Eric Schmidt-backed Augment, a GitHub Copilot rival, launches out of stealth with $252M
The integration of AI into coding processes is gaining momentum, with developers increasingly adopting AI tools to enhance their workflows. According to a recent Stack Overflow survey, 44% of software engineers currently utilize AI tools in their development practices, and an additional 26% plan to do so in the near future. Gartner predicts that over half of organizations are either experimenting with or have already implemented AI-driven coding assistants, with an estimated 75% of developers expected to leverage coding assistants by 2028.
Former Microsoft software developer Igor Ostrovsky envisions a future where AI becomes ubiquitous in developer workflows, improving software quality and team productivity while rejuvenating the coding experience. Motivated by this vision, Ostrovsky set out to create Augment, an AI-powered coding platform designed to streamline development processes.
Augment recently emerged from stealth mode, securing an impressive $252 million in funding and achieving a near-unicorn post-money valuation of $977 million. With backing from prominent investors such as former Google CEO Eric Schmidt and top-tier VCs including Index Ventures, Sutter Hill Ventures, Lightspeed Venture Partners, Innovation Endeavors, and Meritech Capital, Augment aims to disrupt the emerging market for generative AI coding technologies.
According to Ostrovsky, many organizations face challenges with software fragility, complexity, and high maintenance costs. Augment is positioned to address these issues by empowering programmers and organizations to deliver high-quality software more efficiently. With Ostrovsky’s extensive background, including nearly seven years at Microsoft and experience as a founding engineer at Pure Storage, Augment is well-equipped to drive innovation in the field of AI-driven coding technologies.
In 2022, Ostrovsky and Guy Gur-Ari, formerly an AI research scientist at Google, joined forces to develop Augment’s MVP. To bolster the startup’s leadership team, Ostrovsky and Gur-Ari enlisted Scott Dietzen, former CEO of Pure Storage, and Dion Almaer, previously a Google engineering director and VP of engineering at Shopify.
Despite Augment’s significant progress, the company remains largely secretive about its operations. Ostrovsky declined to divulge many details about the user experience or the specific generative AI models powering Augment’s features. However, he did mention that Augment utilizes finely-tuned “industry-leading” open models of some kind.
Regarding revenue generation, Augment plans to adopt standard software-as-a-service subscription models. Ostrovsky indicated that pricing and additional information will be disclosed later this year, closer to Augment’s planned general availability release.
With ample funding secured, Augment is poised for substantial growth, aiming to assemble an exceptional team in enterprise AI. The company is focused on accelerating product development and expanding its product, engineering, and go-to-market functions to facilitate rapid expansion.
However, Augment faces stiff competition in a fiercely competitive industry landscape. Tech giants like Microsoft, Amazon, and Google have already introduced their own AI coding assistants, while numerous coding assistant startups continue to emerge. Despite this, Augment is determined to carve out a niche and make a significant impact in the market.
Ensuring harmonious collaboration among all players in the AI coding assistant space, including Augment, appears challenging. The substantial compute costs associated with AI coding assistants pose a significant obstacle to sustainability. For instance, Kite, a generative AI coding startup, had to cease operations in December 2022 due to overruns related to training and serving models. Even GitHub Copilot, a leading player in the market, operates at a loss, with costs ranging from $20 to $80 per user per month, according to The Wall Street Journal.
Ostrovsky suggests that Augment is gaining traction, with “hundreds” of software developers across “dozens” of companies, including payment startup Keeta (also backed by Eric Schmidt), already using Augment in early access. However, the key question remains: can this initial uptake be sustained in the long run?
Moreover, there are concerns regarding the technical challenges inherent in code-generating AI, particularly regarding vulnerabilities. Analysis by GitClear, a code analytics tool developer, indicates that coding assistants may contribute to an increase in erroneous code being introduced into codebases, posing challenges for software maintainers. Security researchers caution that generative coding tools have the potential to magnify existing bugs and vulnerabilities in projects. Additionally, Stanford researchers have observed that developers who rely on code recommendations from AI assistants tend to produce code with lower security standards. Addressing these technical setbacks will be crucial for Augment and its counterparts to ensure the integrity and security of software development processes.
Then there’s copyright to worry about.
Augment’s models were trained using publicly available data, a common practice in the development of generative AI models. However, some of this data may be subject to copyright or restrictive licenses. While some vendors argue that fair use doctrine protects them from copyright claims, they have also implemented tools to address potential infringement. Nevertheless, coders have filed class action lawsuits alleging open licensing and intellectual property violations.
In response to these challenges, Ostrovsky emphasizes that current AI coding assistants often fall short in understanding the programmer’s intent, improving software quality, facilitating team productivity, and protecting intellectual property. Augment’s engineering team, comprising experts in AI and systems, is committed to introducing innovations in AI coding assistance to developers and software teams.
Headquartered in Palo Alto, Augment currently employs around 50 individuals, with Ostrovsky expecting this number to double by the end of the year.