Startup Deep Dive : Qure.ai — the AI that reads 15 million scans a year, still losing money

Qure.ai’s software now reads scans for more than 15 million patients a year across upwards of 90 countries, a reach few national health systems can match, according to chief executive Prashant Warier speaking to Reuters in May 2025. Yet the Mumbai-founded company earns less than 5% of its revenue from India, the market its own founding story was built around, while the United States alone accounts for a quarter of its topline.

The gap between global scale and local revenue is not the only contradiction. In the fiscal year to March 2025, Qure.ai’s loss widened 87.5% to Rs 90 crore (roughly $9.4 million at $1 ≈ ₹96.0 as of 18 September 2026, per Trading Economics), even as revenue grew nearly a quarter, according to filings reported by Entrackr in October 2025. Being called the world’s most deployed healthcare AI company, it turns out, is not the same as being a profitable one.

Quick facts

What they do

Qure.ai builds artificial intelligence software that reads medical scans and flags what a radiologist or clinician should look at first. Its flagship product, qXR, analyses chest X-rays for signs of tuberculosis, lung cancer, COVID-19-related pneumonia and more than a dozen other findings. A companion product, qER, does the same for CT scans in emergency settings, including a version, qER-CTA, cleared by the US Food and Drug Administration (FDA) in October 2025 to flag large vessel occlusions in the brain that can cause stroke. The buyers are hospitals, diagnostic chains, teleradiology providers and, increasingly, governments and global health bodies running population-level TB and lung cancer screening programmes across Asia, Africa and Latin America. The company describes itself as the world’s most deployed healthcare AI provider, with software installed at more than 3,000 sites across over 90 countries as of September 2024, per its Series D announcement on GlobeNewswire.

The origin

The founding insight was not new algorithms; it was arithmetic. There are roughly one radiologist for every tens of thousands of people in large parts of India, Africa and Southeast Asia, against ratios many times denser in the United States and Europe. A chest X-ray or CT scan taken in a resource-poor clinic can sit unread for hours or days, or never get a qualified second look at all. Prashant Warier, an operations-research PhD from Georgia Tech, had already built and sold one AI company before this: Imagna Analytics, a customer-targeting startup he founded in 2012, was acquired by Indian analytics firm Fractal Analytics in October 2015, according to Crunchbase and FinSMEs’ coverage of the deal at the time. Warier joined Fractal as chief data scientist, and it was inside Fractal, in 2016, that he and Dr Pooja Rao — a physician with an MBBS and a PhD in neuroscience from the Max Planck research school in Germany — co-founded Qure.ai. The pitch was to point deep learning, then newly capable of matching specialists on narrow image-reading tasks, at the exact bottleneck of specialist scarcity, rather than at the well-staffed hospitals of the rich world where most AI-radiology money was already chasing customers.

The struggle years

Two years of Qure.ai’s own audited numbers show a company that has struggled with the economics of its mission long after the product worked. In FY23 (year to March 2023), the company posted a loss of Rs 78 crore against revenue of just Rs 77 crore — spending Rs 2.21 for every rupee it earned, an EBITDA margin of minus 78%, according to financials reported by Entrackr in October 2024. That is a company burning cash almost as fast as a fresh Series C round could refill it: Qure.ai had raised $40 million in that Series C only a year earlier, in March 2022, led by Novo Holdings and HealthQuad with participation from TeamFund, per funding data compiled by Clay.com from Crunchbase and PitchBook records.

FY24 (year to March 2024) looked like the correction: revenue jumped 83% to Rs 141 crore and the loss narrowed 38.5% to Rs 48 crore, cutting the cost-per-rupee-of-revenue to Rs 1.43, per Entrackr’s October 2024 report. Then FY25 undid much of that progress. Revenue kept growing, up 24.5% to Rs 175.5 crore, but the loss widened 87.5% back up to Rs 90 crore as cloud-computing charges nearly doubled to roughly Rs 18 crore, legal and professional fees ran to Rs 37 crore, and employee costs rose to Rs 133 crore — 48% of the company’s total Rs 279 crore cost base — according to Entrackr’s October 2025 filing analysis. Unit economics went the wrong way again, to Rs 1.59 spent per rupee earned. None of this is a hidden crisis; it is sitting in the company’s own regulatory filings, unsoftened by the growth headlines around it.

The turning point

The event that took Qure.ai from a promising Indian AI startup to a globally referenced one was the COVID-19 pandemic. In March 2020, as hospitals worldwide restricted elective imaging and radiology departments were overwhelmed, Qure.ai retrained qXR within weeks to score chest X-rays for likely COVID-19 pneumonia rather than only its original list of findings. The retrained model was validated on a set of 11,479 chest X-rays including 515 PCR-confirmed COVID-19 cases, reaching an area-under-curve of 0.9 with sensitivity of 0.912, according to a study posted on medRxiv in July 2020. Within weeks, the COVID-19 version of the tool was running at roughly 50 sites across the UK, India, Italy and Mexico, and in Pakistan it was deployed on mobile medical vans doing community-level triage, as described in NVIDIA’s own account of the deployment that year. That burst of real-world, multi-country validation under emergency conditions is the clearest evidence trail behind the leap from a company known mainly in Indian hospital circles to one that, by September 2024, could point to more than 3,000 installed sites in over 90 countries and 18 FDA-cleared indications when it raised its $65 million Series D.

