What should a data room actually charge for?

• Perspective
• Pricing

Most software pricing quietly tells you what a vendor is optimising for. Read a data-room price sheet and the incentives show through: controls held back for a higher tier, a per-viewer meter on the audience, a fee to get your own files back out. We think that shape is backwards. So what follows is less a price list than an argument about what a data room should charge for — a question we're still working out in the open, not one we've solved.

Charge for value, not for walls

The core promise of a data room is that it protects a confidential deal and moves it forward. Security, in-boundary AI, redaction, a tamper-evident audit trail — those aren't premium add-ons, they're the product. So none of them should sit behind a higher tier. Paywalling the controls that make a room trustworthy prices safety as a luxury — the wrong signal for a confidential deal.

Our stance is that every core capability ships on every plan. What legitimately differs is capacity, not safety: how many of your own operators and admins run the room, how much storage and AI compute you consume, and how much dedicated support and roadmap influence you want. Those are honest axes — they track real cost and value. And every customer rides the same thing: a team shipping improvements continuously, not a frozen feature set.

The AI point needs a sharp edge, because it's easy to get backwards. We never gate the AI itself — cited retrieval, summarisation, redaction, in-boundary translation, and the in-room agents all ship on every plan. But the compute they burn is a real consumable, like storage: answering thousands of questions or translating a room end to end costs compute to serve. So heavy AI usage costs more the same way heavy storage does — we price the compute, not the feature.

Never tax the audience

Here's the line we won't cross: you should never pay for the people who need to see the room. Investors, counterparties, advisers, guests — the audience of a deal — read for free, without limit. Metering per viewer or per page taxes the very thing a data room exists to enable: letting the right people look at the right documents. A deal doesn't cost us more because you invited three more bidders, so it shouldn't cost you more.

That differs from your own deal-team seats. Your operators and admins — the people building and running the room — are a genuine capacity axis, and it's fair for a plan to scale with your team. The distinction is simple: the people who watch the deal are never a line item; the people who run it are.

Pay for the work, not the exit

The last principle is about honesty of design. You should pay for the platform doing deal work — organising documents, answering questions, redacting, keeping the audit trail current — not for artificial limits built to nudge you into an upgrade. A cap that exists only to create friction isn't a feature; it's a toll booth.

And your data is never the hostage. You can take your documents with you whenever you want — no ransom to leave, no fee to retrieve what was always yours. Pricing that holds your files captive isn't pricing; it's leverage.

None of this is finished — we're still refining how these principles become plans. But the principles we're confident about: charge for value, capacity, and the compute you actually use; never for the audience; never for the exit. If a data room's price sheet punishes you for sharing — or for leaving — it's worth asking what it's really optimising for.