Startup Deep Dive : Mad Street Den — it raised over $50 million, then sold itself for a fraction of that

Mad Street Den raised more than $57.4 million (₹551 crore) from investors including Sequoia Capital, Peak XV Partners and Alpha Wave Global over roughly a decade, building an enterprise AI platform, Vue.ai, that promised to make retailers “AI-native”. In March 2025, it sold almost all of that to a Chennai fintech, M2P Fintech, for as little as $10 million (₹96 crore) in cash and stock, in a deal reported at the time as a distress sale.

That gap — between a decade of blue-chip venture backing and an exit valued at a fraction of the money put in — is the real story of Mad Street Den. It is also a reminder that raising round after round is not the same as building a business that can stand on its own, a distinction this piece traces from a Chennai apartment in 2014 to a fire sale in 2025.

Quick facts

What they do

Mad Street Den built enterprise artificial intelligence software, sold to other businesses rather than to shoppers directly. Its best-known product, Vue.ai, used computer vision to automate the unglamorous parts of running an online store: tagging thousands of product photos, generating personalised recommendations, building visual search, and even producing AI-generated fashion models to replace some traditional photoshoots. Clients that have used its retail tools include Macy’s, Levi’s, Diesel, Farfetch, Rent the Runway, ThredUp, Myntra, MercadoLibre, Tata CliQ and Craftsvilla (Vue.ai press page; Open magazine, 2016). From 2022, a second platform, Blox, carried the same underlying AI into finance, insurance, healthcare and logistics, and the company also did document-processing work for FedEx (Vue.ai press page).

The origin

Ashwini Asokan and Anand Chandrasekaran had spent more than a decade in Silicon Valley — she leading a user-experience team at Intel, he working as a neuroscientist-turned-engineer on Stanford’s Neurogrid project — before they left in 2014 to start a company in Chennai, funding it with roughly $100,000 of their own savings (Open magazine, 2016). They did not start with a product. As Chandrasekaran put it at the time, AI was “a technology looking for a problem to solve”, and the pair spent their first months prototyping across gaming, social networking and image search before noticing something more specific: e-commerce in Asia was exploding, and retailers had no good way to differentiate themselves from one another at the product-catalogue level. Computer vision — the same technology being built for autonomous cars and defence — could be pointed at a much smaller problem: reading a product photo and understanding it well enough to tag, recommend and personalise around it. That reframing, from “AI for anything” to “AI for retail’s catalogue problem”, became the company’s founding insight.

The struggle years

The first struggle was not dramatic, but it was expensive. Before landing on retail, the founders burned through months of their own seed capital testing ideas — gaming, social apps, image search — that led nowhere, a self-funded search phase in 2014 and early 2015 with no institutional backer yet in place (Open magazine, 2016). It was only after that unfocused period that the company raised its first outside cheque, $1.5 million in a seed round from Exfinity Ventures and GrowX Ventures in early 2015 (Open magazine, 2016; Vue.ai press page).

The second struggle came a decade later, and it was terminal for the company as an independent entity. Mad Street Den had by January 2023 raised over $50 million across seven rounds, tripled its headcount to around 275 employees since 2019, and claimed 100% year-on-year growth (Entrackr, January 2023). No further institutional round followed that Series C. By March 2025, with the company reported to be “amid broader market challenges”, M2P Fintech agreed to acquire Mad Street Den’s intellectual property, contracts and employees for $10–15 million in cash and stock — a sum one source close to the deal described to the Economic Times as “a pale shadow of its past in terms of valuation for the assets being acquired” (M&A Critique, March 2025; The Head and Tale, March 2025). By mid-2026, Tracxn recorded the surviving entity’s headcount at 120 — well under half the 275 the company reported at its funding peak.

The turning point

The moment that took Mad Street Den from an unfocused seed-stage experiment to a fundable company was its pivot to retail and the traction that followed. Early customers such as Craftsvilla, Voonik and Tata CliQ adopted its tagging and recommendation tools, and the company said it could lift customer engagement by up to 300% for online retailers using them (Open magazine, 2016 — a company-stated figure, not independently audited). That traction was enough to bring Sequoia Capital in as an investor in the company’s Series A round, its first backing from a marquee global fund. The difference either side of that pivot was stark: before it, a company with $1.5 million in seed capital and no clear vertical; after it, a Sequoia-backed retail AI vendor that, by 2018, said its annual revenue had grown 400% year-on-year, and that in April 2019 raised a $17 million Series B led by Falcon Edge Capital with Global Brain and Sequoia Capital India participating (TechCrunch, April 2019).

