Pricing

Simple, one-time pricing. Installed by hand.

Every system includes hardware, software, on-site installation, Active Directory setup, training, and a full year of support. No per-seat fees, ever.

Low Tier

F-18 Hornet

Small teams and departmental deployments.

Starting at$22,900one-time
Up to ~150 people · Professional offices, small manufacturers, and departments
  • GPUNVIDIA RTX PRO 6000 Blackwell Max-Q · 96 GB
  • CPUAMD Ryzen 7 5700X · 8 cores
  • Memory64 GB DDR4
  • Storage1 TB NVMe · RAID 1 mirrored
Get a quote for the F-18 Hornet
High Tier

F-22 Raptor

Largest models and highest concurrency.

Starting at$59,990one-time
250+ people or heavy engineering use · Larger organizations running multiple models side by side
  • GPU2× NVIDIA RTX PRO 6000 Blackwell Max-Q · 192 GB
  • CPUAMD Ryzen 9 5950X · 16 cores
  • Memory128 GB DDR4
  • Storage4 TB NVMe · RAID 1 mirrored
Get a quote for the F-22 Raptor

Need more? 4-GPU and NVIDIA DGX Station configurations are quoted per project. Prices exclude sales tax.

Every tier includes

Same software. Same controls. Same white-glove install.

Every tier runs on the NVIDIA RTX PRO 6000 Blackwell. Step up for more cores, memory, and storage, or a second GPU for the largest workloads.

  • Hardware, software license, and on-site installation
  • Active Directory / LDAP integration and MFA setup
  • Administrator and end-user training
  • First year of support, software updates, and 24/7 support
  • 3-year limited hardware warranty
  • No per-seat fees · no per-token charges for local models
  • 30-day money-back pilot
  • $1,200/year from year two for software updates and 24/7 support (optional)
3-year cost calculator

What does private AI cost per person?

Slide to your team size. We’ll suggest a tier and compare three years of Enclave with three years of per-seat AI subscriptions.

The number of per-seat AI subscriptions you’d need to buy instead.

Compare against

GCC High AI means Microsoft 365 Copilot for GCC High and similar government-cloud AI add-ons, which also require GCC High licensing underneath.

Comparison rates are rounded from published list and reseller pricing in 2026 and exclude migration and implementation costs. Your quotes will vary.

RECOMMENDED: F-15 Eagle · from $32,990
$0
estimated 3-year savings
Enclave, 3 years (system + 2 yrs support)$0
GCC High AI add-ons, 3 years $0
Enclave cost per employee, per month (3 yrs)$0
Get an exact quote

30-day money-back pilot

Not satisfied within 30 days of installation? We sanitize the drives on-site, remove the system, and refund 100% of what you paid.

Straightforward terms

75% at signing, 25% at delivery. Delivery and installation within 45 days. Quotes are valid for 14 days because GPU prices move fast.

It’s yours

You own the hardware and a perpetual software license. If you skip support renewal, the system keeps running; you just stop receiving updates.

FAQ

Questions buyers ask us.

What exactly is Enclave?
A private AI server that sits inside your building. Your team gets a ChatGPT-style assistant for chat, document Q&A, drafting, and image analysis. The AI models run on the server itself, so prompts and files are processed on hardware you own instead of by an outside AI provider.
Does Enclave make us ITAR or CMMC compliant?
No product can do that on its own, and you should be wary of anyone who claims otherwise. Enclave is designed to support your ITAR and NIST SP 800-171 program: inference runs on hardware inside your facility with no required external connections, and it includes access controls, MFA, and audit logging. We provide System Security Plan materials for NIST SP 800-171 Rev. 2 (used by CMMC Level 2 and DFARS 7012) and Rev. 3. Compliance remains your organization’s responsibility.
Does it need the internet?
No. Once installed, Enclave can run with no internet connection when models and integrations are local, and your firewall blocks outbound traffic. Optional cloud connections exist but are off unless your administrator turns them on.
How do updates work without internet?
Updates arrive as signed, checksum-verified packages. Your IT team moves them onto the server on a USB drive through your normal media-transfer process and runs one command. The update verifies itself, applies, and restarts. We can also apply it on-site.
Is it as good as ChatGPT?
For the work most teams do every day (summarizing, drafting, rewriting, pulling requirements out of documents, and answering questions about your files) it is excellent. The largest cloud models are still ahead on frontier reasoning. That is why every system comes with a 30-day money-back pilot on your real work.
How does the 30-day money-back pilot work?
We deliver and install the system. Your team uses it for 30 days on real work. If you are not satisfied, we sanitize the drives on-site in front of you, remove the system, and refund 100% of what you paid. Low usage by your own employees is not, by itself, grounds for a refund.
Who supports it? You’re a small company.
You work directly with the engineer who built it. We are local to the Twin Cities and western Wisconsin, so we can be on-site the same day. The system is yours: it keeps running without us, it is built on widely used open-source components, and you get full administrator documentation and runbooks.
Does it work with Active Directory and MFA?
Yes. Enclave signs users in through LDAP/Active Directory or local accounts, supports required TOTP MFA, account lockout, password policy, session timeouts, and an acceptable-use banner on first login.
Can we see what employees are asking?
Administrators can review audit logs and chat history according to your policy, and keyword flagging can highlight conversations that mention program names, markings, or other terms you choose.
What models does it run?
Current systems ship with a 30B-parameter vision-language model that reads documents and images. Larger tiers can run 100B+ parameter models side by side. Because it is your hardware, we can add or swap open models as better ones are released.
See it in your building

Not sure which tier? We’ll size it with you.

Tell us your team size and what you’d use AI for. We’ll recommend a system and bring a demo unit so you can try it before you decide.

Or reach us directly: (715) 410-6587 info@enclaveservers.com
Call Book a demo