Startup Deep Dive : Staqu Technologies — built police trust on under 2.1 million dollars

Staqu Technologies has built a facial and voice biometric database of more than 900,000 criminal records for the Uttar Pradesh police, wired its cameras into the Ram Mandir inauguration and the 2023 G20 Leaders’ Summit, and put its analytics behind nine state police forces. Yet across its first nine years, the Gurugram company raised a total that most Series A rounds in Indian SaaS would beat in a single cheque: somewhere between $1.4 million and $2.1 million.

That gap between the stakes of the work and the size of the balance sheet is the story of Staqu. It sells the same core technology, an AI engine called JARVIS that turns ordinary CCTV feeds into searchable, alertable data, to a supermarket chain watching footfall and to a police control room watching for a wanted face. Founded in 2015, it never raised the kind of money its more famous consumer-tech peers did, and it never chased the kind of user base they did either. Instead it built quietly inside government contracts and enterprise pilots, until a Dubai Police mandate and a UP Police rollout gave it a reason to be taken seriously. What follows is what is actually on the public record about how it got there, and where the record runs out.

Quick facts

What they do

Staqu sells one core product, JARVIS, into two very different buyers. To enterprises, in retail, hospitality, manufacturing, real estate and co-working, JARVIS plugs into existing CCTV networks and turns raw footage into structured alerts: whether a worker skipped a safety step, whether a queue is building at checkout, whether a restricted zone has been entered. To governments, the same underlying computer-vision engine powers Trinetra, a Uttar Pradesh Police tool built to search a criminal database using a face, a voice clip or a plain-language query through its newer Crime GPT layer. The company describes itself as converting “conventional cameras into intelligent systems,” and its own materials claim over 200 enterprise and government clients and more than 100,000 cameras connected worldwide, figures that are company-stated rather than independently audited.

The origin

Atul Rai founded Staqu in 2015 with a background in computer vision and neural networks, joined by co-founders who split the early technical load: one on backend and server infrastructure, one on mobile and native app engineering. The founding insight was not law enforcement at all. The team’s first instinct, as reported by Inc42 in 2018, was to build an AI recommendation engine that could read the images, speech and text on a webpage and contextualise advertising against it without hoovering up personal data, an idea the founders pitched as wanting to build “a Google of recommendation.” It was a crowded, capital-hungry lane, and it was not where Staqu’s technology ended up finding a market. The pivot came through a second, parallel project: an AI face-detection tool for identifying criminals and missing persons, launched in November 2017 under the name ABHED. That tool, not the recommendation engine, is the one that survived.

The struggle years

Staqu’s early years were less a single near-death event than a long stretch of running two businesses on very little capital. The company’s first outside cheque was tiny: ₹5 lakh (about $7,200) from the Indian Angel Network in June 2016, according to Inc42’s reporting. On that kind of runway, the company built and shipped ABHED to state police forces starting in 2017, while its original recommendation-engine ambition quietly fell away from the public narrative. By 2018 the law-enforcement product had become the company’s identity, but the funding had not caught up: Staqu would not raise again in any publicly reported round until April 2022, nearly six years after that first angel cheque. In the interim, the company was already running pilots with the Punjab Police, credited by YourStory and Inc42 with helping identify suspects in more than 100 arrests, and by 2018 had reportedly deployed with eight Indian state police departments, all on a balance sheet a single seed round could describe. That mismatch between the seriousness of the deployments and the thinness of the capital base is the defining fact of Staqu’s first seven years.

• June 2016: first funding, ₹5 lakh (~$7,200), Indian Angel Network — thin capital base for a company about to sell into policing (Inc42).
• November 2017: pivot crystallises — ABHED face-recognition tool launched, moving the company away from its original ad-recommendation pitch (Inc42).
• 2018: reported deployment across eight Indian state police departments and 100-plus Punjab Police arrests attributed to the tool, still pre-institutional funding (YourStory; Inc42).
• 2016 to April 2022: a near six-year gap between the seed cheque and the next disclosed institutional round, despite an international police mandate landing in between (Crunchbase; Entrepreneur India).

The turning point

The event that changed Staqu’s credibility, if not immediately its bank balance, was a foreign government mandate. In January 2018, Dubai Police selected Staqu’s predictive-policing technology after reviewing solutions from more than 677 startups, then signed a memorandum of understanding for a system pitched as helping identify criminals and flag risk before an incident occurred. Going from one of 677 applicants to the sole Indian startup with an active Dubai Police MoU is the kind of numeric before-and-after that is easy to verify and hard to dismiss: it is cited consistently across Inc42’s original reporting and subsequent coverage of the deal. What it bought Staqu was not a windfall, disclosed contract values were never published, but a reference customer that made the case to Indian state governments considerably easier to make. The domestic state police relationships that followed, and the National Startup Award Staqu won in the Citizen Security category in 2020, trace back to the credibility that one selection created.

