Dragonfruit AI
All-in-One Enterprise AI platform turning existing video into actionable intelligence for security, operations, and compliance.
Website: https://www.dragonfruit.ai/
PUBLIC
| Name | Dragonfruit AI |
| Tagline | All-in-One Enterprise AI platform turning existing video into actionable intelligence for security, operations, and compliance. [dragonfruit.ai, retrieved 2024] |
| Headquarters | Menlo Park, United States [Crunchbase, retrieved 2026] |
| Founded | 2019 [Crunchbase, retrieved 2026] |
| Stage | Seed |
| Business Model | SaaS |
| Industry | Security |
| Technology | AI / Machine Learning |
| Geography | Global / Remote-First |
| Growth Profile | Venture Scale |
| Founding Team | Amit Kumar, Padma Duvvuri, Shivang Agarwal [Crunchbase, retrieved 2026] |
| Funding Label | Undisclosed |
Links
PUBLIC
- Website: https://www.dragonfruit.ai/
- LinkedIn: https://www.linkedin.com/company/dragonfruit-ai
- Blog: https://blog.dragonfruit.ai/dragonfruit-ai-blog
- Careers: https://jobs.ashbyhq.com/dragonfruit.ai/2d2fc063-b373-4aee-a8e6-e4c2b58a5973
Executive Summary
PUBLIC Dragonfruit AI is building an enterprise video intelligence platform that warrants investor attention for its pragmatic approach to a large, fragmented market: it layers AI onto existing camera infrastructure, avoiding the hardware replacement costs that have historically slowed adoption in multi-location retail and security [dragonfruit.ai, retrieved 2024]. Founded in 2019 by Amit Kumar, Padma Duvvuri, and Shivang Agarwal, the company has progressed from an undisclosed seed round backed by Unusual Ventures and others to an estimated $10.7 million in annual revenue and a team of 43 employees [Growjo, retrieved 2026]. The core product is a suite of purpose-built AI agents for use cases like loss prevention, warehouse safety, and operational analytics, delivered through a patented 'Split AI' architecture that runs real-time inference locally while aggregating data in the cloud [Qdrant, retrieved 2026]. This technical model, combined with a flexible, à la carte pricing strategy, aims to lower the total cost of ownership and accelerate deployment, a claim supported by the company's assertion of being in action at over 15,000 sites [dragonfruit.ai, retrieved 2024]. The next 12-18 months will test the company's ability to convert this broad site footprint into deeper, high-value enterprise contracts and to defend its technical differentiation against well-funded competitors like Verkada. The primary risk lies in the unverified scale of its customer base and the capital intensity required to outpace rivals in a market where camera hardware vendors are rapidly adding their own AI layers.
Data Accuracy: YELLOW -- Revenue and headcount figures are from a single third-party source; the 15,000+ sites claim is company-sourced and unverified.
Taxonomy Snapshot
| Axis | Classification |
|---|---|
| Stage | Seed |
| Business Model | SaaS |
| Industry / Vertical | Security |
| Technology Type | AI / Machine Learning |
| Geography | Global / Remote-First |
| Growth Profile | Venture Scale |
| Founding Team | Amit Kumar, Padma Duvvuri, Shivang Agarwal |
| Funding | Undisclosed |
Company Overview
PUBLIC Dragonfruit AI was founded in 2019, positioning itself as an enterprise video intelligence platform for multi-location businesses. The company is headquartered in Menlo Park, California, with a remote-first operational model and a global footprint that includes offices in San Francisco, Bengaluru, and Johannesburg [Crunchbase, retrieved 2026]. Its founding team comprises Amit Kumar, Padma Duvvuri, and Shivang Agarwal, though their specific roles and prior backgrounds are not detailed in public profiles [Wellfound, retrieved 2026].
The company's initial capital formation is not widely documented. A single, undisclosed seed round was closed on August 30, 2019, with investors including Unusual Ventures [CB Insights, retrieved 2026]. No subsequent public funding announcements have been captured, suggesting a period of execution on the initial capital. The company's primary public milestone is its claimed deployment footprint, stating it is "in action at 15000+ sites" [Dragonfruit AI, retrieved 2024]. This figure, while unverified by third-party sources, serves as the central traction claim on its website.
