How Render Found Product Market Fit: Anurag Goel

It took Render nearly four years to hit $10M ARR — with no marketing spend and no freemium tier. Anurag Goel, former employee #8 at Stripe, explains why that slow burn was actually proof of something rare: genuine product market fit built entirely on word-of-mouth.

Key Facts

• Time to $10M ARR: ~4 years from founding (launched TechCrunch Disrupt 2019) (Anurag Goel)
• Total funding raised: $250M, including $100M at $1.5B valuation (Pablo Srugo)
• Platform scale: 6 million+ developers; hundreds of thousands signing up monthly via word-of-mouth (Anurag Goel)
• Launch team size: 4 people: 1 designer + 3 engineers (including founder) (Anurag Goel)
• Support team size at 6M users: Fewer than 10 people — all technically grounded (Anurag Goel)

What Real Product Market Fit Actually Feels Like

PMF isn't just users liking your product — it's users demanding more of it and telling others. When customers ask for features because they're already invested, not because they're evaluating, that's the signal. For Render, this happened immediately after their TechCrunch Disrupt 2019 launch.

Most founders mistake polite interest for product market fit. Anurag Goel draws a sharper line: real PMF shows up as pull — customers asking for more features because they're already hooked, not just sampling. Before Render launched publicly, the team shipped at their own pace. After launch, everything changed.

The TechCrunch Disrupt Battlefield competition in 2019 was Render's public debut. Winning brought a flood of new users — and with them, a flood of feature requests. The product was unfinished, but users kept coming back. That tension — incomplete product, committed users — was the signal Goel had been waiting for.

But 'people like it' isn't enough on its own. Host Pablo Srugo pushes on the distinction between users who enjoy a product and users who drive hyper-growth. Goel's answer is unambiguous: the difference is whether they tell other people.

"When you build something that is incomplete that people are willing to try, to use, to pay you for, and they want it to get better — that to me is always a great sign of product market fit." — Anurag Goel

"The worst thing that can happen is people don't care about your product. And then they just move on and they don't need new features." — Anurag Goel

Word of Mouth Is the Only Proof That Counts

Word-of-mouth growth — not ads, not referral bonuses — is the clearest proof of product market fit. Users who organically recommend your product to peers do it because the value was obvious fast. For developer tools, this often means blog posts and Hacker News threads, not direct referrals.

Goel makes a clean distinction between virality and word of mouth. Virality requires network effects — you need others on the platform for it to benefit you. Word of mouth is simpler and harder: someone recommends your product because they genuinely want their peers to experience it.

For Render, the developer audience made this especially powerful. Developers love discovering new tools, and when they find something that works, they write about it. Blog posts, Hacker News threads, Twitter threads — these became Render's growth engine. The company signs up hundreds of thousands of developers every month, all organically.

Srugo adds an important nuance: word-of-mouth is linked to time-to-value. If you know a product will wow someone immediately, you tell ten people. If onboarding takes weeks, you hedge. Render's value proposition — connect your GitHub repo, get a live server in minutes — is engineered for that instant wow.

"You don't really have product market fit until you're growing naturally. And that usually happens when your users are spreading the word." — Anurag Goel

"We're signing up hundreds and hundreds of thousands of developers every month. And that's all word of mouth. We're not driving that through ads." — Anurag Goel

• Word-of-mouth ≠ referral programs — it's organic enthusiasm
• Developers write publicly about tools they love (blogs, HN, forums)
• Fast time-to-value lowers the social cost of recommending a product
• Render reaches hundreds of thousands of new developers monthly with zero ad spend

Founder-Market Fit: Why Anurag Chose Developer Infrastructure

Goel spent 18 months evaluating markets before starting Render — exploring healthcare, AI infrastructure, and application hosting. He chose developer infrastructure not because it was the biggest market, but because it was the intersection of his skills, interests, and a genuine problem he kept being pulled back to.

After nearly five years as employee #8 at Stripe, Goel had the financial freedom to work on anything. So he spent 18 months doing something most founders skip: deeply examining which problems he was the right person to solve. He explored healthcare, early deep learning infrastructure, and more before landing on application hosting.

One early prototype gave him a signal. He built a single-click GPU-backed Jupyter notebook for data scientists taking a deep learning course — and it grew to 10,000 users through word of mouth. But he walked away from it, because he cared more about application developers than data scientists. That personal clarity was founder-market fit in action.

Goel is direct about why this matters more than market size: if your startup succeeds, you'll be working on it for a decade or more. Startup life involves constant chaos regardless of how successful you look from the outside — so you better genuinely love the problem.

"You really have to find founder market fit. I think that's crucially important if you're really in it for the long run. Because if you're successful, you're going to be in it for the long run. So you better like what you're doing. Otherwise, you're kind of stuck." — Anurag Goel

"It doesn't matter how successful they are on the outside. Pretty much every company is chaos on the inside when they're growing." — Anurag Goel

Why Hyperscalers Left Developers Underserved

AWS, GCP, and Azure are built around massive enterprise contracts with DevOps-heavy teams. Their organizational DNA — incentives, resources, and culture — is optimized for that customer. Application developers are structurally deprioritized, leaving a large underserved segment that Render was built to serve.

One of the sharpest insights in this episode is Goel's explanation of why AWS never solved the problem Render solves. It's not that they couldn't — it's that they're structurally incentivized not to. Enterprise contracts with massive DevOps teams are where the money is. Application developers are a rounding error on their revenue.

