Deep Sentinel

COMPELLING CPS 33

AI-powered 24/7 security system that delivers personalized guard-like protection at customers' properties with live intervention within seconds.

United States·Founded 2016·~62 emp·PRIVATE · deepsentinel.com ↗ ↓ JSON ↓ MD
Researched 2026-02-18 ● Current
Deep Sentinel — robotics.press intelligence card

Deep Sentinel has built a differentiated AI-plus-human-in-the-loop remote guarding platform with credible momentum—doubling bookings YoY, raising a $15M Series B, and achieving ~50% of sales from BYOC—positioning it well in the underserved proactive security segment. However, the absence of independently verified performance data, undisclosed unit economics for its labor-intensive guard model, and material cybersecurity gaps (no auto-updates, no traditional 2FA) temper conviction and keep it below CONTENDER status until these are resolved.

Moat NARROW

- Proprietary edge AI hub with 50+ fps on-device processing creates latency and privacy advantages over cloud-only competitors - Integrated human-in-the-loop guard operations with two-way audio intervention—difficult to replicate without building both AI and staffed monitoring infrastructure simultaneously - BYOC platform approach creates switching costs once customers integrate existing camera estates under Deep Sentinel's monitoring umbrella - Accumulating proprietary training data from real-world security incidents improves AI model performance over time, creating a data flywheel advantage

Management ADEQUATE

CEO David Selinger demonstrates consistent strategic vision across product announcements, driving coherent expansion from proprietary hardware to platform (BYOC) and new modalities (SentinelNow, mobile trailer). However, the broader leadership bench—heads of operations, AI/engineering, enterprise sales, and security/compliance—is not disclosed in available materials, which is concerning given the operational centrality of guard workforce management and the cybersecurity gaps that persist.

Financials OPAQUE
Bull Case

BYOC accounted for nearly 50% of sales by mid-2025, signaling a successful pivot from hardware-centric to platform-led growth with lower customer acquisition friction and broader TAM

Company reports doubling sales bookings YoY, outpacing the 20-30% industry growth rate, with an oversubscribed $15M Series B from credible investors including Intel Capital and Shasta Ventures

Edge AI hub processing at 50+ fps enables low-latency, privacy-preserving triage that should control guard labor costs while maintaining rapid response—a structural advantage over cloud-only competitors

Product portfolio expansion into SentinelNow (panic button for worker safety), mobile surveillance trailers, and Gen V Hub opens new verticals (construction, logistics, retail EHS) and budget categories beyond traditional security

Proactive deterrence model with claimed 30-second police contact and 100% verified dispatch addresses a genuine market pain point—false alarm fatigue and slow response from legacy alarm providers

Mobile surveillance trailer with solar/battery/cellular enables infrastructure-free deployment, capturing temporary and underserved site segments that fixed systems cannot address

Bear Case

No independently verified performance data exists in the public domain—30-second police contact, 100% verified calls, and crime reduction claims are all self-reported marketing assertions

Labor-intensive guard model creates inherent margin pressure and scale friction; guard-to-camera ratios, utilization metrics, and cost per monitored stream are entirely undisclosed

Critical cybersecurity gaps—no automatic firmware updates, no traditional 2FA, no bug bounty or formal VDP—are material barriers to enterprise adoption and represent fleet-wide vulnerability risk

No revenue, gross margin, churn, or retention data has been disclosed; the $38M total funding and 62-employee base suggest the company is still pre-scale with uncertain path to profitability

References to aggressive deterrents (pepper blasts, smoke bombs) raise unaddressed regulatory, liability, and insurance questions that could create legal exposure

No named customer case studies, deployment counts, or audited outcome metrics are publicly available, creating a significant evidence gap for enterprise procurement cycles

Key Risks

Guard labor scalability: if AI triage is insufficient during event volume spikes, SLAs degrade and margins compress, threatening the core value proposition

Cybersecurity posture: no automatic firmware updates and lack of traditional 2FA are enterprise deal-breakers that could stall commercial growth if not remediated

Capital sufficiency uncertainty: with $38M total funding, 62 employees, and no disclosed revenue or burn rate, runway and path to profitability cannot be assessed

Regulatory and liability exposure from aggressive deterrent features (pepper blasts, smoke bombs) without publicly documented compliance frameworks

Competitive response: major alarm companies (ADT, Vivint) and tech giants (Ring/Amazon, Google Nest) could replicate AI-verified monitoring with vastly greater resources and installed bases

Customer concentration risk: no disclosed customer base diversity or retention metrics; loss of key accounts could materially impact a company of this scale

Catalysts

Publication of third-party audited performance benchmarks (response times, false alarm rates, dispatch outcomes) would materially de-risk the investment thesis and accelerate enterprise adoption

Enterprise security hardening milestones—auto-updates, MFA/SSO, SOC 2 certification—could unlock large commercial accounts currently blocked by security requirements

BYOC channel partnerships with major camera OEMs or VMS platforms could dramatically accelerate distribution and validate the platform model

Successful mobile trailer fleet deployments with measurable ROI data could open a high-velocity lead generation channel feeding permanent system sales

Next funding round or strategic partnership with a major security integrator would validate market positioning and provide growth capital for scaling guard operations

Irreplaceability 3
Market Weight
Tech Differentiation
Operational Deployment
Strategic Momentum
Ecosystem Influence
Coverage Necessity
Fin. Valuation
Fin. Revenue
TypeStandard Research
Published2026-02-18
Length4,060 words · 17 min read
Sources31 sources cited

Generated by automated research. Cross-reference with primary sources before investment decisions.

