Document review pricing has not really changed since 2009. The dominant pricing convention — dollars per gigabyte processed, per month, plus per-document review fees — was designed in an era when "review" meant a contract attorney clicking through TIFFs in a hosted Concordance database. The infrastructure to host that data was expensive, the labor to review it was expensive, and the per-GB / per-doc model was a reasonable proxy for both.
Three things have changed since then. Storage is essentially free — the underlying S3 cost of hosting one gigabyte for one month is well under a cent, and yet legal teams routinely pay $10 to $25 per GB per month for hosted review. Labor has split into two tiers: senior associate time at $400–$900 per hour, and contract review time at $35–$60 per hour, with very little in between. And the work itself has been automated — the responsiveness pass that took a team of contract attorneys six weeks in 2015 can be performed in an afternoon by a competent technology-assisted review workflow, and in minutes by a multi-model GenAI classifier.
The result is a pricing model that is now structurally misaligned with the cost of doing the work. Vendors quote per-GB and per-doc rates that bear no relationship to the marginal cost of processing the data, and clients have no way to challenge those rates because the line items are intentionally opaque. This article unpacks where the dollars actually go, what a fair benchmark looks like in 2026, and how the post-2015 amendments to Federal Rule of Civil Procedure 26(b)(1) give in-house counsel a real lever to push back on inflated review estimates.
Almost every eDiscovery quote you receive will be built from some combination of three pricing primitives. Knowing which one is doing what makes the difference between a $50,000 matter and a $500,000 matter on the same data set.
Per-GB hosting is the recurring fee for keeping data live in the review platform. Industry rates in 2026 range from a low of $5/GB/month for self-serve platforms to $25–$40/GB/month for full-service Relativity hosting bundled with project-management overhead. The single most important question to ask is what the per-GB fee actually includes: processing? OCR? near-duplicate identification? email threading? At the high end of the market, those are line items billed separately.
Per-document review is the labor cost of putting human eyes on each document. Contract review through traditional staffing agencies sits at $1–$3 per document for first-pass responsiveness review and $4–$8 per document for privilege review. On a 250,000-document matter, that single line item alone is $250,000–$750,000 — and it is the line item AI-augmented workflows compress most aggressively.
Project management and "ancillary" fees are the line items that turn a $40,000 quote into a $120,000 invoice. They include processing surcharges (often $75–$150 per GB ingested), production fees ($0.05–$0.15 per page Bates-stamped and converted to TIFF), privilege log preparation (frequently quoted at $5–$15 per logged entry), and project-management hours billed at $200–$300 per hour. None of these are unreasonable in isolation. Stacked together on a single matter, they can exceed the cost of the underlying hosting and review combined.
The most useful way to compare vendors is to price the same hypothetical matter against each model. The benchmark below assumes a typical mid-size commercial dispute: 100 GB of collected data, ~250,000 documents after deduplication, a 6-month active review window, 15,000 documents responsive, and 2,000 documents requiring privilege review and logging. All figures are mid-market US rates as of early 2026.
| Line item | Traditional vendor | AI-augmented platform |
|---|---|---|
| Processing & ingestion 100 GB × rate | $10,000 ($100/GB one-time) | Included |
| Hosting 100 GB × 6 months | $15,000 ($25/GB/mo) | $36,000 ($60/GB/mo, all-in) |
| First-pass responsiveness review 250,000 docs | $375,000 ($1.50/doc, contract review) | Included (AI classification + QC) |
| Privilege review 2,000 docs | $12,000 ($6/doc) | Included |
| Privilege log generation 2,000 entries | $20,000 ($10/entry) | Included |
| Production (Bates, TIFF, load files) | $8,000 | Included |
| Project management ~80 hours | $20,000 ($250/hr) | Optional ($150/hr LegalOps) |
| Total all-in | $460,000 | $36,000 |
The traditional-vendor column is not a strawman. It is what a mid-market AmLaw 100 firm working with a national eDiscovery vendor would quote for exactly this matter today. The AI-augmented column is real DecoverAI pricing — $60/GB/month, all-in, no seat fees, no contracts. The order-of-magnitude difference is not in the storage line; it is in the labor line. AI does not save money on the gigabytes. It saves money on the human hours that the per-document model exists to bill against.
What this benchmark does not include: outside counsel time to direct the review, partner-level privilege adjudication on close calls, deposition prep, and the cost of the review platform's own attorney users (which is zero on platforms without seat fees, and substantial on platforms that charge per user). A complete TCO comparison should add those back in — we work through them in the white paper version of this article.
