"most people think success compounds but rejections were compounding because you were making them compound you were learning something from it" - [Vikram Madhyanathan] [00:13:14]
"the core thing is nobody can care about your company more than you do right" - [Harshil Mathur] [01:04:51]
Disclaimer: Orignal content owned by or sourced from third parties. It does not represent the views of 'Nuggets' platform or it's team. AI is used extensively across this platform including for summaries. Accuracy is not guaranteed, there can be mistakes. Any info or content on this platform is not a financial, legal, or investment advice. Do your own research. Refer for complete disclosures:- Terms of Use · Full Disclaimer
"you have to be in the ecosystem that you're going to serve rather than the ecosystem that going help you sir" - [Harshil Mathur] [00:16:51]
"I think the good part about the YC form is that it asks you the question that most founders should ask themselves they are starting a company but they typically don't" - [Harshil Mathur] [00:19:50]
"you are like a utility company like nobody cares when your electricity is coming... but the five minutes that you're down... you're going to get all the abuses in the world" - [Harshil Mathur] [00:18:17]
"engineers are excited by the complexity of the build than the outcome of the build" - [Harshil Mathur] [00:58:46]
"AI makes the execution speed really fast it makes intelligence available on the fly but the pro and con of it is it's going to make your weaknesses appear far more faster" - [Harshil Mathur] [00:56:08]
Speakers & Credentials
Harshil Mathur: Co-founder and CEO of Razorpay, a $7.5 Billion Indian fintech giant. An IIT Roorkee graduate who left a highly lucrative oil-field engineering job to solve digital payment friction in India.
Vikram Madhyanathan: Managing Director at Z47 (formerly Matrix Partners India). An early believer and seed investor in Razorpay who backed the founders remotely based on their deep structural approach to banking architecture.
Anish: Host of the Z47 Moments podcast, facilitating the deep-dive conversation into Razorpay's operating metrics, cultural crucibles, and strategic evolution.
1. Executive Summary
The Power of 'Lived Empathy': Razorpay’s genesis was not born from a detached market analysis, but from Harshil Mathur’s visceral frustration with incumbent banking APIs while building side projects at IIT Roorkee. This empathy allowed the founding team to architect the product entirely from the merchant's perspective, rather than the supplier's.
Compounding Rejection as Alpha: Breaking into a monopolized, highly-regulated sector required brute-force iteration; the founders faced rejection from nearly 100 bankers, strategically using each failure to iteratively map the unknown architectural constraints of Indian fintech.
The Inescapable B2B Grind: Despite massive funding post-Y Combinator, the company faced a grueling 2-3 year plateau where capital could not artificially accelerate enterprise adoption; they had to painstakingly build trust, ultimately winning over massive clients like Goibibo through relentless support and iteration.
Culture as a Crisis Mechanism: During the severe Yes Bank moratorium crisis, Razorpay’s foundational culture negated complex decision-making. By protecting the survival of thousands of micro-merchants over the capital of ten enterprise clients, they proved that true corporate values are forged, and paid for, during existential threats.
The AI-Driven 'Founder Mode' Renaissance: Observing the paradigm shift brought by LLMs, Harshil broke a six-year coding hiatus to personally re-architect Razorpay’s internal workflows. He advocates that the AI epoch requires founders to plunge back into the operational trenches, viewing AI not as a novelty, but as a ruthless execution engine that amplifies both organizational speed and underlying structural weaknesses.
2. Chronological Table of Contents
[00:00:00] - Introduction & The Genesis of SDS Labs
[00:05:44] - Abandoning Schlumberger & The $100K Trap
[00:11:32] - Compounding Rejections: Pitching to 100+ Banks
[00:16:32] - The Outsider Advantage & Merchant-First Architecture
[00:19:05] - Y Combinator, Matrix Partners, & First Funding
[00:25:04] - The Inescapable B2B Grind & The Goibibo Crucible
[00:37:40] - Historical Crises: Demonetization, COVID, and Yes Bank
[00:42:42] - The Iron Rule of "Being First" in Fintech (UPI Launch)
[00:48:23] - The AI Epoch: Tinkering, LLMs, and Execution Engines
[00:54:01] - Flipping the Org Structure & Harness as a Service
[01:00:50] - Agentic Studios & Bypassing the SaaS Curve in India
[01:03:34] - The Bottom Line: Paul Graham's 'Founder Mode'
3. Detailed Thematic Summary
The Outsider's Alpha & The Lived Empathy Engine
The inception of Razorpay was deeply tied to Harshil and Shashank's time at IIT Roorkee, where they founded SDS Labs to escape the theoretical confines of computer science and mechanically build real products [00:03:05].
