Overflowing Stack: AI Coding Tools [Q2'24]
AI coding tools have proven to be one of the most successful applications of generative AI across industries, with 81% of developers now using AI assistants to streamline their coding.
We initiated coverage on AI coding tools in May’23 and here’s what we noticed then:
- Big tech players looked very well positioned – Microsoft & OpenAI had established a strong base; Google & Amazon were poised for their entry into AI coding.
- Fig, Redocly, Mintlify, Cursor, BlackBox, and Warp stood out, with the first four primed for funding rounds in 2023.
- Open-source participation was thriving – Google’s Gemini and a rumored OpenAI open-source model were set to shake up the market.
Fast forward to mid-2024, the landscape has evolved significantly. While GitHub Copilot continues its impressive trajectory, Google’s AI program has been underwhelming. Open AI models have not become the staple like we thought; with Claude, Gemini being top rated and widely used. Meta’s OS model CodeLlama is popular amongst devs, with many companies finetuning their own versions on top of it. AI agents have been wildly successful in the coding space, unlike the underwhelming response to agents for productivity / writing.
Here’s a look at the latest developments in the AI coding tools space – updated insights on new players, use cases and investors shaping up the future of software development.
Market Map: AI Agents are on the rise, non-US geographies gaining ground
Active playground with 250+ startups, some specialised use cases seeing innovation
The AI coding tools space is booming, with more than 250 startups launching innovative products aimed at several use cases. ~30% of the total startups in the space were founded in 2023-24. Code generation remains the hottest use case, but downstream use cases are gaining ground – such as debugging (CodeRabbit), optimization (Typo), and code maintenance (Grit). Specific use cases such as web crawling (Indexical), frontend dev (Xamun), SQL writing (Vanna AI) are also seeing startup activity.
AI agents have become increasingly valuable and effective for coding
Agentic solutions are rapidly entering the AI coding tools market. These startups focus on empowering developers by automating repetitive tasks, improving code quality, and optimizing workflow efficiency. Unlike traditional AI co-pilots that assist developers, AI coding agents operate autonomously in the background, enabling engineers to supervise multiple “AI engineers” simultaneously. Startups like Factory AI, CodeGen, Pythagora, and Tusk are pioneering this space by developing tools capable of executing multi-step processes, integrating product thinking, and handling tasks based on contextual inputs from sources like Jira tickets and codebases.
Non-US geographies emerging as key hubs for AI innovation
While the U.S. remains a major player, countries like the UK, India, Israel, France, and Germany are emerging as hotbeds for startup activity, with 30% of bootstrapped, pre-seed, and seed-stage companies originating from these non-US regions.
Specialized foundational models for coding hasn’t been a theme
Current text-based AI models have done really well in code generation – GPT, Claude, Mistral, Llama are widely favoured by developers. The focus on specialized coding models hasn’t emerged as a major theme, as base models continue to perform exceptionally well.
Funding Landscape: 60+ deals in the last year; Coding Copilots attract the lion’s share
Robust funding for AI coding tools in the last 12 months
Since May 2023, the AI coding tools market has witnessed 60+ deals, with copilot and assistant tools attracting over 75% of the total funding dollars. Y Combinator has been particularly active in the sector, backing ~30 startups.
Standout startups from our 2023 market map have attracted large deals
Fig, Redocly, Mintlify, Cursor, BlackBox, and Warp stood out in our 2023 research as top candidates for future funding. Since then:
- Cursor raised a $60M Series A, with a16z, OpenAI as investors
- Warp secured a $50M Series B led by Sequoia
- Mintlify has raised an %18M Series A led by a16z
- Fig has been acquired by Amazon
Notable M&A activity taking place in AI coding
The AI coding tools market is also seeing increased consolidation, with notable acquisitions like Fig (by Amazon), CodeStream (by New Relic), and Reviewpad (by Synk).
Github Traction: Zed.dev, Continue.dev and Vanna AI seeing impressive traction
Zed, an AI code editor backed by Redpoint Ventures, has seen remarkable traction on Github – with the highest number of new stars in Q2’24. Zed provides a high-speed, lightweight environment for developers to work on the same codebase simultaneously, with support for various programming languages.
Continue is an open-source AI assistant that integrates into VS Code and JetBrains, offering features like code completion, debugging, and error correction directly within the IDE. The startup has seen some solid developer interest, with >200% growth in Github stars over the last 6 months. Continue is backed by Y Combinator.
Vanna AI is an open source package that allows non-technical users to query datasets using natural language, converting these inputs into SQL commands. The platform is designed to make database interaction more accessible for users without technical expertise.
Web Traction: Factory AI seeing strong consumer interest with >300% QoQ growth
Factory AI creates autonomous agents that handle repetitive coding tasks and optimize developer workflows. Factory AI creates "Droids" that help organisations systematically ideate, plan features, write code, write tests, review code, and write documentation – saving upto 561K engineering hours. The startup has seen a whopping 392% QoQ growth in website traffic in Q2’24, indicating increased interest in its AI-driven approach to task automation.
DocuWriter is an AI-powered tool designed to automate code documentation, testing, refactoring, and conversion. It saves developers substantial time while improving code quality and maintenance.
Text2SQL converts natural language queries into SQL, providing a solution for users who need to interact with databases without using complex SQL commands. The tool aims to simplify database querying for non-technical users.
Looking Forward
- AI Agents delivering strong value in Coding
Across most AI application, “Assistants” i.e. suggesting tools, have been more successful than “Agents” i.e. automated execution. Coding is seeing a strong uptake in agentic tools. Startups like Factory and Pythagora are pushing toward fully autonomous systems that could manage entire coding processes. As these tools evolve, the shift from AI copilots to full-fledged AI “engineers” will be interesting to watch out for.
- AI applications for downstream use cases in software dev are rapidly gaining traction
Use of AI is expanding beyond code writing to use cases across the software development lifecycle, such as areas like code efficiency (Million) and automated bug resolution (CodeRabbit). Specific solutions are also targeting quality-of-life improvement, e.g. frontend development (Xamun), web crawling (Indexical), SQL generation (Vanna AI).
- Winners are emerging, with strong traction from developers as well as investors
Cursor, Codeium continue to show impressive progress and have grabbed investor eyeballs. New startups like Augment, Factory AI and Zed are making their mark in a fast-moving market. Factory AI, Zed are likely to raise funding in the near future. We expect to see downstream innovation from future entrants in the space (for code maintenance, observability).
- Expect intensifying competition in the next few years
Widespread developer adoption, low switching costs, decreasing costs and multiple foundational model options will ensure numerous opportunities for emerging startups as well as indie developers coming up with interesting solutions.
Interested in these startups? Download a detailed list of 75+ high-growth AI coding startups featured in our research – with founder profiles, funding, HQ and more.