The money behind it

• June 2016 — Seed: ₹5 lakh (~$7,200) from Indian Angel Network, a syndicate of individual angel investors rather than an institutional fund (Inc42).
• April 2022 — Pre-Series A: ₹11 crore led jointly by Mount Judi Ventures and listed security-services group SIS Limited; Indian Angel Network took a partial exit in the same round, and Staqu ran a separate ₹1.5 crore ESOP buyback alongside it (Entrepreneur India; IndianWeb2, both citing the company’s own announcement).
• Backer roles: Indian Angel Network supplied the earliest capital and stayed on for six years before partially cashing out in 2022; Mount Judi Ventures and SIS Limited brought the 2022 round’s institutional weight, with SIS in particular a strategic fit given its core business is physical security staffing and services, the exact buyer channel JARVIS sells into.
• Total raised: reported as $1.44 million-plus across three rounds by Inc42, and as $2.11 million across five rounds from 71 investors by Crunchbase and CB Insights; the gap likely reflects smaller, undisclosed angel cheques bundled into the Indian Angel Network syndicate that do not appear as separate named rounds in every database.
• Valuation: not disclosed in any of the sources reviewed for either the 2016 or 2022 rounds.

How it makes money

Staqu runs two revenue lines off one technology stack. On the enterprise side, JARVIS is sold as a subscription or licence to companies that already own CCTV infrastructure, retail chains, manufacturing plants, real estate operators, co-working spaces, and monetises the analytics layer rather than the cameras themselves, which keeps hardware capital expenditure off Staqu’s own books. On the government side, contracts with state police departments for Trinetra and Crime GPT follow a more typical GovTech pattern: a build-and-deploy phase followed by annual maintenance and support billing, with sales cycles and payment timelines set by state government procurement calendars rather than by Staqu. Neither a published take rate nor a disclosed enterprise-versus-government revenue split exists in any source reviewed, so the precise margin structure and which line actually carries the business cannot be stated with confidence; what is verifiable is that government work brings prestige and reference value, per the Dubai Police case, while enterprise contracts likely offer steadier, faster-paying revenue, a common pattern among Indian companies that sell into both public and private-sector security budgets, though this specific inference is not itself sourced to a Staqu filing.

The numbers

Consolidated, audited multi-year financials for Staqu Technologies Private Limited sit behind paid registries (Tofler, Tracxn) as of September 2026, and the free previews of those registries returned inconsistent, unreliable figures for profit and loss that could not be corroborated, so they are omitted here rather than repeated. The two revenue figures below are attributed to a single aggregator, Inc42, and are the only ones this research could independently reconcile across more than one page of that source.

• FY25 revenue: ₹19.6 crore ($2.0 million), up 40.7% year-on-year (Inc42, 2026).
• FY24 revenue: ₹14.0 crore, the base period for that growth figure (Inc42, 2026).
• Profit or loss for either year: not verifiable from any freely accessible source at the time of writing; company financial statements are filed with the Ministry of Corporate Affairs but not reproduced in full on the aggregator pages reviewed.
• Separately, BW Disrupt reported the company describing a 100% year-on-year revenue increase around the close of a prior fiscal year (BW Disrupt, December 2022), a company-stated figure that could not be tied to a specific crore amount and is noted here only as a directional data point, not a verified number.

Where the money comes from

Staqu does not publish a formal revenue-by-segment breakdown, so the split below is built from named clients and deployments rather than disclosed percentages.

• State law enforcement: nine state police departments as customers, including Uttar Pradesh, Punjab, Rajasthan, Bihar, Chhattisgarh, Haryana and Uttarakhand, per Staqu’s own disclosure to investor Mount Judi Ventures at the time of its 2022 funding round.
• Marquee public-security events: security analytics deployed at the Ram Mandir inauguration in Ayodhya (January 2024), the G20 Leaders’ Summit (September 2023), and IPL matches at Bengaluru’s M. Chinnaswamy Stadium, per the company’s own about-us disclosures.
• Enterprise clients: more than 100 companies reported as of the 2022 funding round, named accounts including Rebel Foods, JK Cement, WeWork, GMR and Tata group entities (Mount Judi Ventures portfolio page, 2022); more recent company materials add Chaayos, Embassy, Metro, Crocs, Olive and Borosil, among others.
• The surprise: a company whose public identity is built on policing headlines earns a meaningful share of its named client list from retail and hospitality footfall and safety-compliance analytics, a far less dramatic business than the facial-recognition story that generates most of its press coverage.