Harshil consciously rejected the standard path of Indian engineering prestige, abandoning a massive, life-changing ~$100,000+ salary at Schlumberger in Abu Dhabi after only a brief 4-to-6 month stint because the "oil field" environment actively stifled his desire to build digital products [00:06:13].
The core insight for Razorpay didn't come from a McKinsey market report; it came from a failed side-project. While trying to build a crowdfunding platform, they realized that despite the rapid democratization of social media and e-commerce, digital payments remained drastically less accessible than physical cash in India [00:08:06].
Because they lacked legacy financial networks, they were forced into a merchant-first architectural paradigm. Instead of building a generic wrapper around opaque, antiquated bank APIs (the standard incumbent strategy), they built the absolute ideal interface for a startup, and then brutally retro-fitted the banking backend to match their UI—an approach that fundamentally separated them from competitors [00:15:54].
They quantified this empathy early. The industry standard payment success rate was hovering at an abysmal ~65% [00:17:46]. Razorpay anchored its entire GTM motion on driving this to 85-90%, a metric they have since pushed to an elite 92-93% scale [00:17:53].
The Architecture of the B2B Grind & Compounding Rejection
Early startup literature deeply romanticizes rapid consumer scaling, completely ignoring the punishing realities of B2B infrastructure. Harshil physically walked into a Jaipur bank branch with zero connections, asking for API access, only to be literally laughed out of the building [00:11:59].
They endured rejections from roughly 100 bankers, but treated these rejections as a distributed training dataset. With each denial, a banker would drop a nugget of regulatory or risk-management reality, which Harshil immediately integrated into their next pitch. As Vikram noted, "rejections were compounding because you were making them compound" [00:13:14].
Entering Y Combinator serendipitously shifted their trajectory. Applying with zero expectations and a meticulously completed application, they survived the notoriously brutal 10-minute YC interview format, moving to the US for 3 months to absorb global engineering standards [00:20:23].
Post-YC, despite having raised substantial capital (including a highly unorthodox, remote Skype-pitch investment from Matrix Partners [00:24:04]), Razorpay slammed into a 2-to-3 year GTM plateau. B2B payments operate on pure utility—you cannot aggressively discount or "cashback" your way into an enterprise CFO's tech stack [00:26:06].
Following the mental model of "hunting rabbits vs elephants," they serviced tiny startups for years before finally landing Goibibo in 2017 [00:26:51]. Even then, they suffered a catastrophic false start, being paused by the enterprise client due to infrastructure lag, requiring 6-8 months of frantic, dedicated stabilization before becoming a viable enterprise standard [00:28:03].
Deep-Time Crisis Management & The Crucible of Culture
The true test of Razorpay’s architecture was not scaling, but survival during profound macro-crises. When the Yes Bank moratorium struck in early 2020, over 100+ crores of customer capital was frozen, threatening an ecosystem-wide collapse [00:38:14].
Instead of waiting for executive directives, the Razorpay culture—forged over years of hiring strict "A-Players" and refusing to dilute standards—auto-executed. Within hours, 200 to 300 employees, including facility staff, autonomously repopulated the office at midnight to map the blast radius [00:38:47].
Harshil made a historically defining capital allocation decision: utilizing internal cash reserves to bail out merchants. When forced to choose between satisfying the top 10 enterprise revenue drivers or saving 1,000 micro-businesses from literal bankruptcy within 24 hours, he chose the latter. He accepted furious legal threats from enterprises, proving that cultural DNA is not a wall poster, but a highly expensive tactical decision [00:39:58].
