- MODEL
- LAGUNA S 2.1 / BASELINE
- BASE PLATFORM
- NVIDIA-BASED DESK SYSTEM, 128GB
- INTEGRATION
- SPLAY / STAFFORD, VA
- SERVES
- 1-15 STAFF
PRIVATE AI / OWNED COMPUTE / PRE-ORDER
THE MACHINE IS THE COST.
Compare complete private AI systems built on new commercial NVIDIA-based hardware, installed in your building, and owned by you. No required subscription. No per-seat fee.

05 / THE MACHINES
THE MACHINES
A Node is commercial hardware delivered as a complete private AI system. In the standard installed configuration, external model egress is blocked.
- 01COMMERCIAL NVIDIA-BASED HARDWARE
- 02SPLAYOS + VALIDATED MODEL
- 03SECURED + TESTED + INSTALLED
- MODEL
- DEEPSEEK V4 FLASH
- BASE PLATFORM
- LINKED NVIDIA-BASED SYSTEMS, 256GB
- INTEGRATION
- SPLAY / STAFFORD, VA
- SERVES
- 16-30 STAFF, TWO MACHINES
- MODEL
- GLM-5.2
- BASE PLATFORM
- NVIDIA-BASED CLUSTER, 512GB
- INTEGRATION
- SPLAY / STAFFORD, VA
- SERVES
- UP TO 10 ON GLM-5.2 / UP TO 60 ACROSS SMALLER MODELS
- MODEL
- MINIMAX-M3
- BASE PLATFORM
- NVIDIA PROFESSIONAL GPU SYSTEM, 384GB
- INTEGRATION
- SPLAY / STAFFORD, VA
- SERVES
- 15-75 STAFF, FAST
- MODEL
- GLM-5.2, BEST OPEN MODEL
- BASE PLATFORM
- NVIDIA AI WORKSTATION, 748GB
- INTEGRATION
- SPLAY / STAFFORD, VA
- SERVES
- 15-75 STAFF
If the listed machines do not fit your workload, Splay can build one to order. Memory, speed, and models are set after a call and quoted in writing.
HARDWARE PROVENANCE / SUPPLY CHAIN
SPLAY sources new commercial hardware through established manufacturers and U.S. retail or distribution channels. No grey-market, used, or refurbished components are used unless the buyer approves them in writing.
Every delivery records the platform, configuration, serial numbers, and firmware state. Buyers with contract-specific restrictions can approve the hardware before SPLAY purchases it.
NODE 01 serves 1-15 staff.
A second compatible system and one cable turn NODE 01 into NODE 02.
Two more make NODE 03. Nothing is replaced or migrated.
What you paid counts in full from NODE 01 through NODE 03.
NODE 03 is built for more users. NODE 04 uses a higher-throughput NVIDIA professional GPU configuration for faster answers. NODE 09 is built for the largest models and memory-heavy work. Exact base platforms are disclosed with the quote.
Installed pricing covers the hardware, SplayOS, a validated model, egress configuration, recovery testing, documentation, on-site installation, and 90 days of Launch Care. Hardware warranty, delivery timing, and acceptance terms are stated in each written quote. All machines are pre-order only.
MODEL LEDGER / JULY 2026
Pick a model. See where it lands.
SPLAY BUILD is the standard local software stack installed on every Node. This ledger maps published open models against the machines. The verdict is arithmetic over each model's published size and each machine's installed memory, nothing else. The flag is the developer's home country.
Final shipping configuration is verified during integration before delivery.
The estimate covers the model itself plus 5% overhead; working memory is extra. Sizes are the published figures.
Model score: best closed model = 100 on the Artificial Analysis index, July 2026. Raw scores 38, 40, 51, 44, 51 against the best closed model's 60. Model fit includes 5% overhead; working memory is extra. NODE 04 and 09 are built to order: 50% deposit, quotes hold 30 days. Prices subject to change.
06 / CONTACT
TALK TO US
Splay is based in Stafford, Virginia, and installs on site across Northern Virginia—including Arlington, Alexandria, Fairfax, Loudoun, and Prince William—and Washington, DC. Start with a 30-minute call.