The risks

• Facial-recognition misidentification and legal exposure: a joint investigation by The Wire and the Pulitzer Center found that Delhi Police’s facial-recognition stack, built with vendors including Staqu, Innefu Labs, Amped Software and Pelorus, treats an 80% similarity score as a “positive” match, a threshold flagged by Medianama and the Deccan Herald as loose enough to misidentify people with no criminal record; the same investigation documented a wrongful arrest that led to more than four and a half years of pre-trial incarceration for one individual, though it attributes this to the Delhi Police system generally rather than to any single vendor’s algorithm by name.
• Regulatory vacuum: The Wire’s investigation and multiple explainers (The News Minute; Medianama, 2026) note that India has no dedicated statute or binding guideline governing police use of facial-recognition technology, which leaves vendors like Staqu operating in contracts that carry reputational and legal risk if a deployment is later challenged in court, as has already happened with Delhi Police’s system.
• Government-contract dependency and payment cycles: a company built primarily on state-police and marquee-event contracts is exposed to state budget cycles, change of government, and the long payment timelines typical of Indian public-sector procurement; the nearly six-year gap between Staqu’s 2016 seed round and its 2022 pre-Series A, despite an active international police MoU in between, is consistent with a business that had revenue-generating government work but limited access to fast institutional capital in that period.

The takeaway

Staqu’s arc argues against the idea that credibility has to be bought with capital. It reached a foreign police force, a national election deployment and nine state governments while running on a funding base smaller than many seed rounds, because the thing it was selling, a working system that a buyer with high stakes and low tolerance for error could actually pilot, did more convincing than a large treated balance sheet would have. The cost of that path shows up elsewhere: in a company whose growth is well short of the scale its client list might suggest, and in a business whose core product now sits inside one of the more contested debates in Indian technology policy, the use of facial recognition by police with no dedicated law to govern it. Building trust with government buyers and building a defensible, well-capitalised business have turned out to be two different projects, and Staqu’s record so far is a case study in doing the first while the second still lags behind.

Frequently asked questions

What does Staqu Technologies do?

Staqu builds an AI video and audio analytics platform called JARVIS that connects to existing CCTV cameras to generate alerts and searchable data, sold to enterprises for safety and operations monitoring and to state governments for policing tools such as Trinetra and Crime GPT.

Who founded Staqu, and when?

Staqu was founded in 2015 in Gurugram. Its current leadership lists Atul Rai as CEO, with Anurag Saini and Pankaj Sharma as co-founders; earlier press coverage from 2018-19 also named Chetan Rexwal and Abhishek Sharma among the founding team.

How much funding has Staqu raised, and who backed it?

Reported totals range from $1.44 million-plus (Inc42) to $2.11 million (Crunchbase, CB Insights) across a seed round from Indian Angel Network in June 2016 and a ₹11 crore pre-Series A in April 2022 led by Mount Judi Ventures and SIS Limited. No valuation has been publicly disclosed for either round.

What are Trinetra and Crime GPT?

Trinetra is an app built with Uttar Pradesh Police that uses facial recognition, voice recognition and gang-network analysis against a criminal database reported to hold biometric and identity data on more than 900,000 individuals; Crime GPT, launched in 2024, is a generative-AI layer on top of Trinetra that lets officers query the same database using natural written or spoken language.

Is Staqu Technologies listed, and what is it worth?

No. Staqu is a private company with no stock exchange listing or IPO filing found as of September 2026. Its most recent disclosed revenue is ₹19.6 crore in FY25; no market valuation has been made public.

Sources

Figures are as of September 2026. Currency converted at $1 ≈ ₹96.0 as of 18 September 2026 (Trading Economics).

• Inc42, “Staqu — Funding, Revenue & Investors,” accessed September 2026
• Inc42, “How Staqu Is Making A Case For AI In Policing, And Also Building Better Recommendation Engines,” July 2018
• YourStory, “How this Gurugram startup is helping police, Indian army catch bad guys using AI,” April 2019
• Entrepreneur India, “Staqu Raises INR 11 Crore In Pre-Series A Funding, Led By Mount Judi Ventures And SIS,” April 2022
• IndianWeb2.com, “Staqu Raises INR 11 Cr in Funding Led by Mount Judi Venture and SIS; IAN Makes Partial Exit,” April 2022
• Mount Judi Ventures, portfolio page for Staqu Technologies, accessed September 2026
• Crunchbase, Staqu Technologies company profile, accessed September 2026
• CB Insights, Staqu Technologies financials page, accessed September 2026
• Biometric Update, “Staqu’s biometric Crime GPT helps Uttar Pradesh police take a byte out of crime,” March 2024
• The Wire / Pulitzer Center, “As AI Took Over Policing in Delhi, Who Bore the Brunt?,” accessed September 2026
• Medianama, “Delhi Police tells SC its facial recognition targets only people with criminal records. But is everyone scanned first?,” August 2026
• Deccan Herald, “80% accuracy in facial recognition database is a positive match for Delhi police: report,” accessed September 2026
• Staqu Technologies, “About Us,” company website, accessed September 2026
• BW Disrupt, “Staqu Registers 100% Growth,” December 2022

Found an error? Write to us and we’ll correct it in the open, dated, on the piece.