# eDiscovery Pricing per GB: How Vendors Bill in 2026 > eDiscovery pricing per GB still anchors most bills, but per-user plans and AI per-document fees are changing the math for legal buyers. By The Circle Desk | Legal Ops | Published 2026-10-24 Source: https://legaltechcircle.com/insights/ediscovery-pricing-per-gb-explained/ eDiscovery pricing per GB still sets most bills: practitioners in a Winter 2026 survey reported processing at ingestion mostly below $75 per GB and basic hosting mostly below $10 per GB per month. But the major platforms no longer publish per-GB rate cards, several have moved to volume commitments or per-user plans, and AI review is increasingly priced per document. The useful question is less "what is the rate" than "what is the unit." ## What Do eDiscovery Pricing per GB Benchmarks Show? The most current public benchmark is the [Winter 2026 eDiscovery Pricing Survey](https://complexdiscovery.com/buyers-guide/a-complete-analysis-of-the-winter-2026-ediscovery-pricing-survey/) from ComplexDiscovery and EDRM, published in February 2026 from 53 responses collected between December 2025 and February 2026. The figures are self-reported practitioner views, not audited transactions, and 92.5% of respondents named the United States as their main market, so treat every number below as reported, directional pricing. For processing at ingestion, 39.6% of respondents reported $25 to $75 per GB and 34.0% reported less than $25 per GB. At completion of processing, the most common answer was less than $100 per GB (37.7%), with 15.1% reporting $100 to $150 and 9.4% above $150. Basic hosting without analytics has become cheap: 54.7% reported under $10 per GB per month and 30.2% reported $10 to $20. Hosting with analytics costs more, with 32.1% reporting $15 to $25 per GB per month and 11.3% above $25. Per-GB billing is fading in some lines. For predictive coding and technology-assisted review, the most common answer (35.8%) was an alternative pricing model rather than a per-GB rate. User licenses most often ran $50 to $100 per user per month (41.5%), but 34.0% reported alternative models. ## How Do the Major Platforms Bill Now? None of the platforms we checked publishes a dollar rate, but each publishes its pricing unit, which is the more useful fact for comparison. Relativity's [pricing page](https://www.relativity.com/platform/pricing/), read October 23, 2026, says Relativity aiR is sold on a committed data volume, pay as you go or on one- or three-year commitments, and that its AI tools, including aiR Review, Privilege and Case Strategy, come with that data volume "with no separate AI licenses and no per-user fees." Relativity announced the rename of RelativityOne to Relativity aiR ahead of its RelFest conference, held September 29 to October 1, 2026, according to its [newsroom](https://www.relativity.com/news-events/relativity-unveils-new-pricing-model/). DISCO's [pricing page](https://csdisco.com/pricing), read the same day, says it bills on processed data with unlimited use of its Cecilia AI included in the per-GB rate, while its Auto Review product is priced per document. Logikcull's [pricing page](https://www.logikcull.com/pricing) lists a storage-based plan billed per GB of data each month with unlimited users, and its FAQ says pay-as-you-go plans carry a 10 GB minimum. Nextpoint goes the other way: its [pricing page](https://www.nextpoint.com/pricing/) advertises flat per-user pricing with unlimited data and no hosting charges. Everlaw's [pricing page](https://www.everlaw.com/pricing/) lists unlimited user licenses and marks several AI features as using credits, without publishing rates. Microsoft is a special case for companies already on Microsoft 365. Its [eDiscovery billing documentation](https://learn.microsoft.com/en-us/purview/edisc-billing) says storage of ordinary Microsoft 365 data in eDiscovery cases is included in the enterprise subscription, while data from non-Microsoft 365 AI applications and Graph API exports for E3-licensed tenants are billed pay as you go through Azure. ## What Does AI Review Cost per Document? The Winter 2026 survey found no single model for generative AI review. Hybrid pricing and per-document pricing each accounted for 28.3% of primary models, per-GB pricing for 11.3%, and per-token pricing for just 5.7%. Among all respondents, the most cited AI per-document tier was $0.26 to $0.50 (20.8%), followed by $0.11 to $0.25 and $0.05 to $0.10 (15.1% each). For comparison, respondents most often reported $0.50 to $1.00 per document for human managed review. The fine print matters. Asked how failed or exception documents are billed in AI review, 39.6% of respondents said they did not know. Among those who did, 18.9% said exceptions go to manual review at standard rates. A low per-document headline can rise quickly if many documents fall out of the AI workflow. ## What Should You Watch Next? Watch whether more platforms follow Relativity and DISCO in folding AI into the base data price, or keep charging separately for AI review as DISCO does with Auto Review. Watch whether per-token costs ever pass through to buyers; the survey found few providers doing that so far. And watch the next ComplexDiscovery survey, since the Winter 2026 edition is the baseline most buyers will use in negotiations this year. ## Common Questions **What is a typical eDiscovery processing price per GB?** In the [Winter 2026 survey](https://complexdiscovery.com/buyers-guide/a-complete-analysis-of-the-winter-2026-ediscovery-pricing-survey/), 73.6% of respondents reported ingestion processing below $75 per GB. **What does eDiscovery hosting cost per month?** The same survey reported basic hosting below $10 per GB per month for 54.7% of respondents, and higher rates when analytics are included. **Does Relativity publish its prices?** No. Its [pricing page](https://www.relativity.com/platform/pricing/) describes committed data volume plans with AI included but lists no dollar rates. **Which platforms charge per user instead of per GB?** Nextpoint advertises [flat per-user pricing](https://www.nextpoint.com/pricing/) with unlimited data. --- Legal Tech Circle (https://legaltechcircle.com/). News and analysis, not legal advice.