Replit’s Amjad Masad weighs in on the Cursor deal, the fight with Apple, and why not to sell

Amjad Masad has been building Replit for ten years, but the last 18 months have been something else entirely. The AI ​​coding assistant company is expected to generate $2.8 million in revenue in 2024, ahead of Masad’s billion-dollar annual runaway pace.

TechCrunch sold out The StrictlyVC event On Thursday night in San Francisco, we covered a lot in a short amount of time, starting with the question everyone in the industry is asking right now: a world where rival Cursor is in talks to be acquired by SpaceX. 60 billion dollarsIs Replit bound to sell too? We also looked at Replit’s net profit margin — a measure of how much existing customers are stretching their spending — at 300%, according to Masad, which he says is due to his desire to sue Apple, the App Store battle with Replit, and the company’s start to invest in its customers.

On the issue of independence, Masad gave an unambiguous speech. Unlike Cursor, which he said was operating at a negative 23% gross margin, he argued that Replit had the economics to make it that way — even if it rejected the sale entirely.

The following has been edited for length and clarity:

TC: The SpaceX deal Cursor reported was the talk of the industry last week. What did you get out of it?

AM: Being an independent, small AI company that builds on core models is a bit difficult, especially if you’re burning through a ton of cash. Part of the proposed report is that Cursor has negative 23% margins, and if you also want to invest in training models, that makes it incredibly difficult to stay independent.

For us at Replit, we’ve been able to run things more streamlined, in part because we’re targeting a different set of customers. Gross margin has been positive for more than a year. We are a little more expensive, but we provide more. Our audience is mainly non-technical users who have never created any software before. We provide an end-to-end platform – from query to scalable implementation. We handle security, database, database migration. And we’ve been doing it long enough that we’ve built a lot of those primitives into the platform.

Replit for sale? I assume you talk to potential buyers all the time; it is your fiduciary responsibility.

Yes. We have amazing partners and they sometimes bring up these topics. But we will try to remain independent. I would like us to remain an independent company. We’ve been around for 10 years before it was accepted that you can build apps from just ideas. At YC in 2018, we were talking about creating a billion app creators, and people sometimes laughed at that dream. Now this dream is possible and we started this revolution in September 2024 with our agent coding experience. It just feels like we can take it further.

You work closely with Anthropic, Google and OpenAI. If you had to rank them – who does it best?

Anthropic is still undefeated in the main agent circuit. They have the best tool call; the agent can remain coherent longer. GPT-5 is maturing quickly. Google’s family of Flash models is amazing in terms of price-performance. If you want something fast and cheap, they actually beat open source. We use all three, and honestly, I wouldn’t discount the new labs either. Reflection AI comes out with open source models that we’ve heard great things about. And the Chinese models are impressive – Kimi is as good as the Anthropic generation model from January, so only three months behind.

When you’re ready for an enterprise deal, what makes you do it?

The majority of our sales are inbound or organic – leading multiple products. We’ve gotten customers like Zillow and Meta just through people adopting the product and then reaching out to buy an enterprise plan. When it goes from top to bottom and there is formal baking, we usually win the product. But even in cases where we might miss a feature, once the C-suite and IT team are tapped, Replit wins in security. A lot of vibe-coding tools will create a website and connect it to an external database — great stuff, but it makes security much more difficult because the database is public and you have to configure layer-level security, which is especially difficult for non-technical builders. Replit is full-stack, the database is embedded in the project and not publicly available – making the program inherently more secure.

We’ve also spent 10 years fighting crypto fraudsters and hackers, so our cybersecurity feature is as good as a dedicated cybersecurity startup. Every time you deploy an app to Replit, we create a brand new isolated project in Google Cloud. We inherit Google’s security model.

Can we talk about rot? How long do you keep customers if the best prototypes end up being rebuilt into a company’s existing stack?

The decay is very, very low and the net retention is incredibly high – in some cases 300%. What we hear from customers is that when engineers get frustrated and try to rebuild the software into their own stack, they often end up making it worse. Once enterprises are comfortable with the full Replit stack – especially when we build a single-tenant environment for them – they keep applications in Replit. For example, Bain & Company replaced Tableau and Power BI with Replit and Databricks.

There is growing concern about AI bloat – non-technical users creating more code and burning more tokens. That’s good for you (given the usage-based fees). What about your customers?

We have no expenses which is a pity. Businesses know their ROI very well, and they tell us about their returns. In most cases, they think the investment is totally worth it – often by one, two, three orders of magnitude. If they’re spending $100,000 a month with Replit, they’re usually $2 million, $3 million, $10 million in revenue.

Let’s talk about Apple. Another competitor, Lovable, just won software to create software it was approved by the App Store this week. Replit has been in the App Store, Apple has been blocking your updates for months. How much does this bother you?

It’s not life or death – we could lose the app and it wouldn’t do anything meaningful for our business. But it’s an app that people really love. We’ve been on the App Store for four years. Children in low-income communities learn to code on Replit on their Android devices. Chief Executives use it in meetings.

The reason Replit is blocked when others aren’t is because Replit makes iOS apps. When we launched this capability in December, we had charts showing how many apps were getting into the App Store through us. We think Apple feels threatened by this.

The reason Apple mentions is because you’re uploading new code to the device (after the verification process) which violates their guidelines.

This is a lie. And if necessary, we can prove it in court.

Will this happen?

I hope not. I’m a fan of Apple and would love to collaborate and build something great together. We are happy to refer customers to Xcode (Apple’s own development environment). But you can’t control a market with access to a billion people and make decisions based on discrimination or whim.

I wonder if you would consider investing in your own customers in exchange for equity, like Nvidia, OpenAI and others.

We have thought about this a lot and it should be taken into account. I have personally invested in several startups that started on Replit before they made money. Some of them, like Magic School – a teacher decided to take time to learn some vibe coding during COVID and built an AI app for other teachers. He found the problem is that in America we burn out too many teachers. He wanted to use artificial intelligence to reduce his workload. He did and made $20 million in the first year. Other companies that have started on Replit are valued at, I think, half a billion dollars. The entrepreneurship happening at Replit right now is really exciting. We integrated with Stripe a few months ago and transactions through Replit are growing triple digits per month. Soon our customers will earn more than us.