The 2015 amendments to Federal Rule of Civil Procedure 26(b)(1) rewrote the scope of discovery to make proportionality a primary constraint, not a footnote. The current rule allows discovery of any nonprivileged matter "relevant to any party's claim or defense and proportional to the needs of the case," weighing six factors: the importance of the issues, the amount in controversy, the parties' relative access to information, the parties' resources, the importance of the discovery in resolving the issues, and whether the burden or expense outweighs the likely benefit. The Advisory Committee notes describe this as restoring proportionality "to its original place in defining the scope of discovery."
What that means in practice is that a $460,000 review estimate on a $1.2 million dispute is no longer just expensive — it is potentially disproportionate as a matter of law. In Henry v. Morgan's Hotel Group, Inc., 2016 WL 303114 (S.D.N.Y. Jan. 25, 2016), the court quashed third-party subpoenas precisely because the requesting party had not "satisfied the Court that [the subpoenas'] production is proportional to the needs of the case." That case is one of the first reported applications of the amended rule, and the reasoning has been followed by dozens of district courts since. The proportionality framework is not theoretical — it is a working defense against discovery quotes that exceed the value of the case.
If you are receiving review estimates that look out of line with the dispute, the question to ask opposing counsel and your own vendor is no longer "how do we cut costs?" It is "how would this expense survive a Rule 26(b)(1) proportionality challenge?" Vendors who cannot answer that question on the record are vendors whose pricing is built on the assumption that no one will ask.
Before signing an eDiscovery vendor agreement — or before approving an outside counsel's vendor recommendation — in-house counsel should put the quote through five questions. Each one targets a specific category of hidden cost.
1. What is the all-in price for the most likely outcome of this matter? Vendors often quote the per-GB and per-document rates without committing to a total. Force a single number, with assumptions written down. If they cannot or will not do this, that is itself information — the line items are designed to flex upward as the matter progresses.
2. What is the marginal cost if the data set doubles? The honest answer is that storage and AI-classification costs scale linearly and cheaply. Per-document review and project-management hours scale much faster. A vendor whose marginal cost on a doubled data set is more than double their original quote is using the per-document model to run up the bill on every additional collection.
3. Are you charging me per user, per matter, or per gigabyte? Per-user pricing is the single largest cost driver in legacy eDiscovery platforms, and it is the line item that most aggressively penalizes adding outside counsel, expert witnesses, or junior associates to the review team. Modern platforms have abolished it — DecoverAI charges $60/GB/month with unlimited users, no seat fees, and no contracts — and there is no legitimate reason for a 2026 vendor to still bill on a per-seat basis.
4. What happens to my data if I switch platforms mid-matter? Some vendors charge "egress fees" of $25–$100 per GB to release data, plus per-hour charges to produce a load file. Others release data freely. The contract language to look for is whether export is "included at no additional charge" or "available for an additional fee, to be determined."
5. Can you put your privilege-log methodology in writing? If the vendor's privilege log workflow involves human contract reviewers drafting descriptions one document at a time, you are paying twice — once for the review and once for the typing. A modern AI-assisted workflow drafts the descriptions automatically and uses human reviewers to QC and approve, which is how DecoverAI processed 30,000 documents (including a complete privilege log) in three days for the Tax Credit Investigation case study, with the client achieving a 98% cost reduction relative to the traditional vendor estimate.
DecoverAI replaces the per-GB / per-document / per-hour stack with a single pricing primitive: $60 per gigabyte per month, all-in, with unlimited users and no contracts. That number includes ingestion, OCR, deduplication, AI-powered relevance and privilege classification, Bates numbering, redaction, privilege log generation, full production delivery, and SOC 2 / HIPAA compliant hosting. There is no separate per-document fee, no seat fee, no project-management hourly rate (unless you specifically choose to add embedded LegalOps support at $150 per hour), and no surprise line items at the end of the matter.
The pricing model works because the underlying technology has actually changed. AI-powered classifiers do the responsiveness pass that contract reviewers used to do, at a small fraction of the cost. Auto-generated privilege log entries replace the manual drafting that used to be billed at $10–$15 per entry. Production setup that used to involve a litigation support specialist now runs in the platform with attorney supervision. The cost savings are not magic — they are the consequence of replacing labor-priced steps with infrastructure-priced steps.
For most matters, the practical math works out to $300–$1,200 in total platform cost for the 5–20 GB range that covers the majority of commercial disputes. Larger matters scale linearly: a 100 GB matter is $6,000 per month, a 250 GB matter is $15,000 per month, and so on. No matter the size, the pricing is published, predictable, and easy to defend in a Rule 26(b)(1) proportionality discussion. See the pricing page for the full breakdown, or book a 30-minute demo and we will run your specific matter through the cost calculator on the call.