Deep Sentinel Proprietary Cameras (Wireless and PoE variants) Sensor · FIELDED
└─ Wireless and Power-over-Ethernet camera options designed for integration with the Deep Sentinel AI Hub for video surveillance and threat detection. Wireless variant is DIY-friendly for residential and SMB deployments. PoE variant typically requires professional installation via Deep Sentinel's installer network. Both variants feed into the AI Hub for local edge processing before streaming events to remote guards.
AI Hub Fixed · FIELDED
└─ Local on-premises processing unit performing high-speed AI analysis at 50+ fps for video event filtering and edge-based threat detection before streaming to remote guards. Performs on-device AI triage to limit the volume of events escalated to human guards, thereby controlling labor costs and improving response speed. Edge processing may improve privacy by limiting cloud uploads to events of interest only. On-device vs. cloud fallback behavior during network degradation is not publicly documented.
Mobile Surveillance Trailer Fixed · FIELDED
└─ Solar-powered, battery-backed mobile surveillance platform integrating Gen V Hub, four wide-area cameras, one PTZ camera, dual solar panels, and cellular connectivity for rapid deployment to temporary or infrastructure-limited sites. Designed for construction sites, parking lots, logistics yards, pop-up events, and disaster zones. Provides elevated vantage points with no dependency on site power or fixed infrastructure. Pairs edge AI processing with guard intervention claimed within seconds. Capex-intensive product line with logistics and maintenance considerations including trailering, repositioning, battery/solar performance, and vandalism hardening.
BYOC (Bring Your Own Camera) Integration Platform Software · FIELDED
└─ Platform enabling integration of third-party PoE cameras from various manufacturers into Deep Sentinel's AI monitoring and live guard service ecosystem via RTSP/ONVIF support. Allows customers to onboard existing third-party PoE cameras under Deep Sentinel's AI and live guard monitoring umbrella without rip-and-replace of existing hardware. Represents a strategic shift from hardware-centric to platform-led growth. Series B proceeds (June 2025) earmarked in part for scaling the BYOC program. Enables channel partnerships with camera OEMs and integrators. A compatibility matrix and performance differentials across vendors are not publicly documented.
Deep Sentinel Mobile App Software · FIELDED
└─ End-user application enabling live and recorded footage viewing, two-way audio communication with visitors, and multi-user account management with permission controls. Supports both homeowners and businesses. Authentication requires mandatory passwords; classic 2FA is not supported, though device-based MFA via touch ID is available on supported devices. This authentication posture has been flagged as a risk for enterprise deployments by a 2026 third-party review.
LiveSentinel Monitoring Service Software · FIELDED
└─ AI-enabled remote guarding platform combining edge-based computer vision with human guards for real-time threat verification, two-way audio intervention, and police escalation. Processes video at 50+ fps with claims of contacting law enforcement within 30 seconds. Staffed by trained guards who engage potential intruders via two-way audio, activate sirens and lighting deterrents, and escalate to law enforcement. Company marketing references additional deterrent options including pepper blasts and smoke bombs for specific commercial deployments, though independent corroboration of availability, regulatory context, and adoption breadth of these chemical/smoke deterrents is not available in reviewed sources. Performance claims (30-second police contact, 100% verified calls) are marketing figures without independently published benchmarks as of February 2026.
Gen V Hub Fixed · FIELDED
└─ Enterprise-grade AI hub variant designed to lower barrier to entry for enterprise deployments with improved processing and integration capabilities. Announced alongside SentinelNow and the Mobile Surveillance Trailer as part of Deep Sentinel's enterprise expansion strategy. Serves as the processing hub integrated into the Mobile Surveillance Trailer. Designed to improve processing capability and reduce barriers to enterprise adoption compared to the standard AI Hub.
SentinelNow Handheld · FIELDED
└─ Physical push-button panic device enabling employees to instantly connect with trained Deep Sentinel guards for live camera review, verbal intervention, and emergency escalation during and outside business hours. Extends Deep Sentinel's use case from perimeter and intrusion monitoring into personal safety and workplace incident response. Targets retail associates, office workers, property managers, and other at-risk employees. Guards relay real-time situational information to first responders including suspect count and description, visible weapons, crime details, and medical needs. Broadens addressable market into EHS (environment, health, and safety) budgets.
David Selinger CEO
Deep Sentinel Press Contact
Behavioral analytics L3 · Area Monitoring
C2 / Fleet Management L2 · Autonomy & Software
Multi-sensor fusion L3 · Visual Detection
Threat classification L3 · AI / Analytics
AI / Analytics L2 · Autonomy & Software
Computer vision L3 · AI / Analytics
Patrol & Surveillance L1
Autonomy & Software L1
Detection L1
Anomaly detection L3 · Perimeter Patrol
Area Monitoring L2 · Patrol & Surveillance
Command and control L3 · C2 / Fleet Management
Wide-area surveillance L3 · Area Monitoring
Visual Detection L2 · Detection
Perimeter Patrol L2 · Patrol & Surveillance
Thermal imaging L3 · Visual Detection

News & Analysis

1