The pitch from ProactAI sounds familiar at first—turn your existing security cameras into a retail analytics powerhouse, no new hardware required. It's a promise dozens of computer vision startups have made in recent years, from RetailNext in Silicon Valley to Sensormatic's sprawling global network.
But the Bangalore-based startup, which recently closed what it describes as approximately ₹7.5 crore (roughly $900,000) in its first institutional funding round, is making a different bet: that India needs its own vision foundation models, built from the ground up for the chaos and density of Indian retail environments.
The round was led by Le Travenues Technology, the company behind travel booking platform Ixigo—an investment that brings not just capital but the kind of operational know-how that comes from scaling consumer tech across India's fragmented digital landscape. ProactAI had previously raised a friends-and-family seed round, though the company hasn't disclosed the size or timing of that earlier injection.
Building What the Founders Call "Sovereign AI"
At the heart of ProactAI's strategy is what the founding team—engineers with IIT pedigrees who launched the venture in 2024—calls "sovereign AI." The term has become something of a buzzword in Indian tech circles, but here it signals a specific technical choice: developing computer vision models trained on Indian conditions rather than adapting systems built for Western markets.
The flagship product, internally codenamed "Bhaskara," is designed for person re-identification—tracking individuals as they move through a retail space without traditional identifiers like loyalty cards or phone signals. ProactAI claims accuracy exceeding 98%, a self-reported figure that hasn't been independently verified through benchmark testing or peer review.
What the system does provide, according to the company, is a suite of analytics modules layered onto whatever CCTV cameras retailers already have installed. Footfall tracking. Dwell time by zone. Queue length management. Heat mapping of store traffic patterns. Even staff activity monitoring, a feature that can veer into uncomfortable territory depending on how it's deployed.
ProactAI says it's live in 352 retail stores across 31 Indian cities, with 8,448 cameras feeding data into its platform—figures the company reports in its marketing materials. A case study published in March 2026 by AceCloud, a cloud infrastructure provider that hosts some of ProactAI's production workloads, described deployments across "hundreds of retail locations"—phrasing that aligns with the company's own figures, if not quite confirming them independently.
The startup runs lean. LinkedIn pegs the headcount somewhere between 11 and 50 employees. The company's website mentions an appearance on Shark Tank India, the localized version of the entrepreneurial pitch show, though independent confirmation of that episode—or whether a deal materialized on air—hasn't surfaced in publicly available media coverage.
The "No New Cameras" Promise Is Now Table Stakes
Here's the challenge: ProactAI is entering a market where the core value proposition—plug into existing infrastructure, avoid costly hardware upgrades—has become standard fare.
Sensormatic, a veteran in the space, has been pitching existing-camera integration for years. So have RetailNext, CountPort, and LiveReach AI, among others. The retail computer vision market was valued at around $2 billion in 2024, according to Grand View Research, with projected annual growth of 22.6% through 2030. That's attractive. It's also crowded.
What might set ProactAI apart, at least in the Indian market, is the focus on India-specific training data and model tuning. Retail environments in Mumbai or Delhi look different from those in Los Angeles or London—different lighting, different crowd densities, different clothing patterns that affect how well algorithms can track individuals. Whether that localization translates into a defensible competitive moat is the question investors are betting on.
The company describes its philosophy as "India-First, Global-Ready," though plans for international expansion remain vague. No timelines. No specific geographies. Just the suggestion that what works in Bangalore might eventually scale beyond.
Capital for Compute-Hungry Models
ProactAI says it will channel the fresh funding into further development of its vision foundation models—deep learning systems that require serious computational firepower. According to an AceCloud case study, the company has migrated key production workloads to AceCloud's infrastructure, deploying custom configurations that include four NVIDIA L40 GPUs and roughly 2.7 terabytes of storage to handle the torrents of video data streaming in from thousands of cameras.
Those are not trivial resources for a startup at this stage. Training and running vision models at scale burns through both compute cycles and capital, which makes the Ixigo backing potentially more valuable than the dollar figure suggests. Le Travenues Technology knows what it takes to build and operate consumer-facing platforms across India's complex regional markets. That operational DNA could prove useful as ProactAI tries to sign up more retailers.
Whether that translates into faster adoption—and whether retailers ultimately find enough value in the insights to justify ongoing contracts—remains an open question. For now, ProactAI is placing its chips on the idea that homegrown AI, tuned for Indian conditions, can carve out space in a global market increasingly dominated by a handful of well-funded incumbents.
It's a bet that feels both locally specific and universally familiar. Build something better for your home market. Then see if the world follows.
