Many startups and large tech companies have tried to build artificial intelligence for coding software. Another new player has emerged to solve the problems that will arise if humans and AIs write code together.
Tessl has opened a waiting list for those who are interested in trying out its “AI native” platform.
When we say “is building”, it’s because Tessl hasn’t launched its product yet. The plan is to launch it early next year. The London-based startup has now revealed a little more about its work with some financial fanfare. Tessl quietly raised $125M across a Seed round and a series A, which were both announced for the very first time today. Index Ventures led the latest round, with Accel and GV also participating. Boldstart and GV (Google Ventures), together, led the seed round.
TechCrunch confirmed with multiple sources, Tessl’s post-money value is north of $500m, specifically $750m.
You might have guessed that the reason why a startup without customers or a product shipped is getting so much attention from top-shelf VCs, is because it is being built by a developer whisperer. Guy Podjarny is the founder and CEO of Tessl. He’s a developer whisperer. His last startup, Snyk was a cybersecurity company that was valued at $7.4billion in 2022. He was CTO at Akamai before that. Akamai had acquired Blaze, his first startup which focused on accelerating website loading times. Carlos Gonzalez-Cadenas is the Index partner who led the investment. He said that Podjarny was “incredibly visionary and thoughtful” about his business. “He is very, very good at [at understanding] building developer-oriented business and developing developer communities.”
Podjarny stated in an interview that Tessl was a result of his experience with Snyk.
Snyk focuses primarily on detecting and fixing security vulnerabilities in code. Podjarny noted that a similar problem was becoming more urgent in the software industry, especially because of the rapid growth of code generated by AIs.
He asked himself, “What does AI do to software development?” The answer: it would speed up the process, but also create more software automatically. The complexity of the code and the chances of it breaking would increase as a result of the process of updating and maintaining that code. This has a number of negative implications for organizations (security, uptimes, cost, efficiency). “The more this picture formed in my head, the more I knew that I would build it,” he said.
According to Podjarny, the startup’s name Tessl is a reference “tessellation” as it aims for software and code to fit together neatly, rather than existing in a messy jumble.
Podjarny refused to reveal what applications or code he envisages being developed or maintained on Tessl. It sounds like the project will start small.
He said, “We haven’t shared the full strategy on what this is yet,” in regards to target applications or use-cases. “I would say we are not starting with games. We’re beginning with relatively simple software, which allows us to build a system that is easier for LLMs and humans to specify. We will then evolve from there.”
Tessl’s basic idea is that developers and their teams, which include product managers and other non-coders, can provide specifications in either natural language or code. Tessl then writes code to match the specifications.
Teams are able to test the code in a Sandbox where they can flag and address issues, and continue to modify specifications as needed. Tessl will then automate the maintenance of that code according to the specification. Tessl will then run remediation in order to fix any other issues that may arise as a result of the new code.
From the sound of it, Tessl does not seem to be a walled-garden. Podjarny said that Tessl would initially support Java, JavaScript and Python and will add more languages in the future. Podjarny also said that he is in contact with others who are developing AI coding assistants. He hopes to ensure that the work of these platforms can be maintained using Tessl. This means that, while it may theoretically compete with other platforms such as Anysphere’s Cursor and Poolside, GitHub Copilot, Magic Codeium, Augment or OpenAI, IBM and others, it can also work with any team.
As a side note, many of the startups on this list have valuations that are approaching 10 figures. This is another reason why Tessl’s valuation does not seem out of place. Could this be another bubble, or will these startups be able to justify their numbers on paper?
This extensibility is one of the reasons why investors are so excited about this idea. Gonzalez-Cadenas said that maintaining code is a topic that is “very important” right now. “But he is building a record system here,” he said. “Once you’ve done that, there are a variety opportunities.”