The short answer
First, what the numbers mean — because published figures are scoped so inconsistently that a number without its scope is meaningless. Unless marked otherwise, every band below is the full interactive scope, hardware and software combined: design, software, content, hardware, and integration, delivered working in the space. The bands exclude exhibit fabrication (casework, walls, graphics) and base-building infrastructure (power, data, mounting), which usually live in other vendors’ contracts. Where a public award bundles fabrication in, we say so.
The bands are drawn from the public record — awarded federal contracts and published industry benchmarks — and they are consistent with our own contract history:
- Single kiosk or touch table — roughly $15K–$50K when the purchase is hardware-only: a stock device running vendor software. $50K–$200K for the full scope, once custom software — and, in this class, usually the unit’s own casework — are part of the contract.
- Single gallery interactive or media package — $100K–$300K in public awards for linear-media galleries (content production, AV hardware, and installation together); custom sensing-driven experiences run meaningfully higher for the same full scope — commonly into the mid six figures.
- Multi-station interactive gallery — $500K–$1.5M for the interactive scope alone. Public awards that bundle exhibit fabrication into the same contract run $300K–$3M — and the bundling, not the interactives, is usually what explains the top of that range.
- Building-wide interpretive system — $1M–$10M+ for the media program of a building (interactives, AV, and software across every gallery — still excluding fabrication and construction), inside capital projects that run to tens of millions.
Every figure above traces to a public award or a published benchmark — and when two public numbers for “the same thing” look wildly different, scope bundling is almost always the reason, not vendor pricing. What moves a real project within its band is the next section.
These bands are a summary. The underlying awards — every record, with a source link on every row — are public in The Exhibit Cost Record, the dataset this guide draws from.
Design is the hardest number
Here is the fact that explains most of the confusion in this market: the same firm, pricing honestly both times, will quote a tightly scoped software build meaningfully below an open-ended “design the media for the museum” engagement — a real difference, not a rounding error. It isn’t padding — it’s that design can’t be priced by scope at all.
Software has a scope: features exist or they don’t, and a firm can estimate the work between here and done. Design is subjective — there is no objective “done,” only rounds of exploration until the institution recognizes the answer. So design is really bought as time: an open-ended brief buys many rounds; a tight one buys few. Which means the single biggest lever on your project’s price is not the technology — it’s how well-defined the ask is before anyone prices it. An institution that arrives with interpretive goals, audiences, and constraints settled is buying execution. One that arrives with a blank page is buying the settling, at design-time rates, and that’s the expensive part.
What actually drives the number
Percentages get quoted constantly in this market, and they only make sense if you notice they describe entirely different budgets. Across a whole exhibition’s production budget — fabrication, graphics, everything — digital interactives plus AV hardware and integration typically run 25–45% (kubik maltbie’s published benchmark). Across the exhibition budget as a whole, design fees run 20–25% (Museum Planner’s benchmark). And within a single interactive contract — the number a firm like ours quotes — hardware is roughly a fifth of the total or less: the intelligence is in engineering, not equipment.
Same project, three different pies — and most budget confusion starts with a percentage quoted without its pie. The counterintuitive part holds across all three: hardware is rarely the biggest line. Design, content, and software are where the money actually goes, which is why the previous section — how well-defined the design ask is — matters more than any equipment choice.
The costs that don’t show up in the proposal
Several real costs routinely live outside the interactive firm’s number, and they surprise budgets precisely because each one lands in someone else’s contract: coordination with the exhibit fabricator; infrastructure — power, data, and mounting — that belongs to the general contractor or AV integrator; content licensing; accessibility done late instead of designed in (late always costs more); and spares for hardware that won’t be manufactured in year six. None of these is hidden by anyone in particular. They’re hidden by the seams between contracts — which is why someone on the owner’s side has to own the whole picture.
Why two quotes for “the same thing” come back so far apart
When two bids for the same RFP come back far apart, the usual reaction is that someone is gouging or someone is desperate. Almost always, neither: the two firms priced different projects. One read the open-ended design ask literally and priced the rounds of exploration it implies; the other assumed a tight brief was coming. One priced a productized platform with a decade of operational support; the other priced a custom build warrantied for a year. One included the seams — fabricator coordination, infrastructure, spares; the other left them where the RFP left them, unassigned.
The spread, in other words, is information. Read the cheap bid for what’s missing and the expensive bid for what’s included, and both will teach you what your own RFP failed to say.
What it costs to keep alive
Build cost is half the story — the decade after opening is the other half. We wrote a separate guide on that: What does an interactive cost to own?