Max-Q card
- Network
- 10 Gb
Why the user counts have a context length on them
A model's weights are fixed, but every person talking to it needs their own context, and that is what fills the rest of the card. For a 70B model it costs about 320 KiB per token — so one person at 8K tokens uses 2.5 GiB and the same person reading a long contract at 32K uses 10 GiB. The same machine serves very different numbers of people depending on the work.
That is why every figure here names the context it assumes. A vendor quoting you a user count without one has not done the arithmetic.
See the full VRAM arithmeticWhy this card and not a cheaper one
Two cards with the same memory can answer at very different speeds, because for one person at a time generation tracks memory bandwidth more than anything else. Every card we costed, with the number most vendors leave off the table:
| Card | VRAM | Bandwidth | Power | Price |
|---|---|---|---|---|
| RTX PRO 4000 Single slot and frugal, but 24 GB does not hold a 70B and 672 GB/s is slow to answer. | 24 GB | 672 GB/s | 145 W | $2,799–4,536 |
| RTX PRO 4500 Same memory as a 5090 at half the bandwidth and a similar price. ECC and 200 W are real, but you feel the missing bandwidth on every reply. | 32 GB | 896 GB/s | 200 W | $3,999–5,499 |
| RTX 5090 The bandwidth-per-dollar winner, by nearly double. What the small-team and small-business builds use. | 32 GB | 1,792 GB/s | 575 W | $4,300–5,000 |
| RTX PRO 5000, 48 GB Enough for a 70B with little room left for people. An awkward middle. | 48 GB | 1,344 GB/s | 300 W | $6,500–8,699 |
| RTX PRO 5000, 72 GB The cheapest memory on the table at about $119/GB. Worth asking us about if you want capacity over speed. | 72 GB | 1,344 GB/s | 300 W | $8,600–13,136 |
| RTX PRO 6000, 96 GB In this build Same bandwidth as an RTX 5090 with three times the memory and ECC. You are buying room, not speed. | 96 GB | 1,792 GB/s | 600 W | $14,500–17,000 |
The RTX PRO 4500 is the one worth pointing at. Same 32 GB as an RTX 5090, error-correcting memory, a third of the power, and a blower designed to sit beside another card — at half the bandwidth for a similar price. On a spec sheet it looks like the smarter buy. In use it answers at roughly half the speed. We nearly specified it ourselves.
Tell us what you are trying to run
Describe the workload, how many people need it, and how sensitive the data is. You will get a straight answer about whether owning the hardware makes sense for that situation — including when it does not, and a subscription would serve you better.
Prefer to talk? Call James on 832-338-2926. Please do not send confidential, client-privileged, health, financial-account or credential information through this form.