Petronella Fleet - Prototype Your Client's Private AI Workload On Our Lab, Then Hand Off A Working System
For MSPs with a concrete regulated-SMB deal that needs private AI infrastructure or compliance-grade architecture. We build, validate, and document the prototype on our fleet. Your client procures hardware direct from any vendor using the bill-of-materials we deliver. You resell inference and compliance workflows under your own SOW.
Stage 1 - Fleet Prototyping (4-Tier Ladder)
Pricing disclosed publicly for prototyping only. Deployment and Managed Service are scoped from the prototyping output. All prototyping engagements gate through a free 15-minute Discovery Call so we can match tier to scope before any Stripe link is sent.
Single use case. Hello-world prototype on our fleet with architecture document. No handoff package. Proves feasibility before larger commit.
Buyer signal: budget-first MSP needs proof before scaling spend.
Pay & Start - $35,000Working prototype, runbook, architecture doc, workload profile, bill-of-materials for hardware procurement, managed-service recommendation. Full MSP handoff.
Buyer signal: default tier - most MSP engagements land here.
Pay & Start - $50,000All Fleet Prototype deliverables plus CMMC Level 2, HIPAA, and NIST 800-171 mapping. Includes SSP artifacts and an audit-evidence package for the MSP's end client.
Buyer signal: DIB-adjacent or healthcare MSPs in assessment window.
Pay & Start - $75,000Compliance-Aware prototype plus signed migration guarantee, 90-day post-handoff support, and an embedded CMMC-RP on the MSP's Slack. Invoice and wire/ACH only.
Buyer signal: serious MSPs with $500K+ AI capex committed.
Inquiry Only - Contact UsStage 2, 3, and 4 - Hardware, Deployment, Managed Service
Prototyping is the only stage with public pricing. Deployment and Managed Service scope from the prototyping output. Hardware is MSP-direct; Petronella Technology Group takes zero hardware margin.
Hardware - MSP procures direct
Using the bill-of-materials delivered in Stage 1, your MSP buys hardware direct from any vendor (NVIDIA, Supermicro, Dell, CDW, and others). We take zero markup, hold zero inventory, carry zero warranty exposure. Optional flat $2,500 procurement coordination fee if you want us to run the RFQ and track delivery.
Deployment & Configuration - custom quoted
After prototyping completes, we issue a custom deployment quote based on rig size, network complexity, compliance overlay, and training scope. Most engagements fall in standard regulated-MSP deployment ranges. We provide the exact number in your prototyping delivery packet.
Managed Service - custom subscription
Monitoring, patching, capacity, private-model updates, CMMC-compliant change management. Monthly subscription scoped from actual deployed environment. 12-month minimum post-deployment; 60-day mutual out-clause after month six. Fee ranges align with regulated-MSP managed-AI market norms.
Optional Tier 2 Add-Ons
Publicly priced. Available to active Tier 2 engagements only. Attach to the engagement via admin flow - these are not standalone Payment Links.
How Fleet Differs From Every Other Managed-AI Vendor
We are services-only. That is a strategic choice, not a limitation.
Your MSP owns the client
You invoice. You own the relationship. We deliver engineering under your SOW when you need us.
Prototype before you commit capex
Four tiers let you spend $35K to validate feasibility or $75K to package a compliance-grade audit deliverable before any client hardware ships.
Honest pricing, no hidden cross-credits
Prototyping fees are deliverable engineering work. They do not convert into coupons toward anything. Each stage stands alone.
Capability Deep-Dives
Explore specialized Fleet capabilities in detail.
Private AI Solutions for MSP Clients
Why regulated SMBs need private AI, what MSPs get wrong selling it, and how Fleet prototyping turns AI opportunities into recurring managed-service revenue. Read the deep-dive.
CMMC Compliance for MSP Clients
Deliver CMMC Level 2 to defense-contractor clients with four CMMC-RP engineers, curriculum, and the $75K Compliance-Aware Prototype tier. Read the deep-dive.
Digital Forensics for MSP Clients
Add DFE-credentialed forensics and expert-witness capability as an optional Fleet add-on. Incident response, litigation support, and court-admissible evidence preservation. Read more.
