Startup Deep Dive : SigTuple — the India-built AI microscope that cleared the US FDA on under Rs 10 crore of revenue

In September 2023, a digital microscope built in Bengaluru cleared the United States Food and Drug Administration — the first integrated hardware-plus-AI medical device in digital microscopy to come out of India, as the company and trade press reported. Three years earlier, the same company had lost two of its four founders, including its chief executive, and had told staff that raising growth capital was not realistic in the near term.

That contradiction is SigTuple in one line: a regulatory feat that puts it in a global handful of companies, sitting on top of a business that still booked only ₹9.5 crore of revenue in FY25 against a net loss of ₹17.2 crore (Inc42, citing regulatory filings). This is the story of how a company chasing a very hard problem — teaching machines to read a blood smear the way a pathologist does — survived long enough to matter, and why the hardest part still lies ahead.

Quick facts

What SigTuple does

SigTuple sells AI-powered microscopy to pathology labs, hospitals and clinics. Its flagship is the AI100, a robotic digital microscope that images a physical slide and lets AI models read it, so that the routine, high-volume part of microscopy no longer needs a skilled pathologist standing at the eyepiece. The core applications run on that hardware:

• Shonit: peripheral blood smear analysis — white-blood-cell differential count, red-blood-cell morphology and platelet assessment (company product pages).
• Shrava: urine sediment analysis — detecting cells, casts, crystals and organisms (company product pages).
• Manthana: the AI/ML platform that classifies the imaged data; Mandara handles cloud archiving and remote review (company site and technical specs).

The AI100 is CE-certified for in-vitro diagnostic use, and its Shonit blood-smear application holds US FDA 510(k) clearance (K221309, September 2023). The pitch is scale: a machine reviews and pre-classifies slides, a pathologist reviews the flagged cases remotely, and one specialist can cover many more sites than a hub-and-spoke lab network otherwise allows.

The origin

SigTuple began with a death. Chief executive Tathagato Rai Dastidar has said the company grew out of losing his father to a delay caused by misdiagnosis — the kind of gap that AI reading medical images was meant to close. Dastidar is not a typical health-tech founder: a B.Tech and PhD in computer science from IIT Kharagpur, with technical leadership stints at Yahoo, Gracenote and American Express, where he was Director of its Big Data Labs from 2012 to 2014 (Crunchbase; sciAstra profile).

In 2015 he teamed up with Rohit Kumar Pandey, Apurv Anand and Pranat Bhadani. The founding insight was blunt: a huge share of diagnostic work is a trained human looking at visual data — a slide under a microscope, a retinal image, a chest X-ray — and much of that pattern recognition can be automated, cutting manual error and making a scarce pathologist reachable by telemedicine. Dastidar ran the company for its first two years, then, after closing the Series A late in 2016, stepped back to lead technology while Pandey took the chief-executive seat (Entrackr). The early platform, Manthana, was deliberately broad, aimed at blood, urine, semen, retinal and chest-X-ray data at once.

The struggle years

Breadth turned out to be the trap. Building regulator-grade AI across several imaging modalities at once is slow and expensive, and revenue did not follow. Two documented near-death moments define this period.

The first came in June 2020. After five years, co-founders Rohit Kumar Pandey — then CEO — and Apurv Anand, who had run technology and operations, both left the company. The founders cited the toll on health and family, but the context was financial: as reported at the time, SigTuple concluded that raising growth capital was not possible in the near-to-medium term, so it chose to tighten its belt and cut spending. Dastidar, who had been CEO in the first two years, stepped back into the role (Entrackr; YourStory; Inc42, June 2020).

The second shows up in the accounts. Revenue did not just stall — it collapsed. Operating revenue fell 74.7%, from ₹4.9 crore in FY21 to ₹1.2 crore in FY22, while the FY22 net loss ran to ₹29.1 crore (Inc42, citing filings). A company nearly a decade old was, at that point, burning roughly twenty rupees for every one it earned. The response was to narrow: concentrate the hardware-plus-AI bet on microscopy, lead with the blood smear, and push a single product all the way through hard regulatory clearance rather than spreading thin across modalities.