When hyperscalers have tried to build simpler developer tools, those products don't get the right resources or leadership attention. And they always have an escape hatch: 'if it doesn't work, just use Kubernetes.' That fallback is exactly what Render refuses to offer. If a user needs Kubernetes to solve a problem on Render, Goel considers that a Render failure.

Srugo reframes this as competitive strategy: when you enter a market the hyperscaler is ignoring, you're not competing with AWS — you're competing with the underfunded internal team with the 50th-best PM. That's winnable.

"For AWS and GCP and Azure, all the money is in massive enterprise contracts with companies that have massive DevOps teams. So guess who they're building for? They're building for DevOps engineers." — Anurag Goel

"For render, that was never an option. We said, look, if you have to use Kubernetes for something, that's a failure. That's a thing that render should fix." — Anurag Goel

Getting to $10M ARR: What Worked, What Didn't

Render took four years to reach $10M ARR — no freemium, no marketing spend. The biggest early mistake was skipping a free tier, which curtailed sign-ups significantly. They also had to raise prices in 2023 to achieve unit economics, accepting slower growth to build a sustainable business.

The path from launch to $10M ARR was slower than modern VC timelines would tolerate. Render launched publicly in October 2019 at TechCrunch Disrupt and hit $10M ARR around late 2021 or early 2022 — a 25-30-person team at that point. No paid marketing. No freemium. Just a product people liked enough to pay for and recommend.

Goel admits the freemium decision was a mistake. The team assumed a free trial would attract serious users with real production apps. Instead, they learned that many developers won't try anything — even with a trial — if there's any payment involved upfront. The free tier came after their Series A and changed the trajectory.

But the free tier created a new problem: unit economics. Every new user was increasing burn. In 2022, Render shifted engineering focus from features to infrastructure efficiency — reducing cloud costs so revenue and spend stayed aligned. Then in January 2023, they raised prices. Churn was minimal, but growth slowed. Goel made the call anyway, knowing a sustainable business was the only path forward.

"I made a mistake and I think we should have had a free tier sooner. Everyone else had a free tier, and we decided not to have one... that was severely curtailing our total signups." — Anurag Goel

"Every new user was actually increasing our burn rate at some point. And we had to get out of that before we invested in marketing." — Anurag Goel

• No free tier pre-Series A was a self-identified mistake — curtailed sign-ups
• 25-30 person team at $10M ARR — extremely lean by industry standards
• 2022 focus: infrastructure efficiency over feature development
• Price increase in Jan 2023: low churn but slower growth — a deliberate tradeoff

Engineer-Led Support as a Product and Growth Lever

Render had engineers rotate through one week of full-time customer support each. The tight feedback loop let engineers fix bugs in real time — turning frustrated users into champions. This approach informed the product roadmap directly and helped Render support 6 million users with fewer than 10 support staff.

Before hiring a dedicated support team, every Render engineer spent one full week per rotation doing nothing but customer support. No coding. No other projects. The weekly cadence (rather than daily) was chosen specifically to reduce context-switching costs while maintaining deep customer immersion.

The payoff was twofold. First, engineers could often fix problems on the spot — turning a negative customer experience into a memorable positive one. Second, every engineer developed firsthand intuition about which issues recurred most, which directly shaped roadmap prioritization. No AI classification needed — just pattern recognition from direct exposure.

When Render eventually hired dedicated support staff, they maintained the same bar: only technically grounded people who could understand and reproduce customer issues. These support engineers didn't just resolve tickets — they brought synthesized, solution-ready feedback to product teams. Today, with 6 million developers on the platform, fewer than 10 support staff handle the load.

"In the beginning, it was actually really helpful for engineers to hear directly from our customers who were also engineers. In many cases, we were able to fix the problem right away. And that really quick feedback loop made folks really happy, even though they had run into a problem." — Anurag Goel

"We made sure that the people who were answering customer support requests had enough technical grounding to then give the right kind of feedback to the product team." — Anurag Goel

How AI Supercharged Render's Growth

The generative AI wave massively accelerated Render's growth. More developers are building more apps — including developers who previously lacked time or resources for side projects. DevOps teams can't keep up with the pace of AI-assisted development, pushing companies toward self-serve platforms like Render.

Render is a pre-gen AI company that benefited enormously from the gen AI wave. The number of applications being deployed to Render ballooned starting end of 2024, accelerated through 2025, and spiked sharply in early 2026 — driven by tools like Claude Code and Codex enabling developers to build faster than ever.

Two tailwinds converged. Individual developers could finally build all those side projects they'd always wanted to tackle. And within companies, DevOps teams — already stretched — found themselves completely overwhelmed by the volume of AI-assisted code being produced. Self-serve platforms like Render became the logical answer.

Render is also adapting its product to match the new primitives AI apps require. Render Workflows allows agentic applications to run through multi-step processes with retries, concurrency limits, and observability. Sandboxes — for running untrusted, machine-generated code in isolated environments — are next. Goel notes Render has been running untrusted code at scale for years, giving them a natural advantage here.

"The number of applications being created has ballooned, exploded. People are able to build things now that they always wanted to build, but didn't have the time for, the resources for." — Anurag Goel

"Our revenue has just changed. It's like even at a much higher scale of revenue, we're growing much faster than we did before." — Anurag Goel

Render vs. Hyperscalers: Who They're Really Built For