← Sticker Wall AGENT · DETAIL

Dataminr Real-Time Event Verification Agent: Pentagon $318M Contract Secured

Workflow: The input end ingests public data streams such as social media, news, sensors, and commercial telemetry. Multi-agents co

AGENT

Key Fields

FIELD STAMPS

IndustryFintech

RegionUS

ScaleSME

ChannelOnline

🔧 Workflow

The input end ingests public data streams such as social media, news, sensors, and commercial telemetry. Multi-agents collaborate to run event extraction and cross-verification models, escalating single-source signals to early warnings based on confidence thresholds. Human analysts review-in-the-loop before outputting structured threat briefings and real-time alerts, which are pushed to defense and financial clients via IDIQ task orders in a 24/7 continuous cycle. In August 2026, the Pentagon awarded a multi-year framework contract worth $318 million (per company disclosures as of August 2026).

🛠 Setup Requirements

Requires a public data scraping pipeline, LLMs combined with event extraction models, and at least one dedicated human reviewer/analyst role. Building from scratch takes about 3 to 6 months. The real challenges are data licensing, low-latency pipelines, and suppressing false positive rates to operational levels, rather than the models themselves.

🧰 Toolchain

🔧 Dataminr First Alert Advanced (Agentic real-time cross-verification and event impact prediction)
🔧 Fivecast ONYX (AI data fusion and behavioral change detection)
🔧 Riverside Research RavenEye (Agentic intelligence loop framework)
🔧 Semantic Visions (Entity tagging and early risk signals)

💰 Revenue

① Government and defense clients settle via IDIQ task orders (primary company revenue): The US Department of Defense and others pay per task order. On August 27, 2026, Dataminr was awarded a multi-year framework contract worth $318 million (~$318M), averaging roughly $26.5 million monthly, accounting for approximately 100% of its monthly revenue (company public disclosures as of August 2026); ② Replicator tier—local security teams/think tanks/financial institutions via monthly subscription: Individuals and small teams use open-source OSINT pipelines to deliver real-time threat briefings (pricing and client counts undisclosed, share of total pie unknown); ③ Annual licensing: Following Primer.ai's model of charging military clients such as the US Air Force via annual prepaid licenses, charged annually at undisclosed rates, with overall share unlisted (company financial reporting basis); ④ Opportunity item—commercial-side situational intelligence licenses: Individuals can replicate the pathway of selling public signal tagging and situational intelligence; scale and revenue share are undisclosed.

💸 Cost

Data source licensing and real-time pipelines constitute the primary costs; tool subscriptions and API fees are undisclosed. Fulfilling government contracts also requires passing FAR and DFARS compliance audits and undergoing prolonged free validation periods, resulting in substantial compliance costs.

⏱ Time Investment

Systems operate automatically 24/7; human teams conduct three-shift reviews of high-risk alerts daily. In the early startup stage, a single-person role can initially run the daily briefing closed-loop.

🚀 Getting Started

The first step is to use open-source pipelines (such as GitHub OSINT-MONITOR projects) for real-time warning validation of a single region or event, establishing the closed-loop from data to confidence scoring to human confirmation to briefing; then, using auditable source-chain delivery records, pitch small-scale paid pilots to local security teams or think tanks to build a track record for future entry into government framework contracts.

🔑 Keys to Success

✅ Human-in-the-loop review, rejecting fully automated decision-making to preserve credibility
✅ Real-time cross-verification rather than single-source triggers, suppressing false positive rates to acceptable levels
✅ Government framework contracts like IDIQ and security compliance qualifications determine the revenue ceiling
✅ Vertical scenario first: establish the early-warning closed-loop for a single conflict zone or threat type before horizontally expanding to adjacent domains like finance and supply chain risk

⚠️ 风险

⚠️ High concentration of government clients; framework contracts do not guarantee minimum procurement volumes
⚠️ False positives or latency may impact critical operational decision-making, with liability boundaries difficult to define
⚠️ Litigation risks surrounding data licensing and privacy issues

📌 Real Cases

📌 On August 27, 2026, the Pentagon awarded Dataminr a $318 million contract, describing its First Alert Advanced as agentic real-time verification: automatically cross-validating events, predicting impacts, and generating alerts (DefenseScoop reporting)
📌 Primer.ai (Primer Delta and Primer Command) charges military clients such as the US Air Force via annual prepaid license models, having secured ABMS/JADC2-related contracts (contract ceilings in the hundreds of millions) and SBIR projects for machine-speed content generation and intelligence automation (disclosed on primer.ai official website)
📌 Hangzhou startup MizarVision, during the US-Iran conflict in February 2026, tracked and tagged movements of US warships and aircraft via public signals and sold situational intelligence services to enterprise clients, proving that the OSINT analysis license sales pathway is equally viable on the commercial side (Tencent News June 2026 report)

Sources