Custom AI Development
When standard prototyping is not enough: multi-model orchestration, custom fine-tuning, domain-specific RAG, and compliance-grade inference pipelines. Entry at $75K Compliance-Aware tier. Read more.
How a Fleet Engagement Runs, Step by Step
Every prototyping engagement follows the same five steps. You always know what happens next and what you owe before it happens.
Discovery Call (free, 15 minutes)
You bring the client's use case, the compliance overlay (CMMC, HIPAA, NIST 800-171, or none), and the timeline. We qualify scope and recommend one of the four prototyping tiers.
Scope confirmation and payment
After the call we send the Stripe Payment Link for the agreed tier. Prototyping fees are paid upfront and are non-refundable, which is why the call comes first: it exists so neither side commits money to the wrong tier.
Prototype build on our fleet
The prototype runs on Petronella Technology Group's private fleet in a dedicated tenant. Your client's data is processed under NDA with security controls matched to its classification. This is the engineering phase: model selection, pipeline construction, and the compliance overlay where the tier includes it.
Deliverable handoff to your MSP
You receive the deliverables for your tier: working prototype, runbook, architecture document, workload profile, bill-of-materials, and where the tier includes it, SSP artifacts and an audit-evidence package.
Your choice: MSP-direct hardware or our coordination
Hand your MSP the bill-of-materials and buy direct, or attach the optional procurement coordination fee and we run the RFQ and track delivery. Deployment and Managed Service are then quoted from what the prototype actually revealed, so the scope is grounded in a validated build rather than a guess.
How to Decide Which Prototyping Tier Fits Your Deal
The four tiers are not good-better-best upgrades. They answer different questions, and the cheapest tier that answers your client's question is the right one.
Your client wants proof that private AI works for their data before anything else is decided. Nobody has promised a production system yet, and no compliance officer has asked for artifacts. One use case, one architecture document, feasibility answered. If it proves out, the next conversation is a full Fleet Prototype.
The deal is real: your client has signed intent and your MSP needs the full handoff package to scope deployment and open the managed-service conversation. The runbook, workload profile, and bill-of-materials are what let you quote the follow-on work with confidence.
The client sits in defense, healthcare, or another regulated space and the compliance officer or an assessor will ask how the AI system maps to CMMC Level 2, HIPAA, or NIST 800-171. The SSP artifacts and audit-evidence package are the deliverables that answer those questions.
The client has committed serious capex, the migration has to work on a date, and your MSP wants a named CMMC Registered Practitioner available in your Slack through transition. This tier is inquiry-only because it is scoped per engagement, not self-serve.
How Fleet Fits With the Rest of the MSP Partners Program
Fleet is Tier 2 of a four-tier ladder, and every tier stands alone. Nothing here requires buying anything else.
Start with MSP Stack
If your team wants the frameworks and curriculum before the first deal lands, the self-serve Stack membership covers CMMC, private AI, and compliance training on the member's account. Stack and Fleet are independent purchases.
Compare all four tiers
Stack, Fleet, Operator Council, and the Strategic Partnership are each priced and scoped separately on the pricing page. There are no cross-credits between tiers.
See how onboarding works
New to the partner program? The onboarding page walks through the application, fit review, agreement, and first-deal steps. Joining as a Petronella Partner costs nothing.
Frequently Asked Questions
Direct answers to the questions MSP owners ask before committing a prototyping fee.
Who owns the prototype and the work product after the engagement?
Why is the fleet capped at 5-6 engagements per year?
Does my MSP need AI engineers to run the handoff?
What happens after the prototype ships?
Can Petronella take over the end-client relationship?
What data does Petronella see during prototyping, and where does it go?
How do the add-ons attach to an engagement?
Are prototyping fees refundable if the deal falls through?
Can we run Fleet work alongside an active CMMC engagement for the same client?
Is there a way to see the fleet before committing?
Does Petronella mark up the hardware or hold inventory?
What commitment does the managed service carry after deployment?
What the Handoff Package Contains, Deliverable by Deliverable
The deliverables are the product. Here is what each one is and what your MSP does with it after the engagement ends.
Working prototype
The validated build itself: the model running the client's use case on the fleet, with the pipeline exercised end to end. This is what you demo to the client's decision maker to close the production conversation. It is proof of feasibility, not a production system, and the tiers that include handoff package it accordingly.
