The AI Coding Race: Can Emergent Outcode Claude And Gemini?

“The future belongs to those who can imagine it.” — Dr. A.P.J. Abdul Kalam

For Mukund Jha and Madhav Jha, that line is more than just a maxim. The brothers have always believed in this vision from the first time they touched a computer keyboard in the mid-90s.

This passion for coding and programming — born out of a love for video games — led to their journey to Silicon Valley and back to India, where Mukund eventually became a cofounder at hyperlocal startup Dunzo.  

Armed with the formative experience of working at Google, Amazon and Dropbox, one could say that the Jha brothers have always had one eye on the future. But neither of them could have foreseen today’s reality — coding and programming are increasingly becoming the domain of AI models. 

Their shared love for coding and developing software has led to Emergent Labs, a GenAI coding platform that is looking to capitalise on the ‘vibe coding’ wave. 

Essentially, Emergent Labs enables anyone, even individuals with zero technical or coding knowledge, to build fully functional apps using AI-generated code. 

Launched in mid-2025, Emergent Labs has already raised $30 Mn and amassed 3 Mn users, but Mukund Jha calls it ‘barely scratching the surface’. In the era where software drives nearly every business and idea, he believes that app development should not be limited to software engineers. 

Different Paths, Shared Obsession

After completing their post-graduation in India, the Jha brothers went to the US to pursue PhDs in computer science at separate universities. While Madhav completed his doctorate, Mukund dropped out after one semester to become an intern at Google in 2010. 

Years later, Google became an investor in Mukund’s major entrepreneurial venture, Dunzo, which eventually shut down after a turbulent time in 2023 and 2024. 

“For both of us, Google helped us understand how to build things at scale, what good tech engineering actually means. Google helped us get into search, data, machine learning, pattern recognition, basically all the fundamentals for modern day AI,” Mukund recalled.

One thing never changed for the brothers though: their curiosity about the potential of software and programming.  As Mukund told us, the two often exchanged notes, ideas, code snippets, and kept up with the pace of machine learning and AI evolution. 

Their fascination peaked in 2018, when OpenAI released early LLMs, and in the past five years, all manner of AI-powered platforms have emerged in the aftermath. 

“We were super nerds about these. Along with our daily chores, our interest in AI and machine learning made us burn the midnight oil to learn more about the tech. I called up Madhav and said, ‘let’s build something together’,” Mukund told Inc42.

As Dunzo’s fortunes turned, Mukund looked at the new AI opportunity. By early 2024, the brothers converged on the audacious idea of automating coding and software testing. 

The first step was applying to Y Combinator in the hope of raising funds and getting acceleration support for an AI platform that automates software testing. 

Even without a business plan, Emergent was accepted into the Summer 2024 cohort of Y Combinator, and soon after also raised $5 Mn from Girish Mathrubootham’s Together Fund. 

And this also allowed the duo to narrow focus. Instead of automating software testing entirely, Emergent honed in on a simpler question: What is the smallest component of software development that can be automated? 

The answer was app development, paving the way for Emergent as it currently stands.

Building Emergent Out Of Thin Air

In the early days, Emergent was an AI agent that could solve bugs in lines of code. The startup benchmarked it using the newly introduced Software Engineering Benchmark (SWE) for AI models. In this, an AI agent must autonomously solve 500 GitHub bugs.

Within two months, Emergent took the top spot on the SWE leaderboard. This validation cemented the conviction of the founders that software automation wasn’t a distant dream.

“Seeing our agents solve these challenges made us realise where the world was heading. It was clear that software automation was the next logical step,” Mukund, who is the CEO of Emergent, said.

A private beta was launched in February 2025, followed by an official launch in June 2025. 

Emergent’s vision is different to the likes of Claude Code, OpenAI’s Codex, Google’s Gemini AntiGravity or Cursor, other AI-powered coding platforms and integrated developer environments. Simply put, the two brothers wanted to abstract the entire software development lifecycle for non-technical users.

This led to a ground-up approach for the platform — non-technical users, Mukund claims, do not need to understand compiler errors or navigate through code logs. So Emergent has deployed AI agents to fix these bugs or problems silently in the background. 

“As a user, you only need to focus on what app you are building. The cost of building an app for an individual is so high that most will not build, but what we are trying to do is democratise it,” Mukund said.

In just eight months after launching in beta, Emergent has grown to a team of 40, with employees in India and the US. The company claims to have 3 Mn active users, of which around 60,000 are said to be paying users.

Emergent’s user base largely consists of solo founders, small business owners, and first-time entrepreneurs, with popular use cases being logistics apps, marketplace apps, and chatbots.

The Evolution Of AI Coding

One of the biggest questions being asked of AI platforms is how they will monetise the hype around AI products. 

Emergent charges a monthly subscription fee ranging from $20 to $200, depending on how much of the platform an individual user wants to utilise. Users also pay a hosting fee to deploy their apps on Emergent’s infrastructure.

With the current traction, the startup expects an ARR of around $25 Mn for FY26. 

Mukund revealed that the startup is also exploring additional revenue streams, such as enabling developers to pay for improved discovery of their apps, similar to an app marketplace.

One big friction for Emergent at the moment is that deploying AI-coded apps on the Google Play Store and Apple’s App Store is perhaps not as easy as creating them. This challenge is universal across so-called ‘vibe coding platforms.

Mukund acknowledges this gap. “We are ironing out this issue,” he said, noting that 15%-16% of Emergent’s users currently have mobile apps.

Looking ahead, Emergent plans to deepen its infrastructure capabilities, strengthen data security, and stay ahead of rapid AI advancements.

As AI transforms every layer of software development, Mukund believes that the future belongs to platforms that allow anyone, regardless of technical background, to build, launch, and scale apps.

But there is one big caveat: the competitive intensity is particularly high and growing for AI applications. While the likes of Emergent have shown that having a clear go-to-market strategy can unlock early revenue, building on this means competing with the likes of Anthropic’s Claude, OpenAI’s Codex, Gemini and others. 

These companies have dominated the global AI narrative in the past three years, and many see them as challenges for this new wave of applications. In particular, the deep pockets of tech giants and consistent support from the investor and private equity ecosystem has led to a situation where many new startups might fall by the wayside, unless they consistently innovate and create more value faster than bigger platforms. 

After all, once AI coding models become more efficient and accurate, there’s no moat between players. This is a big risk not just for Emergent but other applications built on AI models. 

Emergent Labs may be young and with a clear mission — to ease app development. As AI reshapes how software is built, can the startup keep up with the AI giants?

Edited By Shishir Parasher & Nikhil Subramaniam