The money behind it

• Seed, early 2015 — $1.5 million (₹14.4 crore): Exfinity Ventures and GrowX Ventures backed the company right after its retail pivot (Open magazine, 2016).
• Series A, 2016/17 — amount undisclosed: Sequoia Capital India joined as a new investor alongside existing backers Exfinity Ventures and GrowX Ventures, validating the retail-AI thesis (Vue.ai press page; TechCrunch, April 2019).
• Series B, 24 April 2019 — $17 million (₹163 crore): led by Falcon Edge Capital, with Global Brain (Japan) and Sequoia Capital India also investing; the round funded expansion into Seattle, Japan and Spain (TechCrunch, April 2019).
• Inorganic growth, November 2022: acquired Inturn, a New York-based inventory-monetisation platform, to extend into excess-inventory management (Vue.ai press page; M&A Critique, March 2025).
• Series C, announced January 2023 (round dated 12 October 2022 by Tracxn) — $30 million (₹288 crore): led by Avatar Growth Capital, with existing investors Peak XV Partners (formerly Sequoia Capital) and Alpha Wave Global (formerly Falcon Edge Capital) participating; earmarked for expansion of the finance, logistics, insurance and retail lines across the US, UK, India, the Middle East and Latin America (Entrackr, January 2023).
• Total raised: over $50 million by the company’s own account (Entrackr, January 2023); Tracxn’s independent count puts it at $57.4 million (₹551 crore) across seven rounds and 46 investors.
• Exit, 4 March 2025 — $10–15 million (₹96–144 crore): M2P Fintech acquired Mad Street Den’s IP, contracts and employees in a cash-and-stock deal, with the proceeds reported to be earmarked largely for staff costs and other liabilities rather than a payout to shareholders (The Head and Tale, March 2025; M&A Critique, March 2025).

How it makes money

Despite a consumer-friendly name and an “AI stylist” framing in its early marketing, Mad Street Den never sold to shoppers. It sold enterprise software and services to the businesses that sell to shoppers — an easy point of confusion the company itself had to keep correcting. Its revenue came from licensing and subscription arrangements with retailers, and later with financial and healthcare businesses, for access to its AI tooling; none of the specific pricing, per-SKU fees or take rates were ever made public, so they cannot be stated here.

Its product lines, as described on its own site, were organised into distinct commercial “hubs” that map to different stages of a retailer’s or enterprise’s workflow:

• Data Hub — automated product-catalogue tagging, data cleanup and content moderation, sold to retailers managing large, fast-changing catalogues (Vue.ai, company-stated).
• Customer Hub — personalisation, audience building and customer-journey orchestration, used by clients such as Tata CliQ and MercadoLibre (Vue.ai press page).
• Automation Hub — intelligent document processing and workflow automation, the line that served FedEx and, after 2025, was folded into M2P Fintech’s banking-technology stack (Vue.ai press page; YourStory, March 2025 coverage of the M2P deal).
• Optimization Hub — sales-efficiency and lead-generation tools (Vue.ai, company-stated).
• Post-acquisition value-added services — Recon360, a KYC Suite, fraud and anti-money-laundering (FRM & AML) tools and a rewards product, added once Vue.ai’s computer vision was absorbed into M2P’s fintech infrastructure (Vue.ai, 2026).

The part people consistently got wrong: this was never a shopping app, and its costs looked like an enterprise software business’s costs — engineering and machine-learning headcount, cloud compute for image processing at scale, and a direct enterprise sales team working multi-month contract cycles — not the marketing-heavy cost structure of a consumer brand. Where its margin actually sat, and what its gross margins or customer-acquisition costs were, was never disclosed and cannot be verified here.

The numbers

Mad Street Den never published consolidated, group-level revenue or profit-and-loss figures for any year, and no audited financials for the US-headquartered parent are publicly available. The only regulatory filing that is public is for its small Indian subsidiary, Mad Street Den Systems Private Limited — and that filing is a poor proxy for the group’s real business, since most of its enterprise contracts with clients such as Macy’s, Meta and FedEx would have been billed through its US entity, not the Indian one. Two data providers give different reads of even that limited filing, so both are named rather than one being invented as fact.

Cut from this section: any figure for consolidated group revenue, EBITDA, burn rate or cash runway. None of these were disclosed by the company, and none could be verified through a filing, and so none is stated as fact here — the honest answer is that Mad Street Den’s real financial health has to be read through its fundraising and exit history instead, not a income statement.

Where the money comes from

Mad Street Den was pitched, and covered in the press, mainly as a retail AI company — but by the time of its Series C, its revenue base was already split across sectors and geographies, and the eventual buyer came from a sector it had treated as a secondary bet.

• Retail and e-commerce (the founding vertical): Macy’s, Levi’s, Diesel, Farfetch, Depop, Off-White, Rent the Runway, ThredUp, Myntra, MercadoLibre, Tata CliQ, Craftsvilla and Voonik (Vue.ai press page; Open magazine, 2016).
• Financial services (BFSI): document-processing work for FedEx pre-dated the eventual sale; after March 2025, Vue.ai’s tooling was integrated directly into M2P Fintech’s banking-technology stack (Vue.ai press page; YourStory, March 2025).
• Insurance, healthcare and logistics: served through the Blox platform launched in May 2022, part of the stated plan to extend beyond retail (Vue.ai press page; Entrackr, January 2023).
• Geography: headquartered in Redwood City, California, with an engineering base in Chennai, India; the Series C round was earmarked in part for expansion across the US, UK, India, the Middle East and Latin America (Entrackr, January 2023).