Runbook
Step-by-step operating instructions for your team: how the system is started and stopped, what the routine checks are, what a degraded state looks like, and what to do first when something breaks. Written for a working MSP technician, not for a data scientist.
Architecture document
The system design: model choices and why, serving topology, storage layout, network path, and where each compliance control attaches. This is the document your MSP uses to scope the Deployment quote and to answer the client's technical review board.
Workload profile
The measured behavior of the prototype that informs the bill-of-materials.
Bill-of-materials
The vendor-neutral parts list your MSP or the client takes to NVIDIA, Supermicro, Dell, CDW, or any other source. Because Petronella Technology Group takes zero hardware margin, the BOM carries no incentive toward any vendor.
Compliance deliverables (Compliance-Aware and above)
The compliance mapping to CMMC Level 2, HIPAA, or NIST 800-171, plus SSP artifacts and an audit-evidence package. These are the documents that sit beside the rest of the client's compliance posture when an assessor asks how AI is governed in the environment.
Questions to Ask Any Private AI Prototyping Partner, Including Us
An MSP owner signing a five-figure prototyping engagement should ask these questions of any vendor, Petronella Technology Group included. The answers below are ours.
Where does the prototype actually run?
Ask whether the work happens on the vendor's own infrastructure or on rented multi-tenant cloud. Our prototype runs on Petronella Technology Group's private fleet in a dedicated tenant. The fleet is Petronella's own prototype and demonstration lab.
Who sells the hardware after the prototype?
Ask whether the vendor marks up the production hardware, because that is where an engagement can quietly change character. We are services-only: your client buys direct from any vendor using the BOM, we take zero hardware margin, hold zero inventory, and carry zero warranty exposure.
Who does the client call at 2 a.m.?
Ask who holds the pager. In the Fleet model your MSP holds the client relationship and the ongoing contract; Petronella Technology Group operates under your SOW. If you want after-hours depth behind your own team, the partner program's co-managed SOC coverage is the layer for that conversation.
What happens to the model weights and the data?
Ask for the ownership and retention terms in writing before payment. Our default: the client owns custom code and fine-tuned weights per the SOW, and no client data is retained beyond the engagement retention window. Get the equivalent answer from any vendor you compare us against.
Is the fee a credit toward later work?
Ask whether the prototyping fee converts into a discount on deployment. Ours does not: fees are non-refundable and each stage stands alone, which keeps every quote honest about what that stage costs.
Who signs the compliance documents?
Ask who stands behind an attestation letter. Documents signed by a credentialed practitioner, such as a CMMC Registered Practitioner, submissions to regulators or assessors, and Business Associate Agreements always carry the Petronella Technology Group, Inc. name. That is true under the free partner agreement too.
How MSPs Position a Fleet Engagement With Their Client
Process guidance only: one way to sequence the conversation with a regulated client. Your pricing to your client is yours to set.
Frame the client conversation around the use case
The strongest first question for a regulated client is not "do you want AI?" It is "which workflow, and what data does it touch?" A document-review assistant touching controlled data has a different compliance shape than an internal policy Q&A bot, and the prototyping tier follows from that answer.
Use the prototype as the decision gate
Present prototyping to your client as the step that de-risks the capex decision. The client is not committing to production hardware; they are paying for a validated answer to "will this work on our data, under our controls?" The bill-of-materials then turns that answer into a procurement plan.
Attach the compliance overlay where the client needs it
If the client's compliance officer is already in the conversation, the Compliance-Aware tier's SSP artifacts and audit-evidence package give them something concrete to file alongside the rest of the client's compliance program, rather than an AI system that exists outside it.
Plan the recurring layer before the demo
The durable economics sit in managed operations: monitoring, patching, model updates, capacity management. Partners who scope that layer early, using the runbook and workload profile from the handoff, walk into the client demo with a recurring-revenue proposal ready instead of a one-time project.
Ready To Scope Your Client's Private AI Prototype?
Book a free 15-minute Discovery Call. We qualify scope, recommend a prototyping tier, and send the Stripe Payment Link only after the call. Questions? Call (919) 348-4912 or contact us. Ready for a peer cohort instead? See Petronella Operator Council.