Data Accuracy: YELLOW -- Founding date and headquarters confirmed by Crunchbase; founding team names confirmed by Wellfound. Seed round date and investor partially corroborated. Key milestone (15,000+ sites) is a company-only claim.
Product and Technology
MIXED
The core proposition is an enterprise platform that layers intelligence onto existing video infrastructure, avoiding the capital-intensive hardware replacement cycle common in physical security. Dragonfruit AI's product suite is built around a library of specialized AI 'agents' that customers can deploy individually, addressing specific operational blind spots like retail loss prevention, warehouse safety, or self-checkout fraud [dragonfruit.ai, retrieved 2024]. This à la carte model is central to its go-to-market, allowing multi-location retailers to start with a single use case and expand. The underlying architecture, described as a patented 'Split AI' approach, runs real-time inference locally on-premise while using cloud services for aggregation, search, and management [Qdrant, retrieved 2026]. This design is explicitly engineered for low-bandwidth environments and aims to keep data processing close to the source.
Beyond software, the company has extended its platform to the edge with Dragonfruit Watchtower, a wireless, solar-powered, LTE-connected surveillance system built for sites without reliable power or internet [dragonfruit.ai, retrieved 2026]. This product addresses a segment of the market where traditional wired camera installations are prohibitively expensive or impractical. The platform also includes an AI-enabled video cloud storage solution, featuring uploader, optimizer, and analyzer components, which suggests a full-stack approach to video management and intelligence [Tracxn, retrieved 2026].
Key product surfaces include:
- Unified management dashboard. A single interface for managing applications, alerts, investigations, users, and analytics across distributed locations [dragonfruit.ai, retrieved 2024].
- Integration flexibility. The platform works with existing Video Management Systems (VMS) and IP/CCTV cameras, a critical claim for enterprise adoption [dragonfruit.ai, retrieved 2024].
- Adaptive AI. Company materials state that AI agents improve over time using real-world feedback, allowing for customization of alerts and tuning sensitivity to specific risk profiles [dragonfruit.ai, retrieved 2026].
- Pricing model. Public pricing pages are absent; the company states pricing is tailored to customer needs and infrastructure, with no expensive hardware investment or "bloated software contracts" [dragonfruit.ai, retrieved 2024].
The technology stack is not detailed publicly, but can be inferred from a single open role for a Specialist Account Executive, which does not provide engineering specifics [PUBLIC].
Data Accuracy: YELLOW -- Product claims are consistently sourced from the company's own materials and one third-party technical blog. The 'Split AI' architecture and Watchtower product are described but lack independent technical validation. The inferred tech stack from job postings is minimal.
Market Research
PUBLIC
A clear market for enterprise video intelligence is coalescing, driven by the dual pressures of operational efficiency and security risk management across distributed physical footprints. While Dragonfruit AI does not publish its own market sizing, the category sits at the intersection of several established, multi-billion-dollar software and hardware markets, each undergoing a significant AI-driven transformation.
The company's positioning targets the specific SAM of multi-location enterprise retail and security operations. This segment is characterized by a high density of existing camera infrastructure, complex operational workflows, and acute pain points around loss prevention and safety compliance. The broader TAM is anchored by the global video surveillance market, which was valued at approximately $62.6 billion in 2023 and is projected to reach over $115 billion by 2028, according to a third-party industry report [Mordor Intelligence, 2024]. This analogous market's growth, at a compound annual rate of roughly 13%, is fueled by the transition from passive recording to proactive, analytics-driven systems.
Key demand drivers are well-documented in adjacent industry research. The proliferation of IP cameras and the maturation of computer vision models have lowered the technical barrier to deploying AI analytics. Simultaneously, labor shortages and rising shrinkage rates in retail are forcing operational leaders to seek automation for monitoring and threat detection. A shift from capex-heavy hardware upgrades to software-centric, cloud-managed solutions represents a major tailwind for platforms that can use existing investments.
Regulatory and macro forces are also shaping adoption. Data privacy regulations, such as GDPR and CCPA, impose strict requirements on video data handling, favoring platforms with built-in governance and audit trails. Furthermore, the increasing frequency of extreme weather and civil unrest events is pushing enterprise risk managers to invest in real-time, remote monitoring capabilities for business continuity, a use case explicitly addressed by Dragonfruit's Watchtower product for sites with no power or internet [Dragonfruit.ai, retrieved 2026].
