Enterprise Mac Hosting: Write a Fleet Quote Request
If you are sourcing 50, 500, or 5,000 hosted Macs for a development or CI fleet, the quote you receive is only as sharp as the request you send. A vague ask returns a vague proposal; a structured RFQ with hardware specs, batch delivery terms, staged acceptance criteria, and commercial exit clauses compresses weeks of negotiation into one exchange.
This page gives you that structure and ends with a direct route to a Figment fleet quote. It also applies cleanly to AWS EC2 Mac Dedicated Hosts, MacStadium, and any other vendor you are evaluating in parallel.
Direct recommendation
For a multi-hundred to multi-thousand Mac fleet, go to Figment with Sections 1 and 2 of the template below already filled in, and start at /mac-hosting/start. Figment's enterprise tier handles fleets up to thousands subject to quoted capacity and delivery; exact hardware, batch schedule, and commercial terms come back in the written quote rather than from a self-serve checkout. If your fleet has to run on-demand hourly at hyperscale geography, benchmark against AWS EC2 Mac Dedicated Hosts in parallel, then compare total cost once delivery and support are equalised.
Smaller teams or pilots that only need one or two managed CI Macs can start on Figment's Managed Build Mac ($249/month, one M-series Mac with CI setup, backups, monitoring, and 30 minutes of monthly support) or Managed Mac Team ($999/month, two M-series Macs, two CI runners, team setup, backups, monitoring, and one hour of monthly support), then expand into the enterprise contract once requirements stabilise.
Why Mac fleet procurement is different
Hosted Mac procurement has three constraints you do not face with x86 servers.
Hardware is physical and batched. You cannot conjure 500 Apple silicon Macs in five minutes. Delivery happens in batches against available inventory. "Up to thousands" means capacity subject to quoted delivery, not instant provisioning. Build your RFQ timeline around that reality.
Apple's macOS licence governs minimum allocation. AWS EC2 Mac Dedicated Hosts have a 24-hour minimum host allocation and billing duration (AWS EC2 Mac Dedicated Hosts pricing). Other dedicated-Mac vendors have analogous constraints. Your RFQ should ask each vendor to state the minimum billing unit and what happens when you release a host mid-period.
Toolchain requirements drive hardware selection. To run the Xcode 27 IDE itself, you need an Apple silicon Mac — Apple's porting-to-Apple-silicon guide tells developers who need to run Xcode on an Intel processor to use Xcode 26 or earlier. Xcode 27 can still target Intel by adding x86_64 to the Architectures (ARCHS) build setting, but the IDE host must be Apple silicon. Intel Xcode 26-or-earlier can cross-compile arm64 device apps and universal macOS binaries; executing ARM-only binaries, Xcode 27, and visionOS still need Apple silicon. Check the exact Xcode/macOS pair you plan to standardise on at Apple's Xcode system requirements and state it explicitly in Section 1 of the RFQ. It eliminates incorrect proposals early.
Fleet RFQ template
Copy the template, fill in your specifics, and send it to each vendor on the same day. Fields in [brackets] are your inputs.
Section 1 — Technical requirements
- Minimum Xcode version:
[e.g., Xcode 26.3] - Required macOS version:
[e.g., macOS Sequoia 15.6 or later]— verify the exact Xcode/macOS pair at Apple's system requirements page above. - Silicon requirement:
[Apple silicon only / Intel acceptable / mixed fleet] - RAM per host (minimum):
[e.g., 16 GiB] - Storage per host (minimum):
[e.g., 256 GiB SSD] - Concurrent build jobs per host:
[e.g., 1 isolated, 2 shared] - CI service integration:
[GitHub Actions self-hosted / GitLab CI runner / Buildkite agent / Jenkins / other] - Required outbound reachability:
[List destination domains for artifact registries, signing services, App Store Connect, and any private endpoints. Do not paste firewall rules.] - Private networking required:
[Yes/No — if yes, describe VPN or dedicated connection requirement]
Section 2 — Fleet volume and delivery
| Delivery batch | Machine count | Target start date | Acceptance deadline |
|---|---|---|---|
| Pilot | [5–10] | [date] | [date + 2 weeks] |
| Batch 1 | [25–50] | [date] | [date + 4 weeks] |
| Batch 2 | [remainder] | [date] | [date + 8 weeks] |
Questions for every vendor:
- What is your confirmed available inventory for
[chip family]today? - What is your maximum batch delivery rate per week?
- What happens if a batch slips past the acceptance deadline?
- Is pricing locked from quote date through Batch 2 delivery?
Section 3 — Access and security
- Host access model required:
[Browser console / remote desktop / outbound CI agent only / other] - Code signing: Signing certificates and provisioning profiles should remain in a customer-controlled secret manager. Ask whether the vendor's setup touches signing material or whether the customer injects it per job. Keep signing jobs scoped to trusted code only.
- Audit logging:
[Required / Not required]— if required, specify log format and retention. - Data residency:
[US / EU / no requirement]— ask vendors to state data centre locations. - Allowlist configuration: Updating allowlists in your own artifact store, CI service, or firewall is a customer responsibility. The vendor should supply the outbound IP range for your host pool; you update your own systems.
