A new AI orchestration startup from the founders of Lithuanian unicorn Nord Security has set out to help companies put AI projects into production, initially focusing on providing greater visibility, security and adaptability to large-scale language models (LLMs). is leaving.
The startup, called Nexos.ai, is the brainchild of Tomas Okmanas (pictured above) and Eimantas Sabaliauskas, who have built one of the most popular brands not only in Lithuania but across Europe. Nord Security, best known for its flagship VPN product NordVPN, bootstrapped for its first decade before succumbing to a massive $100 million investment in 2022 at a $1.6 billion valuation (it later raised its valuation to $3 billion during a follow-up round of funding). achieved).
Their new company is coming out of secrecy today, having received $8 million in funding from a number of high-profile backers, including major investor Index Ventures, which has now invested for the first time in Lithuania.
“We have known Thomas and the work he has been doing for many years, so as soon as we heard that he was building a new company in the field of AI, we were finally willing to accept venture capital funding at this (early) stage. , we were very eager,” Hannah Seal, partner at Index Ventures, told TechCrunch.
Other notable investors include Creandum and Dig Ventures, and prominent angels including CEOs of Datadog, Klarna, Supercell, and Wix also participated.
Catalyst utilization
Today, teams looking to put AI into production must connect numerous tools that involve recruiting and building a team with the necessary skills. This is where Nexos.ai wants to get involved.
“We saw that there was a huge gap between running AI in pilot and going into production,” Okmanas told TechCrunch in an interview. “It can work and be useful when testing AI in the lab, but how do you ensure high availability, especially when enterprises want to put it into production? How do you ensure security? How do you manage your costs?”
Nord Security has been around for more than a decade, but five years ago it transitioned into an umbrella company called Tesonet, an incubator with a portfolio of more than 20 companies. One of them is Hostinger, a web hosting company. Hostinger recently added AI-powered smart features to its website building tool. Okmanas, a Hostinger board member and shareholder, said some of the challenges they faced were the catalyst for what would eventually become Nexos.ai.
“We wanted to use AI in our website builder, so we turned on OpenAI, started testing, and put it into production,” said Okmanas. “In August, I received a bill for $150,000. For what? Why was it so expensive? “There was absolutely no visibility.”
And when OpenAI went down several times, Okmanas became convinced that something had to be done to make it easier to deploy, manage and optimize the “increasingly complex ecosystem of AI models” that organizations might need.
Customers can access more than 200 AI models, from big names like OpenAI and Anthropic to small, niche LLMs, through a simple application programming interface (API). The idea is that if OpenAI goes down, companies can temporarily (and automatically) switch to another provider without having to make much progress. Or, if the cost of accessing a particular LLM is exploding for any reason, companies may switch to a different LLM to keep costs down.
Nexos.ai also introduces “intelligent caching” into the mix. If multiple users are repeating a particular question, the system can switch to its own database rather than continuing to use LLM, which can be costly.
In terms of security and compliance, Nexos.ai prevents individuals from sending personal data to LLM providers, and if an employee leaves the company, that access can be terminated immediately.
But there’s no avoiding the elephant in the room. One of the reasons companies have been hesitant to embrace AI is the thorny issue of data security. Healthcare companies, banks or insurance companies cannot entrust all of their sensitive information to their LLM provider. . It’s worth noting that Hostinger itself suffered a data breach in 2019, and NordVPN has also been hacked in the past. This is the type of attack that every company faces today.
This raises the question of how Nexos.ai handles such data, considering that it hosts everything on its own infrastructure. Okmanas said the company will likely offer self-hosting in the future and is already supporting integration with the company’s internal LLM.
We also have guardrails in place to detect when data, such as personally identifiable information (PII), is being transferred. In these cases, the data can be routed back to the originating company’s own LLM or database. However, if the query is general (e.g. a customer asks an AI agent for details about location and business hours), the query will be processed on the Nexos.ai side.
From idea to start
It took Nexos.ai about six weeks to go from idea to formal integration, and while the speed of securing funding largely depended on the founders’ pedigree, a big part of it was simply timing.
“It seems like we’re finally getting beyond the hype around AI and the real applications are now emerging,” Seal added. “Every large company is realizing that this really makes sense and they need to adopt AI at scale. “Now is the time for infrastructure to catch up with the model.”
However, the speed of execution is largely due to Tesonet’s extensive organizational set-up, with approximately 4,000 employees across the entire portfolio. This allowed Okmanas to quickly build a team of about 30 people he knew and trusted who could work full-time on Nexos.ai.
“We have a team that can really join forces. They’ve been working together for years, so we don’t have to tell them what’s what,” Okmanas said. “We plan to hire externally as well, but it will take much more time.”
Nexos.ai’s platform is scheduled to launch in late March, but Okmanas said it is already working with a number of “beta customers and design partners.”