The turning point

The single event that changed SigTuple’s standing was the US FDA 510(k) clearance for AI100 with Shonit, granted in September 2023 (FDA record K221309). Peripheral blood smear examination is the reference test for hematological disorders — blood cancers, anemias, infections — and it had stayed stubbornly manual, needing a skilled pathologist on site. SigTuple’s clearance covered a device that images the smear and uses AI to locate and characterise cells for the WBC differential and RBC and platelet morphology.

The numbers on either side of that milestone tell the story. Before it, FY22 revenue was ₹1.2 crore and the loss was ₹29.1 crore. After it, operating revenue recovered to ₹5.81 crore in FY24 (Tofler) and ₹9.5 crore in FY25, up 64.1% year-on-year, while the net loss narrowed to ₹17.2 crore (Inc42). Clearance did not make SigTuple profitable. What it did was rare and durable: it put an India-built device into the small global group cleared to sell an integrated hardware-and-AI microscopy product in the United States, and gave the company a credential that no amount of marketing can substitute for.

The money behind it

SigTuple has been backed by marquee investors from the start, but the cheques got smaller as the market for deep-tech health hardware cooled. The documented rounds:

• Seed, October 2015: about $0.74 million, led by Accel with others (Inc42).
• Series A, February 2017: $5.8 million, with Accel and IDG Ventures (now Chiratae Ventures) among backers (Inc42; press reports).
• Series B, June 2018: $19 million (about ₹130 crore), led by Accel Partners and IDG Ventures, with Endiya Partners, pi Ventures, Flipkart co-founder Binny Bansal, Axilor Ventures and VH Capital; venture debt from Trifecta Capital (Inc42; YourStory).
• Series C, April 2019: $16 million, led by Trusted Insight with Accel, Chiratae and pi Ventures; Binny Bansal joined the board (Entrackr).
• Series C extension, February 2023: about $4.19 million, with Endiya Partners and Accel (Inc42; Entrackr).
• Series C extension, August 2024: about ₹33 crore ($4 million), led by SIDBI Venture Capital with Endiya Partners (Entrackr).

Total capital raised is put at roughly $49.7 million by Inc42 and about $54.7 million by Crunchbase and Tracxn, which count angel and grant tranches differently; Entrackr described the cumulative figure as over $40 million at the 2024 round. Inc42 estimates the August 2024 round valued the company at about $29.37 million (roughly ₹282 crore) — a figure the company has not confirmed, and one to read as a data-provider estimate rather than a disclosed post-money valuation. The trajectory is telling: the big money arrived in 2018-19, and the last two rounds were small extensions to keep the company funded through clearance.

How it makes money

SigTuple’s model has two revenue engines wrapped around one device:

• Hardware: sale or placement of the AI100 digital microscope at labs, hospitals and diagnostic chains.
• Per-report / per-test charges: a recurring fee for each analysis run through the AI, described in earlier reporting as a per-report charge on a hub-and-spoke model (Entrackr, 2019).

The margin logic is where the business either works or does not. The AI and cloud layer (Manthana, Mandara) is the high-margin, scalable part; the microscope hardware and field support are the capital-heavy, lower-margin part. The part outsiders get wrong is assuming the device sale is the business. It is closer to a razor-and-blade shape: the recurring per-report revenue, plus the ability to have one remote pathologist cover many machines, is what is supposed to turn a slow hardware roll-out into a scalable diagnostics network. Until installed devices and test volumes are large, though, fixed R&D and regulatory costs dominate — which is exactly what the loss figures show.

The numbers

Reported figures for Sigtuple Technologies Private Limited (₹ crore; operating revenue and net loss as compiled from filings by Inc42 and Tofler):

Two things stand out. Revenue is real but small — under ₹10 crore even in the best year on record — and it is volatile, having fallen by three-quarters in FY22 before rebuilding. And the loss, while still large relative to sales, is moving the right way: from ₹29.1 crore in FY22 to ₹17.2 crore in FY25, even as total expenses stayed elevated at about ₹26.8 crore in FY25 (Inc42). The FY25 net margin is roughly minus 180% — a company still spending far more than it earns, but spending less to earn more than it used to.

Where the money comes from

SigTuple does not publish an audited segment or geography split, so the honest picture is directional:

• By product: the blood-smear application (Shonit) is the lead, being the one with both CE and US FDA clearance; urine sediment (Shrava) is the second application on the same AI100 platform (company product pages).
• By geography: the core installed base is India, but the entire strategic point of the 2023 FDA clearance and the CE mark is access to the United States and Europe — the markets where a per-report AI microscopy service can command far higher prices than in India.
• By revenue type: a mix of device sales and recurring per-report charges, with the recurring layer meant to grow as the installed base does.