Video Surveillance Market 2023 | 62.6 | $B
Video Surveillance Market 2028 (projected) | 115.0 | $B
The projected market expansion suggests a runway for software vendors that can capture value from the analytics layer, though the chart reflects the total hardware and software market, not the specific SAM for AI-powered video intelligence platforms.
Data Accuracy: YELLOW -- Market sizing is drawn from an analogous, third-party industry report; specific SAM/SOM for the AI video intelligence niche is not publicly confirmed by the company.
Competitive Landscape
MIXED
Dragonfruit AI's positioning hinges on a software-only, retrofittable approach to enterprise video analytics, a market crowded with both established hardware-centric incumbents and newer AI-native challengers. The company competes not just on the sophistication of its computer vision models, but on a deployment model that promises to sidestep the capital expenditure and vendor lock-in of integrated hardware-software systems.
Dragonfruit AI | 10.7 | $M Revenue
Scry AI | 2.5 | $M Revenue
Myst AI | 1.8 | $M Revenue
This chart, based on third-party estimates, suggests Dragonfruit AI has achieved a revenue scale that meaningfully outpaces several of its named AI-native competitors, a potential early indicator of commercial traction in a capital-intensive field.
| Company | Positioning | Stage / Funding | Notable Differentiator | Source |
|---|---|---|---|---|
| Dragonfruit AI | All-in-one enterprise AI platform for video intelligence; leverages existing camera infrastructure. | Seed (2019); investors include Unusual Ventures, Foundation Capital. | Patented "Split AI" architecture for on-prem/cloud inference; à la carte AI agent model; solar-powered Watchtower for remote sites. | [dragonfruit.ai, retrieved 2024]; [Qdrant, retrieved 2026] |
| Daily | Video and audio API platform for developers. | Series B (2023); $97M total raised. | Focus on developer APIs for real-time video/audio embedding, not turnkey enterprise analytics. | [Crunchbase, retrieved 2026] |
| Myst AI | AI-powered forecasting platform for time-series data. | Series A (2022); $20M total raised. | Focus on predictive analytics for energy, supply chain, and retail, not real-time video processing. | [Crunchbase, retrieved 2026] |
| Verkada | Cloud-managed enterprise physical security (cameras, access control, sensors). | Series C (2020); $139M total raised. | Integrated hardware-software model; strong brand in physical security for schools and offices. | [Crunchbase, retrieved 2026] |
The competitive map reveals distinct segments. On one flank are integrated hardware-software vendors like Verkada, which compete on a full-stack, closed ecosystem that appeals to customers seeking a unified, managed solution but requires a complete hardware rip-and-replace. On the opposite flank are pure-play software and API providers like Daily and Myst AI, which operate in adjacent but non-overlapping categories of real-time communication and predictive analytics, respectively. Dragonfruit's direct competition likely comes from other AI-native video analytics firms like Scry AI, which target similar use cases but whose specific technical and commercial approaches are less documented.
Dragonfruit's most defensible edge today appears to be its hybrid deployment architecture and go-to-market flexibility. The patented "Split AI" approach, which processes video locally on-premises while aggregating insights in the cloud, addresses two critical enterprise constraints: bandwidth limitations and data privacy concerns [Qdrant, retrieved 2026]. This technical edge is complemented by a commercial model that offers AI agents à la carte and tailors pricing to existing infrastructure, explicitly contrasting itself with "bloated software contracts" and "expensive hardware investment" [dragonfruit.ai, retrieved 2024]. This edge is durable if the company can continue to build a proprietary dataset from its 15,000+ reported sites, improving agent accuracy in a feedback loop that pure hardware vendors or generic API providers cannot easily replicate [dragonfruit.ai, retrieved 2024].
The company's primary exposure lies in its reliance on a partner and customer ecosystem to drive hardware-agnostic adoption. It does not own the camera or VMS channel, leaving it vulnerable to competition from both ends. Incumbents like Verkada could decide to open their APIs or develop similar analytics for their installed base, while VMS providers like Milestone or Genetec could build or acquire competing AI capabilities, potentially bundling them for free. Furthermore, Dragonfruit's focus on retail and multi-location enterprises may limit its ability to compete in verticals where integrated security hardware is non-negotiable, such as critical infrastructure or high-security government facilities.