The money behind it

Qure.ai’s disclosed fundraising has moved in four visible steps. Peak XV Partners (then Sequoia Capital India) led a $16 million round in February 2020, just before the pandemic hit — capital that funded the clinical validation work behind the COVID-19 pivot. Novo Holdings and HealthQuad then led a $40 million Series C in March 2022, alongside TeamFund, backing the company’s push from pandemic-era proof of concept into paid, recurring hospital and government contracts. The largest round to date, a $65 million Series D, closed in September 2024, led jointly by Lightspeed Venture Partners and 360 ONE Asset, with Merck’s Global Health Innovation Fund and Kae Capital joining returning backers Novo Holdings, HealthQuad and TeamFund, according to the official announcement on GlobeNewswire. Each backer brought something specific: Peak XV’s early India-network capital, Novo Holdings’ healthcare-sector patience and global health connections, and Lightspeed’s later-stage capital aimed explicitly at US market entry and acquisitions, per Lightspeed partner Dev Khare’s comment in the funding announcement. Including a $8.2 million grant from the Bill & Melinda Gates Foundation recorded in January 2026, Qure.ai’s disclosed funding across rounds and grants totals roughly $133 million, per Clay.com’s aggregation of public funding data; Reuters had put the cumulative figure at $125 million as of May 2025, before that grant. The company’s valuation has never been officially disclosed; Tracxn estimated it at $264 million as of November 2024, a figure Reuters repeated in its May 2025 report on the company’s IPO plans — short of the $1 billion mark that would make it a unicorn, despite periodic industry chatter framing it as close to one.

How it makes money

Qure.ai sells software, not devices or radiologist time. Buyers license its algorithms, typically bundled with the reading workflow, either as standalone software layered onto a hospital’s existing X-ray or CT machines, or through partnerships with device makers and teleradiology networks that resell the AI read as part of their own service. Sales of tools and software made up 86% of operating revenue in FY25, about Rs 151 crore of the Rs 175.5 crore total, with the remainder coming from other healthcare products, according to Entrackr’s October 2025 breakdown. On the cost side, the company is engineering- and regulatory-heavy rather than sales-heavy: employee expenses, at Rs 133 crore in FY25, are nearly half of all costs, reflecting the data-science and clinical-affairs headcount needed to keep clearing new FDA indications and country-specific approvals; cloud computing, at roughly Rs 18 crore and rising fast, reflects the compute cost of running inference on scans at scale. The part outsiders tend to get wrong is assuming Qure.ai’s business is charitable or grant-funded because its most visible use cases are TB screening in poor countries. In practice, according to CEO Prashant Warier’s own account to Reuters in May 2025, the United States is Qure.ai’s single largest market at about 25% of revenue, while India — the market that inspired the company and hosts many of its highest-profile public health deployments — contributes under 5%. The margin, such as it exists, sits in software sold into well-resourced health systems that can pay recurring licence fees; the population-screening work in lower-income countries builds evidence, regulatory credibility and, in cases like Gates Foundation funding, grant revenue, more than it builds gross margin today.

The numbers

Three consecutive years of audited financials, as reported by Entrackr, show revenue climbing steadily while profitability has moved in the opposite direction more often than not.

Revenue nearly doubled from FY23 to FY24 and grew again in FY25, a three-year compound path that would please most enterprise software investors. The loss line is the harder story: it fell sharply in FY24 only to nearly double again in FY25, meaning Qure.ai has now posted a larger loss in FY25 than in either of the prior two years bar FY23. Cash and bank balances stood at Rs 35 crore at the end of FY25, per Entrackr — a runway figure investors will be watching closely against a Rs 90 crore annual burn, even with fresh Series D capital sitting on the balance sheet from the prior year.

Where the money comes from

Geographically, Qure.ai’s revenue has essentially inverted from where the company is headquartered. International revenue rose 39.6% to Rs 174 crore in FY25, more than 99% of the operating total, while domestic Indian revenue fell 80% year on year to just Rs 1.3 crore, according to Entrackr’s October 2025 analysis. Reuters’ May 2025 reporting puts the United States alone at about 25% of revenue, the single largest national market, with India in the low single digits. The surprise for a company whose entire founding narrative is about serving radiologist-starved, lower-income health systems is that the paying business now runs almost entirely on higher-income, better-reimbursed geographies, while India and comparable markets carry a disproportionate share of the public health and screening-programme visibility — the TB active-case-finding work, the UNICEF and Save the Children partnerships on paediatric screening, the CE-marked toddler TB tool cleared in October 2025 — without yet carrying a matching share of the revenue.