This agility was also displayed as an offensive maneuver in 2016. When the UPI protocol was announced, massive incumbents (SBI, HDFC) were paralyzed. Razorpay, servicing a meager 10,000 businesses, aggressively integrated UPI 6 months ahead of the curve, absorbing a massive wave of enterprise traffic fleeing legacy systems that couldn't adapt to the sovereign protocol shift [00:45:49].
The Generative AI Paradigm Shift & A Return to 'Founder Mode'
Harshil recognizes that the post-2022 AI shift is an extinction-level event for complacent codebases. In late December and January, tinkering with models like Claude and ChatGPT, he realized AI had evolved from a passive assistant into a relentless "execution engine" [00:48:23].
Breaking a six-year management hiatus, his GitHub repository lit up green. He weaponized LLMs to build agentic dietary planners and meeting notes, plunging the CEO back into the coding trenches [00:52:11].
He forced a sweeping organizational mandate: Razorpay employees are now actively rated from L1 to L5 on AI proficiency, with explicit demands that L1-L3 talent must forcibly upscale. He inverted traditional corporate hackathons, forcing the executive leadership team to build and present AI tools to junior engineers (Harshil placed 4th) [00:54:52].
Harshil warns of the "Garbage at Scale" vulnerability. In a test run, an autonomous D2C company-building AI agent rapidly drifted into hallucinated failure states because it lacked strict architectural boundaries [00:56:24]. He notes that AI makes you faster, but also exponentially magnifies existing organizational weaknesses.
Embracing Paul Graham’s "Founder Mode," Harshil argues that delegating core company pillars—tech architecture, brand, and culture—to managers during an epochal technological shift is corporate suicide. The founder must remain inextricably linked to the machine [01:04:38].
The Reference Vault
4. Data & Figures
Data Point
Value
Context
Timestamp
First Oil-Field Salary
~$100,000+
Life-changing salary at Schlumberger that Harshil walked away from.
The Outsider's Advantage (Merchant-First Architecture): [00:16:32]
In established industries, incumbents are severely handicapped by their intimate knowledge of supplier constraints. Because Razorpay had zero inside access to banks initially, they were forced to treat the bank as a black box and build their API wrappers purely based on what startups actually wanted. This outsider ignorance forced an obsessively customer-centric architecture; by the time they plugged the banks in, the superior merchant UI was already cemented.
Hunting Rabbits vs. Elephants (B2B GTM Strategy): [00:26:51]
In enterprise B2B sales, founders are perpetually seduced by the illusion of the "Elephant"—the single massive client that will validate their valuation and solve revenue churn forever. Vikram injected this model to force Razorpay to abandon their crippling enterprise greed. By hunting "Rabbits" (startups with high failure rates but massive velocity), Razorpay built an unshakeable operational base. They survived the grueling three-year plateau because a swarm of rabbits provides far more systemic resilience than betting the company’s runway on an 18-month elephant sales cycle.
The Utility Company Axiom: [00:18:17]
Unlike consumer apps (social media, content) that obsess over maximizing "Time on Platform" via dopamine-driven UI features, infrastructure SaaS must optimize for sheer invisibility. Harshil drives this ethos deep into product managers: nobody celebrates their electricity functioning, but five minutes of a blackout triggers unmitigated rage. Razorpay ruthlessly strips away animations and UI friction because every additional second a user spends checking out is a measurable failure in value delivery.
Harness as a Service (The AI Shift): [00:59:37]
The raw intelligence of LLMs (like GPT-4 and Claude) is rapidly commoditizing. The moat of the next decade is no longer building the intelligence itself, but building the "Harness." By wrapping chaotic, non-deterministic LLMs in highly deterministic, workflow-specific boundaries (like Razorpay's Cart Abandonment Agent), companies translate raw computational theory into measurable SME value. The bottleneck has shifted away from code execution and entirely toward architectural planning and error bounding.