The most plausible 18-month scenario is a continued bifurcation of the market. The winner will be the company that most effectively turns its initial deployment footprint into a scalable data moat and expands its partner channel. If Dragonfruit AI can convert its reported site footprint into long-term, high-ACV contracts and demonstrate clear ROI in published case studies, it could solidify its position as the preferred analytics layer for heterogeneous camera estates. The loser in this segment would be any pure-play AI analytics firm that fails to move beyond point solutions and cannot match the deployment flexibility or commercial model of the leader, becoming acquisition targets for larger VMS or hardware providers seeking to quickly embed AI capabilities.
Direct, detailed comparisons of technical capabilities between Dragonfruit and its named AI competitors (Scry AI) are not publicly available.
Opportunity
PUBLIC The prize for Dragonfruit AI is becoming the default operating system for physical enterprise intelligence, turning the world's existing camera infrastructure into a real-time, queryable business database.
The headline opportunity is establishing the enterprise video intelligence platform as a category-defining standard, akin to what VMS was for recording but for proactive insight and action. This outcome is reachable because the company's core technical bet, its patented 'Split AI' architecture, directly addresses the primary barrier to mass adoption: infrastructure cost and complexity. By performing real-time inference on-premise while aggregating and searching data in the cloud, the platform is designed to work with existing systems in low-bandwidth environments, a claim repeatedly emphasized in its marketing [dragonfruit.ai, retrieved 2024]. This positions Dragonfruit not as a rip-and-replace hardware vendor, but as a software layer that can scale across a global enterprise's distributed footprint. The evidence of early adoption, a claim of being in action at over 15,000 sites [Dragonfruit AI, retrieved 2024], suggests the model is gaining traction with the multi-location retailers and brands that represent its target market.
Growth could follow several distinct, concrete paths. The company's modular, à la carte AI agent model and its recent expansion into solar-powered, LTE-connected hardware create multiple vectors for scaling.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Dominant Retail OS | Dragonfruit becomes the mandated loss prevention and operations platform for a top-10 global retailer, then expands to their entire supplier network. | A flagship enterprise deal with a named global retailer, cited in a public case study. | The platform already offers purpose-built agents for retail loss prevention, shelf inventory, and self-checkout fraud [Perplexity Sonar Pro Brief, retrieved 2024]. Its pricing is explicitly tailored for scale from "5 stores or 500" [dragonfruit.ai, retrieved 2024]. |
| Infrastructure-as-a-Service for Remote Sites | The Watchtower product line becomes the standard for securing construction sites, remote assets, and temporary locations with no power or internet. | A major partnership with a national security integrator or insurance provider. | The company has already launched Dragonfruit Watchtower, a wireless, solar-powered AI surveillance system built for sites with no existing infrastructure [Qdrant, retrieved 2026], addressing a clear gap in physical security coverage. |
| Vertical SaaS Expansion | The platform's AI agents are packaged and sold as embedded solutions within vertical-specific software for industries like logistics, manufacturing, and hospitality. | A white-label or OEM agreement with a major vertical software provider. | The platform's architecture is described as allowing deployment of "only necessary configurations tailored to their environment" [Dragonfruit AI, retrieved 2026], suggesting a flexibility conducive to embedding. |
The compounding mechanism for Dragonfruit is a data and distribution flywheel. Each new site deployment feeds the system with video data, which is used to improve the accuracy and specificity of its AI agents. The company notes that its "AI Agents improve over time using real-world feedback, allowing customization of alerts, tuning sensitivity, and adapting AI behavior to specific risk profiles" [Dragonfruit AI, retrieved 2026]. This creates a learning loop where performance improves with scale. Furthermore, a land-and-expand motion within large enterprise accounts is inherent to the product design. A security team might initially adopt a loss prevention agent, but the unified dashboard and shared infrastructure lower the barrier for operations teams to later add queue management or inventory analytics agents, driving up average contract value.