The surprise is the direction the story ultimately ran in: a company that spent its first years explaining that it was not “just” a retail vendor, by trying to sell the same computer-vision stack into finance, insurance, healthcare and logistics, ended up being bought whole by a fintech infrastructure company — its “side bet” sector became its only way to survive as a business, just not as an independent one.

The risks

• A funding-dependent growth model with no further round after Series C: the company raised new capital roughly every two to three years from 2015 through October 2022, but no subsequent institutional round is on record between then and the March 2025 sale — a funding gap of well over two years that coincided with its need to sell rather than raise again (Entrackr, January 2023; M&A Critique, March 2025).
• Working-capital and liquidity pressure baked into the deal terms: reporting on the M2P transaction said the $10–15 million consideration was largely earmarked to cover staff costs and other liabilities, not to return capital to shareholders — a structure consistent with a company that needed the acquirer’s balance sheet to meet its own obligations (The Head and Tale, March 2025).
• Shrinking headcount as a leading indicator of shrinking scale: the company said it had grown to around 275 employees by January 2023; by mid-2026, Tracxn recorded the surviving entity at 120 employees, a fall of well over half, consistent with reduced ability to service the large enterprise contracts the business depended on (Entrackr, January 2023; Tracxn, accessed 2026).

The takeaway

Mad Street Den’s decade is a useful correction to the idea that a strong investor list is itself a sign of a durable business. Sequoia, Peak XV, Alpha Wave and a string of other funds backed the company across four rounds and over a decade, and each round bought time and headlines, not proof that the underlying enterprise contracts were generating enough cash to stand alone. The company also spent real effort broadening its addressable market — from retail into finance, insurance, healthcare and logistics — without any one of those verticals visibly reaching the scale needed to carry the business by itself. The lesson is not that diversification or venture backing are mistakes; it is that neither substitutes for a shorter, harder question a growing enterprise software company has to keep answering for itself: could this business cover its own costs if the next funding round did not arrive on schedule. For Mad Street Den, in 2025, the answer turned out to be no.

Frequently asked questions

What did Mad Street Den do?

It built enterprise AI software — chiefly a product called Vue.ai — that used computer vision to automate product tagging, personalisation, visual search and AI-generated fashion imagery for online retailers, later extending similar tools into finance, insurance, healthcare and logistics through a platform called Blox (Vue.ai press page).

Who founded Mad Street Den, and when?

Ashwini Asokan and Anand Chandrasekaran founded it after leaving jobs in Silicon Valley — she at Intel, he in neuroscience-linked engineering at Stanford — moving to Chennai in 2014 with about $100,000 of personal savings (Open magazine, 2016). Tracxn’s company record dates formal incorporation to 2015.

How much funding did Mad Street Den raise in total?

The company said it had raised over $50 million as of January 2023 (Entrackr); Tracxn’s independent tally puts total funding at $57.4 million (₹551 crore) across seven rounds and 46 investors, including Sequoia Capital/Peak XV Partners, Alpha Wave Global, Avatar Growth Capital, Exfinity Ventures and GrowX Ventures.

Is Mad Street Den still an independent company?

No. On 4 March 2025, M2P Fintech acquired Mad Street Den’s intellectual property, contracts and employees for $10–15 million (₹96–144 crore) in cash and stock, in a deal reported as a distress sale relative to the company’s earlier fundraising history (Entrackr, March 2025; M&A Critique, March 2025).

Did Mad Street Den sell directly to shoppers?

No. It was a business-to-business company. It licensed AI tools to retailers, and later to financial and healthcare businesses, rather than operating a consumer-facing shopping product itself, despite marketing language that sometimes suggested an “AI stylist” for individual consumers (Vue.ai about-us page).

Sources

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

• Entrackr, January 2023 — “AI firm Mad Street Den raises $30 Mn in Series C”
• Entrackr, March 2025 — “M2P Fintech to acquire Mad Street Den in distress sale”
• TechCrunch, April 2019 — “Vue.ai raises $17M to equip online retailers with AI smarts”
• Open magazine, 2016 (Wealth Issue) — “Artificial Intelligence: Ashwini Asokan, Anand Chandrasekaran… Founders of Mad Street Den”
• Vue.ai official press page (vue.ai/press), accessed September 2026
• Vue.ai official about-us page (vue.ai/about-us), accessed September 2026
• M&A Critique (via Economic Times), 5 March 2025 — “M2P Fintech in talks to acquire AI startup Vue AI in distress deal”
• The Head and Tale, March 2025 — “M2P Fintech acquiring Mad Street Den in a distress sale”
• Inc42 Datalabs — Mad Street Den company financials page, accessed September 2026
• Tracxn — Mad Street Den company profile and funding/investors page, accessed September 2026

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