In 2014, a young U.S. Army officer named Grant Demaree was in Liberia, watching an Ebola outbreak outrun the machinery meant to stop it. Treatment units were being built with care and precision - in the places the virus had been three months earlier. The plan was not wrong so much as late. By the time a decision worked its way through email threads, slide decks, and rotating in-person meetings, the crisis had already moved on. Demaree, the story goes, sketched the outline of a fix on a napkin. The realization underneath it was blunt: people were dying because decisions were slow.
That napkin is now a company called Onebrief, and in January 2026 it closed a $200 million Series D that valued it at $2.15 billion. It is one of the more quietly consequential defense-tech stories going - a business built not on drones or satellites, but on the least glamorous thing in any large organization: how a staff actually makes up its mind.
The ProblemMilitary staffs are decision engines. They run slow.
Inside the Department of Defense, the unglamorous truth is that wars are planned by staffs - large groups of officers coordinating movement, logistics, intelligence, and intent across many teams and many echelons. For decades, the tools of that trade have been email, PowerPoint, shared drives, and meetings where the participants rotate in and out. A plan is a stack of slides. When one thing changes, someone updates a slide - and then twenty other slides quietly go stale.
Onebrief's pitch starts from a customer's own vocabulary rather than a founder's vision statement. It focuses on three numbers a commander already cares about: decision speed, decision quality, and staff efficiency. According to the company's own research, staffs using the platform operate roughly twice as fast and twice as efficiently as they do with the traditional approach. Whether or not you take an internal figure at face value, the framing is the product's whole personality: pick the metric the buyer believes in, and move it.
The ProductWhat it actually does
Onebrief is collaborative, AI-enabled software purpose-built for military planning. Instead of a slide deck that has to be manually kept in sync, it treats a plan as structured, living data. Change one element and the platform propagates the update across the orders, briefs, and views that depend on it - the difference between editing a document and editing a spreadsheet that everyone is looking at simultaneously. People who have used it tend to reach for the same comparison: it is a bit like Jira or a shared workspace, but for winning wars.
The harder engineering problem is where it lives. Onebrief is deployed across the military's classified networks - NIPR, SIPR, JWICS, and CENTRIX - up to top secret, and it works offline, which matters when the network you are on is a tent in the Pacific rather than an office in Virginia. It handles maps and GIS, imports and exports DoD data formats through an open API, and encrypts to the FIPS 140-3 standard. None of that is a demo feature; it is the price of admission to the buildings where this work happens.
The old workflow
- Plans live in PowerPoint and email
- One change means updating 20 slides by hand
- Meetings with rotating participants
- Versions drift; the "current" plan is a guess
- Decisions lag the situation
The Onebrief workflow
- Plans are structured, living data
- Change once, it propagates everywhere
- Real-time collaboration across echelons
- Version control and a single source of truth
- AI drafts and updates orders and briefs
On top of that sits the newer layer: an AI Co-planner. Onebrief's AI Assist is built on Anthropic's Claude and trained partly on open-source U.S. military doctrine, historical orders, and past staff outputs. Using the live context of a plan and doctrinal templates, it can draft and update orders, plans, and briefs while staying tethered to the commander's intent. It is one of the clearer examples of a large language model being pointed at a genuinely specialized workflow rather than a chatbot.
The TractionWho is actually using it
Onebrief says it is used across U.S. military headquarters worldwide - the Army, Navy, Marine Corps, Air Force, Space Force, and Special Operations Forces - and that it is now integral to three of the four largest operational plans on the planet. It has spread organically across four of the seven geographic combatant commands, including in the Indo-Pacific, which is not a coincidence given the company is headquartered in Honolulu. Its own team is heavy with people who lived the problem: veterans of the Joint Staff, NATO, all three Marine Expeditionary Forces, the 75th Ranger Regiment, and special operations.
The growth pattern is the interesting part. Rather than a single top-down mandate, adoption has moved bottom-up, unit by unit - the company has cited a 5,700% annualized usage-growth figure. Software spreading by word of mouth through classified networks is a strange sentence to write, but it is roughly what happened.
The MoneyFour rounds, one direction
The funding history reads like a company that found product-market fit and then kept pouring fuel on it. A $53 million Series B in 2024 was followed by a Series C that brought total capital to roughly $103 million, then a Series C extension led by Battery Ventures. In 2025 Onebrief crossed a roughly $1.1 billion valuation to become a defense-tech unicorn. Then came the $200 million Series D in January 2026, led by Battery Ventures and Sapphire Ventures, with Salesforce Ventures joining and existing backers General Catalyst and Insight Partners returning.
The Series D came bundled with an acquisition: Battle Road Digital, a company that builds real-time operational decision support using gaming technology. The logic is that planning and wargaming should not be separate tools. Onebrief wants a single system where you draft a plan, wargame it, simulate it, and support the live decision - what it calls an operating system for command. "Integrating with Onebrief allows us to rapidly bridge the gap between simulation and live operations," said Battle Road founder Josh Henderson.
The TimelineFrom napkin to command OS
- 2014The problem, seen firsthandDemaree watches decisions lag the Ebola crisis by months in Liberia and sketches a fix.
- 2019Onebrief is foundedHe starts building collaborative planning software for military staffs.
- 2024Series B and CA $53M Series B and a Series C push total funding to roughly $103M as usage climbs.
- 2025Unicorn statusOnebrief crosses ~$1.1B and expands its AI Co-planner and offline mode.
- 2026$200M Series D at $2.15BBattery and Sapphire lead the round; Onebrief acquires Battle Road Digital.
The MarketPalantir gets the headlines. Onebrief gets the OPLANs.
The obvious comparison is Palantir, and there is overlap in the world of defense data and command-and-control. But Onebrief is aimed at a narrower, deeper seam: the daily craft of staff planning. Its most durable competitor is not a rival startup at all - it is the incumbent workflow of slides, inboxes, and meetings that has run military planning for a generation. Displacing a habit is harder than beating a product, which is part of why the bottom-up adoption curve matters more than any feature list.
There are conditions under which the story gets harder. Defense procurement is slow and political; software that touches classified networks carries accreditation burdens that would sink a lighter company; and an internal "2x faster" figure is a claim, not an audited benchmark. AI that drafts military orders will - and should - face scrutiny about where the human sits in the loop. Onebrief's answer is to keep the AI tethered to doctrine and commander's intent rather than letting it freelance, but that is a design stance to watch, not a settled question.
The TakeawayWhat a builder can steal from this
Strip away the classified networks and the defense context, and Onebrief is a clean case study in a few transferable moves. Build for a market you actually lived in, so you know which problems are real and which are theater. Anchor the entire product on a single metric the buyer already believes in, and make everything ladder up to moving it. Let adoption spread bottom-up from the people doing the work, not top-down from a logo on a contract. And when a large language model finally makes sense, point it at a narrow, high-stakes workflow with real ground truth - here, decades of doctrine - rather than a general-purpose chat box.
The napkin sketch turned out to be right about one thing above all: in a crisis, the scarcest resource is not information or even courage. It is the speed at which a group of people can look at a changing situation and decide, together, what to do next. Onebrief made a company out of that gap - and a $2.15 billion one at that.