AI-based coding has exploded in popularity with the promise that developers will work faster and easier. But it brought something else: a significant increase in the line of code, and therefore the possibility of bugs that lead to crashes and other accidents. Today, an Israeli startup called Lightrun has built an observability platform to identify and debug (fix) code before these issues arise – is launching a $70 million Series B. The funding highlights not only the market gaps for such tools, but Lightrun’s traction in meeting its demand.
New backer Axel co-leads the round alongside former investor insight partners, along with participation from City, Grillot Capital, GTM Capital and Sorenson Capital. Lightrun is currently raising $110 million, including a Series A led by Insight, covered in 2021.
The startup has not disclosed its ratings, but there are some strong indications that it is doing well.
First, there are customers. Citi is a strategic backer and is one of an impressive list of well-known clients, including ADP, AT&T, ICE/NYSE, Inditex, Microsoft, Priceline, Salesforce and SAP.
Secondly, there is the product and company timing as to how it fits into the current market landscape. In July 2024, Lightrun announced a new AI-based debugging tool for use within your organization’s Integrated Developer Environment (IDES), aptly referred to as the Runtime Autonomous AI Debugger. The company’s platform already provided impressive results, but this was a product that actually spoke to the current predicament that many companies face. AI has led to more coding and more problems, and Lightrun has built AI tools to address it.
The company said it has 4.5 times its revenues have been inaugurated, which is why it knocked investors. Andrei Brazoveanu, the Axel partner who led the company’s investment, said he had been looking at Riggtulun (who observes, and even observes) for years before, and he eventually rushed after its launch.
“It all came together last year,” he said. “They saw acceleration in businesses because of AI.”
Timing is something that Ilan Peleg, CEO who co-founded the company with CTO Leonid Blouvshtein, knows. Peleg turned his attention to further education and ultimately built Rigg Turn, a champion middle distance runner, won four national championships in Israel and ranked in the top 16 of all mid-distance runners in Europe.
As Peleg sees, there are many companies building observability tools in today’s market (the most notable include DataDog and App Dynamics, among others).
However, no one has yet reached the “Holy Grail” of such work. Not only will you get a full picture of all the code shipped in production, but you can also understand how it interacts with what is already in use, and how to predict where the problem will arise. And do so by minimizing organizational disruptions and minimizing costs.
“The code is cheaper, but the bugs are expensive,” he said.
Meanwhile, the problem reached an “inflection point,” he said. “Developers can now ship more code than ever,” with all the automation used thanks to AI. “But it’s still a very manual process to fix it when things go wrong.”
Lightrun’s breakthrough was building an observability toolset that could monitor the code in the same way as an IDE and understand how it works along with the code in production. The code can then be adjusted automatically as it moves into production and continues to work without interruption or crashing. This does this by creating an AI-based simulation to understand its behavior and allowing you to fix the code before any issues arise.
“This is the unique part of us,” Pered said.
Given how it is close to other activities within the organization, there are many options for how Lightrun develops.
One of these is building more concrete tools for cybersecurity teams, given the obvious security implications that arise from bugs. The other is to build some of the tools that are even closer to the point of writing code, and build more to make possible bugs even more efficient.
For now, Peled said the plan is to focus on building its tools, talent and business, especially in IDEs. “Everything poses risk to resilience is mitigated,” he said, but he did not rule out more purpose-specific tools than in the future.
Regarding the code assistant, “these may be in our future,” he said. He said it would be difficult to predict what kind of code writing will look like in the future. Today, between 30% and 60% of all production problems are estimated from problems with code generated by both humans and machines, providing a way to observe and fix everything – whatever it is, regardless of how it was created, Lightrun is trying to fix it.