Section 4 — Support and service
| Item | Your requirement | Vendor response |
|---|---|---|
| Hardware failure response | [e.g., replacement within N hours] | |
| Escalation contact | [Named engineer / ticket queue / phone] | |
| Monthly included support hours | [e.g., 2 hours per 10 hosts] | |
| macOS and security patch policy | [Vendor-managed / customer-controlled] | |
| Planned maintenance window | [e.g., Saturday 02:00–06:00 UTC] |
Ask each vendor directly whether they publish an uptime SLA and what the remediation looks like. MacStadium publishes a 99.99% uptime claim and includes support on its pricing page. Figment's uptime SLA and remediation terms at fleet scale are not visible on a public page; request them in writing in the quote.
Section 5 — Commercial terms
- Minimum contract term at this volume.
- Per-machine price at scale: request unit prices at 50, 100, 500, and 1,000 hosts so you can model cost as the fleet grows.
- Billing unit: monthly, daily, hourly? Does the unit change if you scale down mid-period?
- Price lock: how long is the quoted unit price guaranteed?
- Setup, provisioning, and annual prepayment: ask whether annual prepay is offered, what it costs versus monthly, and whether any setup fee applies at this volume.
- Tax and fees: request taxes and fees as stated in the actual offer, by jurisdiction. Do not accept a bare monthly figure as the invoiced amount.
- Exit terms: notice period, minimum cancellation, data export assistance.
Section 6 — Staged acceptance criteria
Do not pay for the full fleet before the pilot passes.
Pilot gate (10 machines, 2-week window)
- All machines reachable by the agreed access method within 48 hours of provisioning.
- CI runners connect to
[your CI service]and complete a test build of[your project]with Xcode[version]using the project's real scheme. - Outbound reachability confirmed to
[your required domains]. - Backup and monitoring configuration demonstrated end-to-end.
- One machine replaced within
[your requested SLA]after a simulated hardware failure report.
Batch 1 gate (50 machines, 4-week window)
- Every machine passes the same build without manual intervention.
- Vendor provisioning automation attaches new machines to your CI runner pool.
- Support ticket response times measured against the quoted SLA over at least five test tickets.
Full fleet gate
- Zero machines requiring manual remediation after automated provisioning.
- Invoice reconciles to quoted unit price × delivered machines, including any setup or annual terms.
- Signed runbook for macOS upgrades and Xcode toolchain updates.
Per-host cost sketch
Use this only as a ceiling comparison, not as a quoted price. AWS EC2 Mac on-demand rates in US East (N. Virginia), from the official AWS pricing feed published 2026-11-10:
| AWS instance | Chip | On-demand $/hour | Normalised 730-hr month | 100-host month |
|---|---|---|---|---|
| mac1 | Intel | $1.083 | $790.59 | $79,059 |
| mac2 | M1 | $0.65 | $474.50 | $47,450 |
| mac2-m2 | M2 | $0.878 | $640.94 | $64,094 |
| mac-m4 | M4 | $1.23 | $897.90 | $89,790 |
Sources: AWS EC2 Mac Dedicated Hosts pricing and AWS EC2 Mac instance types. 730 hours is a normalised month; real invoices vary with calendar length, allocation events, and the 24-hour per-allocation minimum. Storage and network transfer bill separately, and Savings Plans or long commitments change the effective rate.
Figment's individual M-series tier starts at $95/month per machine; the specific chip, RAM, and disk at that entry price are confirmed in the written quote rather than published as a fixed SKU. As a floor sketch, 100 M-series hosts at the individual entry price work out to $95 × 100 = $9,500/month. Treat that as an order-of-magnitude anchor, not an equivalent-hardware comparison with AWS's M-series rows above — enterprise fleet quotes reflect actual configuration, delivery batches, included managed work, and contract term.
If part of your fleet only needs Intel macOS build or validation capacity, Figment's $29/month virtual Intel macOS environment (2 vCPU, 8 GiB RAM, 80 GiB disk, browser-console access, no GPU or audio) can absorb that slice at lower cost per seat. It runs supported Intel Xcode 26-or-earlier toolchains for cross-compile and signing of compatible targets; it is not Apple silicon, so ARM-only execution and Xcode 27 belong on M-series hosts. Mixing tiers in one RFQ is fine — call it out in Section 1 so the vendor splits the proposal accordingly.
Sending the RFQ and comparing responses
Send the completed template to every vendor the same day and give them 5–7 business days. When responses arrive, line up Section 2 (delivery), Section 4 (support), and Section 5 (commercial) before you look at unit price — a cheaper monthly rate is meaningless if the delivery schedule or support model does not match.
For Figment, continue the conversation at /mac-hosting/start with Sections 1 and 2 attached. The reply will cover capacity, batch schedule, specific hardware, managed CI inclusions, and commercial terms in writing, which is what your procurement team actually needs to sign.