Founder Mode (vs. Manager Mode): [01:04:38]
Drawing on Paul Graham's modern essay, this framework rejects the classical MBA doctrine that scaling requires founders to retreat into detached "Manager Mode." Harshil argues that during epochal technological shifts (like the generative AI wave), delegating core structural pillars is a death sentence. Founders must aggressively micromesh with product engineering, corporate culture, and brand defense, because executives optimize for quarterly stability, whereas founders are required to optimize for existential survival.
6. Anecdotes
The Jaipur Bank Branch Comedy: [00:11:59]
Context: Highlighting the extreme naivety and pure hustle of early-stage founders, Harshil literally walked into a local retail bank branch in his neighborhood in Jaipur and asked the teller how to build a payment gateway and get API access. He was laughed at, but that sheer audacity earned him his first connection to a corporate Delhi office, initiating the compounding sequence of 100+ rejections that ultimately trained them in banking architecture.
The Y Combinator Skype Violation: [00:24:04]
Context: Venture Capital firms operate on rigid heuristics, the most sacred being: never invest without meeting the founders in person. Matrix Partners intentionally shattered this rule for Razorpay. Realizing that Harshil and Shashank possessed an almost savant-level understanding of banking failure rates, Vikram and his team pushed a term sheet entirely over an international Skype call while the founders were at YC. It illustrates how deep domain expertise allows founders to bypass institutional friction.
The Post-TechCrunch Outage & The Abusive Customer: [00:36:18]
Context: Shortly after completing Y Combinator and riding the high of TechCrunch press, Razorpay's core partner bank completely pulled the plug. With their entire system down for a week, Harshil forced his tiny 7-8 person team to manually call all 20-30 early merchants to deliver the bad news. While they faced immense abuse on the phone, Harshil notes that one of the most aggressively abusive customers remained a loyal merchant years later, proving that extreme transparency during failure breeds unshakeable loyalty.
The Yes Bank Midnight Mutiny: [00:38:24]
Context: When the RBI froze Yes Bank, Razorpay had roughly 100+ crores locked. Harshil rushed back to the office expecting to fight a solitary executive war. Instead, he found the office organically flooded with 200-300 employees—including housekeeping staff brewing coffee—who had returned at midnight unprompted. Harshil sat back, realizing his only job was to authorize the bailout of 1,000 small merchants over 10 screaming enterprise clients. The anecdote proves that true corporate culture acts as a decentralized autonomous system during extreme duress.
The Goibibo False Start: [00:28:03]
Context: Razorpay spent years chasing their first "Elephant," Goibibo. When an incumbent payment gateway collapsed, Razorpay got their shot. They threw dedicated teams at the integration, went live, and celebrated. A month later, Goibibo ruthlessly paused them. The product scaffolding simply wasn't robust enough for enterprise load. The anecdote is a brutal reminder that startup "hustle" and white-glove support cannot mask a failing core product; it took 8 months of sheer engineering grind to permanently win back the account.
The Autonomous "Garbage" Company: [00:56:24]
Context: Fascinated by the new era of AI, Harshil deployed an autonomous framework designed to build a complete D2C company by itself. He watched as the AI hired sub-agents, assigned marketing tasks, and began coding. Within minutes, the AI drifted into absolute hallucinations, producing highly complex, utterly useless garbage. Harshil uses this to warn his engineers against the "Complexity Trap"—without strict human architectural planning, non-deterministic AI will confidently build beautiful bridges to nowhere.
7. References & Recommendations
Geopolitical & Institutional Entities
Yes Bank / RBI Moratorium: [00:37:45] - Referenced as the ultimate macro stress-test of Razorpay's internal culture and financial resilience when massive capital blocks were frozen.
UPI (Unified Payments Interface): [00:45:34] - Mentioned as a sovereign protocol shift. Razorpay adopted it 6 months ahead of megabanks, capturing massive market share by refusing the "wait and see" approach.
IIT Roorkee: [00:01:54] - The prestigious Indian Institute of Technology where Harshil and Shashank met. Noted as a place where the theoretical computer science curriculum forced them to seek hands-on coding side projects.
