Highlights:

  • Poolside has created generative AI coding models to assist developers with code auto-completion and context-specific code suggestions.
  • The startup is also developing a custom software stack, or toolkit, for AI training tasks that include features allowing researchers to manage the datasets for their machine learning projects.

Recently, artificial intelligence coding startup Poolside Inc. raised a USD 500 million Series B funding round. Poolside is a competitor to GitHub Inc.’s Copilot.

Bain Capital Ventures led the recent funding. It included participation from several globally recognized companies, such as eBay Inc., Nvidia Corp., Citi Ventures, HSBC Ventures, and LG Technology Ventures. Additionally, venture capital firms like DST Global, StepStone Group, Schroders Capital, Premji Invest, Dorsal Capital, BAM Elevate, Adams Street, and Fin Capital also joined the round. The startup’s existing investors, Felicis Ventures and Redpoint Ventures—who previously led its USD 100 million Series A and USD 26 million seed rounds, respectively—were also part of the extensive list of backers.

The funding round brings Poolside’s total capital raised to USD 626 million, valuing the company at an impressive USD 3 billion, according to sources.

The extensive list of investors and the significant capital raised indicate that Poolside is onto something exceptional with its generative AI-powered coding assistant models, positioning them as a strong alternative to GitHub’s popular Copilot. There’s reason to believe its technology could be equally good, if not superior, given that Poolside’s Co-founder and CEO, Jason Warner, previously served as GitHub’s Chief Technology Officer and led engineering teams at Canonical Ltd. and Heroku Ltd.

Co-founder Eiso Kant, who currently serves as Poolside’s CTO, also boasts an impressive background, having founded several startups focused on developer technology in the past.

Poolside has created its own generative AI coding models aimed at assisting developers with tasks like code auto-completion and suggesting relevant lines of code for specific contexts or codebases. The startup reports that it has several Global 2000 companies as clients, although it has only revealed a few of them on its website.

Poolside distinguishes itself from other AI coding competitors with its reinforcement learning approach to training its models. This method allows the models to learn continuously from feedback based on code execution, enabling them to improve consistently as they handle millions of tasks in real-world software projects.

The startup is developing a custom software stack or toolkit for AI training tasks equipped with features that allow researchers to manage the datasets used in their machine learning projects. Additionally, the software includes tools for automating tasks like measuring the accuracy of newly created large language models.

In a statement, Warner expressed his belief that only a select few companies will ultimately attain what is referred to as “artificial general intelligence” (AGI), a more advanced type of AI that can think and learn independently, akin to human capabilities. The potential of AGI remains a topic of controversy, and it is precisely what Poolside asserts it is working to develop.

“We believe software development will be the first broad capability where AI will reach and surpass human-level intelligence. Through our team, our applied research and a powerful revenue engine, Poolside will bring AI for software development so that anyone in the world can build,” Warner said.

Poolside reported that the funds from the recent round are already being put to use, with an investment in 10,000 Nvidia graphics processing units to train its next-generation models. The company also noted that the capital will support its go-to-market strategies and research and development efforts.

Despite ongoing concerns regarding the security and reliability of AI code generation tools—many of which are believed to plagiarize existing code—developers have expressed significant enthusiasm for these technologies. Sources said that a large majority of respondents are using AI tools to support their development efforts. Additionally, GitHub has garnered over 1.8 million paying users for its Copilot assistant.

Poolside is not the only AI coding startup attracting significant investment from venture capitalists. Its competitors include Magic AI Inc., which secured USD 200 million in new funding in July, as well as Augment Inc., which completed a substantial USD 227 million Series B round that raised its valuation to USD 977 million. Additionally, Cognition AI Inc., the creator of Devin, finalized a USD 175 million funding round led by Founders Fund in March, increasing its valuation to USD 2 billion.

Companies like Google LLC and Amazon Web Services Inc. also compete in this space with their CodeWhisperer and Gemini Code Assist tools. Moreover, some startups are in the mix, like Tabnine Ltd., Codegen Inc., Refact, Laredo Labs Inc., and TabbyML Inc.