AI-based coding explodes in a promise that it will make developers work faster and easier. However, AI encoding also leads to other things: a substantial increase in lines of code, thus the possibility of crashes or other unfortunate errors.

On Monday, an Israeli startup called Lightrun (Lightrun) built an observable platform to identify and debug (remedial) code and announced a $70 million Series B series B before these issues arose. Funding not only highlights the market gaps in similar tools, but also the traction of Lightrun in meeting demand.

New Backer Accel will lead with former investors Insight Partners, which is the involvement of Citi, Glilot Capital, GTM Capital and Sorenson Capital. So far, Lightrun has now raised $110 million, including the Insight-led Series A Series A that we cover in 2021.

The startup didn't disclose its valuation, but there are some strong signs that it's doing well.

First, there are its customers. Citi is a strategic supporter and one of the impressively well-known clients, which also includes ADP, AT&T, ICE/ICE/NYSE, Inditex, Microsoft, Priceline, Priceline, Salesforce and SAP.

Secondly, there is the time for products and companies to adapt to the current market structure. Back in July 2024, Lightrun announced a new AI-based debugging tool for use in an organization's integrated developer environment (IDES), appropriately known as runtime autonomous driving AI DEBUGGER. Although the company's platform has achieved impressive results, this is a product of the current dilemma faced by many businesses: AI leads to more coding and more problems, and Lightrun builds AI tools to solve this problem.

The company said revenue has more than fourfold since revenue started, which is why investors knocked on the door. Andrei Brasoveanu, an acceleration partner who led the company's investment, said he had been following Lightrun (even observing) in the years before that, and after the launch, he finally got stuck.

“It all came together last year,” he said. “They saw the acceleration of the enterprise, all due to AI.”

The timing is something Ilan Peleg, CEO of co-founding the company with CTO Leonid Blouvshtein, knows. Before shifting attention to further education and eventually establishing Lightrun, Peleg was a champion midrange runner, winning four national titles in Israel and ranking in the top 16 of all midrange runners in Europe.

As Bellig has seen, dozens of companies in the market today have built observability tools. Some of the most prominent ones include DataDog and App Dynamics.

But no one has been able to reach the “holy grail” of such work: not only has the ability to fully understand all the code being shipped in production, but also how it will interact with what is already in use, and how to expect where the problem arises - and deal with it with minimal interruptions and minimal interruptions and minimal costs to the organization.

"Code is getting cheaper, but errors are expensive," he said.

At the same time, he said, the problem has reached a "turning point". Thanks to AI, developers can now ship more than ever because of all automation. "But when the problem arises, it's still a very manual process. ”

Lightrun's breakthrough is to build an observability tool set that monitors code like an IDE in an IDE and understands its behavior with actively produced code. Lightrun is then able to automatically adjust the code as the code goes into production to continue running without interruption and crashing. It does this by being able to create AI-based simulations to understand the behavior and then fix the code before the problem arises.

"This is our unique part," Belig said.

Given how Lightrun develops, there are many options, given the observable proximity to other activities in the organization.

One of them is to build tools more specifically for cybersecurity teams, considering the obvious security impact. Another person may be closer to the code creation point to make discovery and fix possible errors more efficient.

Currently, the program is focused on tools, talents and businesses specifically for IDEs, Belig said. "Everything brings the risk of resilience, and we are mitigating it," he said.

As for the code assistant, “those may be our future,” he said, “but even after execution, it’s complex and extensive to focus on software fixing issues.” He said it’s hard to predict what future code creation will look like. Today, estimating 30 to 60% of all production problems comes from human and machine-generated code problems, providing a way to observe and fix everything - no matter how it was created - Lightrun is working hard to solve.