A startup whose product competes with GitHub Copilot and different AI-powered coding assistants has achieved unicorn standing.
On Thursday, Codeium stated it closed a $150 million Series C spherical led by General Catalyst that values the corporate at $1.25 billion post-money. The spherical, which additionally noticed participation from current traders Kleiner Perkins and Greenoaks, brings the corporate’s whole funding raised to almost a quarter-billion {dollars} ($243 million) a mere three years since its launch.
Codeium’s co-founder and CEO, Varun Mohan, advised TechCrunch that Codeium hasn’t even touched the $65 million Series B tranche it raised in January but. Back then, simply eight months in the past, Codeium was valued at half a billion {dollars}.
“Even although we’ve barely made a dent in our current funding, we consider that this injection of capital will permit us to considerably ramp up R&D and progress whereas making even bigger strategic bets,” he stated.
Codeium was based in 2021 by Mohan and his childhood buddy and fellow MIT grad, Douglas Chen. Prior to Codeium, Chen was at Meta, the place he helped to construct software program instruments for VR headsets just like the Oculus Quest. Mohan was a tech lead at Nuro, the autonomous supply startup, chargeable for managing the autonomy infrastructure crew.
The startup started as a radically totally different firm referred to as Exafunction, centered on GPU optimization and virtualization for AI workloads. But in 2022, Mohan and Chen sensed an even bigger alternative in generative coding and determined to rebrand — and pivot.
“Despite the inflow of generative AI instruments, builders are nonetheless scuffling with time-consuming coding duties,” Mohan stated. “Many of the AI-driven options present generic code snippets that require vital handbook work to combine and safe inside current codebases. That’s the place our AI coding help is available in.”
Codeium’s platform, powered by generative AI fashions educated on public code, serves up ideas within the context of an app’s whole codebase. It helps round 70 programming languages and integrates with various in style improvement environments, together with Microsoft Visual Studio and JetBrains.
To appeal to devs away from Copilot and different rivals, Codeium has launched a beneficiant free tier to start out. The technique appears to have labored: Today, the startup has greater than 700,000 customers and over 1,000 enterprise prospects, together with Anduril, Zillow and Dell.
Quentin Clark, managing director at General Catalyst, implied that Codeium received a few of its bigger contracts by embracing a steadfastly client-centric method to product analysis.
“The crew’s method has all the time been to observe its prospects, main the corporate to construct options on their phrases — deployable in any atmosphere and supporting extra languages than anybody else,” Clark stated in an announcement. “What Codeium has created isn’t only a demo, an announcement, or an concept — this can be a absolutely scaling enterprise, with massive enterprises adopting the product throughout their whole organizations.”
Businesses are sometimes cautious of exposing proprietary code to a 3rd occasion — as an example, Apple reportedly banned workers from utilizing Copilot final yr, citing issues about confidential knowledge leakage. To try to allay such fears, Codeium started providing a self-hosted set up choice alongside its customary software-as-a-service plan.
Companies can now deploy Codeium’s service on their very own {hardware} if they need. Or they will undertake a hybrid setup, preserving their knowledge on their very own gadgets whereas utilizing Codeium’s servers for computing wants.
There’s all the time some threat concerned in knowledge transfers to the cloud, however Mohan claimed that Codeium leverages robust encryption. “We by no means prepare our proprietary generative autocomplete mannequin on consumer knowledge, by no means promote knowledge and guarantee all knowledge transmission is encrypted,” he added.
Codeium has additionally taken steps to take away “non-permissively” licensed code (e.g., code below copyright) from the datasets it used to coach its AI fashions. Some code-generating instruments educated utilizing restrictively licensed or copyrighted code have been proven to regurgitate that code when prompted in a sure means, posing a legal responsibility threat (i.e., builders that incorporate the code could possibly be sued). Mohan stated that’s not the case with Codeium, due to its coaching knowledge prep and filtering method.
“We additionally take away any remaining knowledge that appears much like code that’s explicitly non-permissively licensed simply in case different folks copied code with out offering the right attribution and licensing,” he added. “On high of this, now we have state-of-the-art, post-generation attribution filtering and logging within the case that these massive probabilistic fashions produce code that’s much like public code, whether or not permissively or non-permissively licensed.”
But what about hallucinations? Most AI coding instruments are infamous for making stuff up, which might be fairly damaging in an enterprise atmosphere.
An evaluation by developer tooling startup GitClear discovered that generative AI instruments have resulted in extra mistaken code being pushed to codebases over the previous few years. And a Purdue research discovered that over half the solutions that OpenAI’s ChatGPT offers to programming questions are incorrect. Security researchers have warned of the potential for such instruments to amplify current bugs in software program.
A current survey from cybersecurity agency Snyk discovered that 9 in ten builders fear in regards to the broader safety implications of utilizing AI coding platforms. But Mohan claimed that Codeium’s supposedly superior, deep context-rich tech yields extra reliable outcomes than most.
“Our context consciousness engine is ready to floor ends in what’s already current in a consumer’s codebase, resulting in ideas with fewer hallucinations and extra adherence to current syntax, semantics and requirements,” he stated.
Whether benchmarks again that up or not, Codeium’s gross sales pitch appears to be resonating with the suitable execs: Revenue hit eight figures this yr. Mohan stated the 80-person, Mountain View-based startup plans to broaden headcount to 120 by 2025 because it goals to make an even bigger dent in a market with formidable opponents like Tabnine, Anysphere and Poolside.
Catching as much as Copilot, which had over 1.8 million paying customers as of April, in all probability isn’t within the playing cards for Codeium — no less than not imminently. It doesn’t must be. As Mohan rightly famous, given the widespread adoption of AI coding instruments amongst builders (regardless of their reservations), even a small slice of the nascent section is certain to be profitable.
Polaris Research tasks that the AI code instruments market will probably be price $27.17 billion by 2032.
“An overabundance of hype is a problem the trade faces,” Mohan stated. “This will make it more durable for each firm to actually persuade finish customers that they’re on the forefront of risk. But we consider that truth-seeking and sensible AI corporations like Codeium will finally minimize by means of this noise.”