Companies & Products
SDS Labs: [00:03:05] - The collegiate coding group Harshil and Shashank founded at IIT Roorkee to build practical software, serving as the foundational testing ground for their entrepreneurial partnership.
Schlumberger: [00:06:01] - The premier oil-field services company. Harshil abandoned a highly lucrative career here to suffer the friction of early-stage software building.
Hotstar / Napster: [00:03:17] - Mentioned as contextual markers for the era when SDS Labs was illegally streaming cricket matches on the campus intranet, drawing parallels to early internet pirate ingenuity.
Ola: [00:09:48] - Cited as an example of a B2C business where founders can easily empathize with the end user (booking a cab), contrasting with the difficult empathy required for B2B fintech.
Y Combinator (YC): [00:19:05] - The global startup accelerator. Harshil noted that the true value wasn't the tactical advice, but the psychological exposure to global engineering standards.
Matrix Partners India (Now Z47): [00:22:27] - The VC firm that broke their own due-diligence rules to seed Razorpay via Skype, providing the necessary runway for the multi-year B2B grind.
Tiger Global: [00:25:51] - Mentioned as one of the massive growth funds that capitalized Razorpay after YC, creating immense pressure to scale.
Goibibo: [00:27:35] - Razorpay’s first critical enterprise client, serving as the crucible that forced them to mature their infrastructure from startup-tier to enterprise-tier.
TechCrunch: [00:36:23] - The iconic tech publication. Reaching their front page after YC demo day was immediately followed by Razorpay's banking partner pulling the plug, creating an ironic PR disaster.
Stripe: [00:14:41] - Mentioned as the global paradigm for payment processing, which Razorpay chose to rival in India rather than waiting for Stripe's localized deployment.
Claude / OpenAI: [00:47:49] - The foundational LLMs that triggered Harshil to break his 6-year coding hiatus and re-enter "Founder Mode," viewing these tools as execution engines rather than toys.
GitHub: [00:52:11] - The developer platform. Harshil noted his commit history turned fully "green" for the first time in six years thanks to AI-accelerated coding workflows.
Apple / Amazon: [00:41:44] - Cited as historical beacons where dominant corporate culture and operational intensity was permanently hardcoded by the founders (Steve Jobs, Jeff Bezos).
People & Intellectuals
Shashank Kumar: [00:02:23] - Razorpay Co-Founder. Harshil credits him with maintaining the unyielding standard of engineering hires, refusing to dilute the team to a "B-Team" even under pressure.
Rajinder & Rajat: [00:23:01] - Founders / investors adjacent to the Matrix ecosystem who acted as the critical networking bridge that tipped off Matrix Partners to Razorpay's stealth YC entry.
Steve Jobs & Jeff Bezos: [00:41:44] - Referenced as the gold standards of leaders who built culture manually, proving that "culture eats strategy" is only true if founders lead from the absolute front.
Paul Graham: [01:04:38] - YC Founder. His essay on "Founder Mode" is cited by Harshil as the definitive blueprint for navigating existential technological shifts, demanding founders remain intimately connected to core operations.
8. The Bottomline (by AI)
The era of passive, "wait-and-see" leadership is officially dead; the AI disruption dictates that founders must re-enter the operational trenches or face rapid obsolescence. Razorpay’s trajectory proves that while compounding rejection builds unshakeable infrastructure, maintaining an uncompromising, empathy-driven culture is the only true defense against macro-crises. Executives must immediately weaponize AI as a ruthless execution engine, shifting their focus from basic output generation to designing strict, deterministic boundaries ("Harness as a Service") that convert raw computational chaos into measurable enterprise value.
Jul 16, 2026
Dr. Robert Wachter | A Giant Leap: How AI Is Transforming Healthcare... | 14 Jul 2026 | Talks at Google
"don't get me wrong US healthcare delivers miracles every day particularly when it comes to cutting edge and intensive care... but the health care system itself is a headache wrapped in red tape inside the nightmare that France Kofka himse…
Current Success Rate
92-93%
The optimized transaction success rate achieved by Razorpay today.