The surprise is the mismatch between reputation and revenue. SigTuple carries a global-first regulatory credential and blue-chip investors, yet its total revenue is smaller than a single mid-sized diagnostic lab’s. The credential is an option on a large future market; it is not yet a large business.

The risks

• Cash and dilution: the company remains loss-making (₹17.2 crore net loss in FY25) and its last two rounds were small extensions, not fresh growth capital. If revenue does not scale faster, it faces either a hard raise in a cautious health-tech market or further cost cuts — the same squeeze that preceded the 2020 founder exits.
• Commercialisation gap: regulatory clearance is necessary but not sufficient. Turning an FDA 510(k) into US and European sales means clinical validation, distribution, reimbursement pathways and a sales force in markets where incumbents such as established hematology-analyser makers are entrenched. Clearance in September 2023 has, so far, lifted revenue only to ₹9.5 crore.
• Concentration and adoption: the business now rests heavily on one modality (microscopy) and, within it, one flagship application (blood smear). Pathologist trust, workflow integration and the medico-legal question of who signs off on an AI-assisted report all gate adoption, and any one of them slowing down hits the whole revenue base.

The takeaway

SigTuple is a case study in the difference between a hard technical win and a business. Narrowing from a broad AI-diagnostics platform to a single device pushed all the way through US FDA clearance was the right call — it is what kept the company relevant after its founders left and its revenue collapsed. But clearance is a starting line, not a finish. The transferable lesson is that in regulated deep tech, the moat and the milestone are often the same thing, and getting there can consume so much time and capital that the commercial engine still has to be built afterwards, from a standing start, against the clock of the next fundraise. SigTuple has done the part almost no one can do. Whether it can do the part everyone underestimates — selling it — is the open question.

Frequently asked questions

What does SigTuple do?

SigTuple builds AI-powered automated microscopy for pathology. Its AI100 digital microscope images slides and uses AI to analyse them, with applications for peripheral blood smear (Shonit) and urine sediment (Shrava), aimed at labs, hospitals and clinics.

Who founded SigTuple and who runs it now?

It was founded in 2015 by Rohit Kumar Pandey, Tathagato Rai Dastidar, Apurv Anand and Pranat Bhadani. Pandey and Anand left in 2020; Tathagato Rai Dastidar, an IIT Kharagpur PhD and former American Express Big Data Labs director, is the current CEO.

Is SigTuple profitable?

No. It reported a net loss of ₹17.2 crore in FY25 on revenue of about ₹9.5 crore, though the loss has narrowed from ₹29.1 crore in FY22 (Inc42, citing filings).

What is the significance of SigTuple’s US FDA clearance?

In September 2023 its AI100 with Shonit received US FDA 510(k) clearance (K221309), reported as the first integrated hardware-plus-AI medical device in digital microscopy from India, opening access to the US market.

How much has SigTuple raised and who are its investors?

Estimates range from about $49.7 million (Inc42) to $54.7 million (Crunchbase/Tracxn) across its rounds. Backers include Accel, Endiya Partners, Chiratae Ventures, pi Ventures, Trusted Insight, SIDBI Venture Capital and Binny Bansal.

Sources

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

• Inc42 — SigTuple company, funding and profit/loss profiles (2026)
• Tracxn — SigTuple company profile and legal-entity financials (2025-2026)
• Tofler — Sigtuple Technologies Private Limited, company and financials (CIN U74110KA2015PTC081804)
• Entrackr — “SigTuple’s two co-founders move on; Tathagato Dastidar new CEO” (June 2020); “$16 Mn Series C; Binny Bansal joins board” (April 2019); “$4 Mn extended Series C led by SIDBI Venture” (August 2024)
• YourStory — “SigTuple CEO says will persevere after co-founders exit” (June 2020); Series B coverage (June 2018)
• Biospectrum India / GenomeWeb / US FDA 510(k) database (K221309) — AI100 with Shonit FDA clearance (September 2023)
• Crunchbase — SigTuple funding and investor profile
• SigTuple company website and AI100 technical specifications — product details (Shonit, Shrava, Manthana, Mandara; CE certification)

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