The risks

Three risks sit close to the surface of Qure.ai’s own disclosures and its market position. First, cash burn against a thin balance sheet: a Rs 90 crore FY25 loss against Rs 35 crore of cash and bank balances, per Entrackr, means the company is dependent on either reaching the breakeven CEO Prashant Warier told Reuters he expects “next financial year” (FY26) or raising further capital before its runway tightens. Second, regulatory fragmentation: each new market requires its own clearance — FDA in the United States, CE marking under the EU’s Medical Device Regulation in Europe, and separate national approvals across the African and Southeast Asian countries where its TB programmes run — so growth is gated country by country rather than scaling in one motion, and the pace of clearances (19 FDA indications as of October 2025, per the company’s own press release) has to keep outrunning the cost of getting each one. Third, competitive convergence: independent market analysis from IntuitionLabs in 2025 places Qure.ai among a “Tier II” group of disease-specific AI vendors, alongside Aidoc, Arterys, Gleamer and Viz.ai, that differentiate today by acting as a primary screening read in radiologist-scarce settings rather than the second-opinion role most competitors play in well-staffed US hospitals; if better-funded rivals push into the same low-resource, high-volume screening markets, or if health systems standardise on a single triage vendor per hospital network, that differentiation narrows.

The takeaway

Qure.ai’s most transferable lesson is not about artificial intelligence at all. It built its core product for the hardest possible customer first — a clinic with one X-ray machine, no resident radiologist, and a public health mandate to find tuberculosis cases it might otherwise miss — and only later carried the same underlying algorithms into the best-resourced, most reimbursement-friendly health system in the world. Designing for scarcity produced a technically rigorous, widely validated product; it took roughly a decade, a pandemic, and four funding rounds before that product’s revenue caught up with its reach. The gap between deployment scale and financial scale, visible in Qure.ai’s own filings, is the honest cost of building for the underserved market first and monetising the well-served one later.

Frequently asked questions

What exactly does Qure.ai sell, and to whom?

Qure.ai sells AI software that reads chest X-rays, CT scans and ultrasound images to flag findings such as tuberculosis, lung cancer, stroke-causing blockages and cardiac issues. Its customers are hospitals, diagnostic chains, teleradiology providers, and governments or global health organisations running population-level screening programmes, with software installed at more than 3,000 sites across over 90 countries as of September 2024, per its Series D announcement.

Is Qure.ai profitable?

No. Its loss widened 87.5% to Rs 90 crore in FY25 (year to March 2025) even as revenue grew 24.5% to Rs 175.5 crore, according to financials reported by Entrackr in October 2025. CEO Prashant Warier told Reuters in May 2025 that the company expects to break even in the following financial year.

How much funding has Qure.ai raised, and at what valuation?

Disclosed rounds and grants total roughly $125–133 million depending on the cut-off date, per Reuters (May 2025) and Clay.com’s aggregated funding data, across a $16 million round in 2020, a $40 million Series C in 2022, a $65 million Series D in 2024 and a Gates Foundation grant in 2026. Its valuation has not been officially disclosed; Tracxn estimated it at $264 million as of November 2024, a figure Reuters repeated in May 2025.

Is Qure.ai a unicorn?

Not based on any publicly reported figure. Its most recent reported valuation estimate, $264 million as of November 2024, is well short of the $1 billion unicorn threshold, despite periodic industry chatter about its growth trajectory.

How is Qure.ai different from competitors like Aidoc or Lunit?

Qure.ai and peers such as Aidoc, Lunit and Viz.ai all sell AI radiology triage software, but market analysis from IntuitionLabs in 2025 places Qure.ai’s differentiation in radiologist-scarce settings, where its software often functions as the primary screening read, versus the second-reader role AI tools typically play in well-staffed hospitals in the United States and Europe.

Sources

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

• Entrackr (FinTrackr) — “Qure.ai revenue soars 83% to Rs 141 Cr in FY24, slashes losses” — October 2024
• Entrackr (FinTrackr) — “Qure.ai’s losses widen 87% to Rs 90 Cr in FY25” — October 2025
• GlobeNewswire — “Global healthcare AI innovator Qure.ai completes $65 million Series D funding round led by Lightspeed and 360One Asset” — September 2024
• Reuters, carried on Yahoo Finance — “Indian healthcare AI startup Qure.AI aiming for IPO in two years, CEO says” — May 2025
• Clay.com — Qure.ai funding dossier, aggregating Crunchbase and PitchBook data — accessed September 2026
• Crunchbase and FinSMEs — coverage of Fractal Analytics’ acquisition of Imagna Analytics — October/November 2015
• Qure.ai newsroom — “Qure.ai Earns 19th FDA Clearance for Neurocritical AI in U.S.” — October 2025
• HLTH.com — “CE Mark For India-Made AI For Screening TB In Toddlers” — October 2025
• NVIDIA Blog — “Qure.ai Helps Clinicians Answer Questions from COVID-19 Lung Scans” — 2020
• medRxiv — “Computer-aided covid-19 patient screening using chest images (X-Ray and CT scans)” — July 2020
• IntuitionLabs — “Top Imaging & Pathology AI Companies: 2025 Market Analysis” — 2025

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