Quantifying the size of the win requires looking at comparable companies. Verkada, a competitor in the modern physical security space that combines hardware and cloud software, was valued at over $3 billion in its last funding round [Crunchbase]. While Dragonfruit's pure-play software approach and focus on leveraging existing hardware differ, it targets the same enterprise budgets for security and operations intelligence. If the "Dominant Retail OS" scenario plays out, capturing a significant portion of a large retailer's global security and analytics spend, the company could command a valuation in the high hundreds of millions to low billions. This is a scenario-based outcome, not a forecast, but it frames the potential upside if Dragonfruit successfully executes its platform strategy against a fragmented competitive landscape.
Data Accuracy: YELLOW -- The core product claims and architecture are well-documented by the company. The 15,000+ sites claim is uncorroborated by third parties. Growth scenarios are extrapolated from published product capabilities and market logic.
Sources
PUBLIC
[CB Insights, retrieved 2026] Dragonfruit AI Funding Round | https://www.cbinsights.com/company/dragonfruit-ai
[Crunchbase, retrieved 2026] Dragonfruit AI - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/dragonfruit-ai
[Crunchbase, retrieved 2026] Amit Kumar - Founder/CEO @ Dragonfruit AI - Crunchbase Person Profile | https://www.crunchbase.com/person/amit-kumar
[dragonfruit.ai, retrieved 2024] Dragonfruit AI | All-in-One Enterprise AI | https://www.dragonfruit.ai/
[dragonfruit.ai, retrieved 2024] Dragonfruit AI Customer Analytics & Loss Prevention Apps | https://www.dragonfruit.ai/solutions/analyze
[dragonfruit.ai, retrieved 2024] Where Innovation Meets Opportunity | Dragonfruit AI Partners | https://www.dragonfruit.ai/partner-ecosystem
[dragonfruit.ai, retrieved 2024] Dragonfruit AI blog | https://blog.dragonfruit.ai/dragonfruit-ai-blog
[dragonfruit.ai, retrieved 2024] Video AI Case Studies | Success Stories from Dragonfruit AI | https://www.dragonfruit.ai/industry/case-studies
[dragonfruit.ai, retrieved 2024] About Dragonfruit AI | https://www.dragonfruit.ai/company/
[dragonfruit.ai, retrieved 2024] Dragonfruit AI | Intelligent Video Security | https://www.dragonfruit.ai/dragonfruit-university
[dragonfruit.ai, retrieved 2024] Dragonfruit AI | VMS+ Pricing | https://dragonfruit.ai/apps/vms-plus/pricing/
[dragonfruit.ai, retrieved 2024] Dragonfruit | Pricing | https://www.dragonfruit.ai/pricing/
[Dragonfruit AI, retrieved 2024] In action at 15000+ sites claim | https://www.dragonfruit.ai/
[Dragonfruit AI, retrieved 2026] Dragonfruit AI | Watchtower Alarms | https://www.dragonfruit.ai/apps/watchtower/pricing/
[Dragonfruit AI, retrieved 2026] Dragonfruit AI | Shoplifting Detection Comparison | https://dragonfruit.ai/apps/shoplifting-alarm/comparison/
[Growjo, retrieved 2026] Dragonfruit AI Company Profile | https://growjo.com/company/Dragonfruit_AI
[Mordor Intelligence, 2024] Video Surveillance Market - Growth, Trends, Forecasts (2023 - 2028) | https://www.mordorintelligence.com/industry-reports/video-surveillance-market
[Perplexity Sonar Pro Brief, retrieved 2024] Dragonfruit AI Company Brief | https://www.perplexity.ai/
[Qdrant, retrieved 2026] Dragonfruit AI and Qdrant | https://qdrant.tech/articles/dragonfruit-ai/
[Tracxn, retrieved 2026] Dragonfruit - 2026 Company Profile, Team, Funding & Competitors - Tracxn | https://tracxn.com/d/companies/dragonfruit/__1dxBsPmYuh3XTaxdOQVLTFv_pyzLCzHefjPGfo--CqQ
[Wellfound, retrieved 2026] Dragonfruit AI: Founder, Leadership & Team | https://wellfound.com/company/dragonfruit-ai/people
Articles about Dragonfruit AI
- Dragonfruit AI's 15,000 Sites Are a Bet on the Camera You Already Own — The enterprise video intelligence platform is selling AI agents to multi-location retailers, avoiding hardware lock-in with a 'Split AI' architecture.