techcrunch.com
A new AI orchestration startup from the founders of Lithuanian unicorn Nord Security is setting out to help enterprises put their AI projects into production, with an initial focus on bringing greater visibility, security and adaptability to large language models (LLMs).Nexos.ai, as the startup is called, is the handiwork of Tomas Okmanas (pictured above) and Eimantas Sabaliauskas, who built one of the most recognizable brands not only in Lithuania, but in all of Europe. Nord Security, best known for its flagship VPN product NordVPN, bootstrapped its way through its first 10 years before succumbing to a bumper $100 million investment in 2022 at a $1.6 billion valuation. Their new company is exiting stealth today with $8 million in funding from a slew of high-profile backers, including lead investor Index Ventures, which has now made its first ever investment into Lithuania.Weve known of Tomas and the work that hes done for many years, so as soon as we heard that he was building a new company in the AI space, and was finally willing to take venture capital money at this [early] stage, we were very eager, Index Ventures partner Hannah Seal told TechCrunch. Other notable investors include Creandum and Dig Ventures, and prominent angels such as the CEOs of Datadog, Klarna, Supercell, and Wix also participated.Capitalizing on a catalystCurrently, teams that want to put their AI into production have to connect myriad tools, which likely involves recruiting and building teams with the necessary skills. This is where Nexos.ai wants to step in.Ive seen that theres a big gap between running AI as pilots and going into production, Okmanas told TechCrunch in an interview. When youre testing AI in your lab, it might work and it can be useful, but when you want to put it into production, especially in enterprises, how do you ensure high availability? How do you ensure security? How do you manage cost?Nord Securitys been around for more than a decade, but five years ago, it was folded into an umbrella company called Tesonet, an incubator with a portfolio of more than two-dozen businesses. One of these is web-hosting firm Hostinger, which recently added AI-enabled smarts to its website building tool. Okmanas, a Hostinger board member and shareholder, said some of the issues they encountered served as a catalyst for what would eventually become Nexos.ai.We wanted to employ AI in our website builder, so we turned on OpenAI, we started testing it, and we put it in production, Okmanas said. In August, we had $150,000 billed. For what? Why was it so expensive? There was no visibility.AI website builder on HostingerImage Credits:HostingerAnd when OpenAI went down a handful of times, Okmanas was 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 may need.Through a simple API (application programming interface), customers can access more than 200 AI models, from big-name incumbents like OpenAI and Anthropic to smaller, niche LLMs. The idea is, if OpenAI goes down, a company can temporarily (and automatically) switch to a different provider without breaking stride. Or if the costs involved in accessing a specific LLM explode for whatever reason, a company can transition to another one to keep their costs down.Nexos.ai also ushers intelligent caching into the mix if a particular question is repeated by multiple users, the system can turn to its own database rather than continuing to engage the LLM, which can get expensive.On the security and compliance fronts, Nexos.ai also prevents individuals from sending private data to LLM providers, or if an employee leaves a company, their access can be terminated immediately.Nexos.aiImage Credits:Nexos.aiTheres no escaping the elephant in the room, though: One of the reasons enterprises have been hesitant to embrace AI is the thorny issue of data security healthcare companies, banks, or insurance firms cant simply trust LLM providers with all their sensitive information. Its worth noting that Hostinger itself was hit with a data breach in 2019 and NordVPN has also been hacked in the past the sort of attacks that all companies face today.This raises questions around how Nexos.ai handles such data, given that it is hosting everything on its own infrastructure. Okmanas said the company will likely offer self-hosting in the future, and that it already supports integrations with companies own internal LLMs. It also has guardrails in place to detect when data, such as personally identifiable information (PII), is sent to it in such cases, it can re-route the data back to the originating companys own LLMs or database. But if a query is generic, like a customer asking an AI agent for details about their location and opening hours, then the query will be handled on the Nexos.ai side.From idea to inceptionGoing from an idea to formal incorporation took Nexos.ai around six weeks, and while the speed of securing the funding was largely down to the founders pedigree, a big part of it was simply the timing.I feel like weve finally gone beyond the hype of AI, and now the real-world applications are coming, Seal added. All the large enterprises are realizing this is really meaningful, and they need to adopt AI at scale. And now is the time for the infrastructure to catch up with the models.The speed of execution, though, was substantively due to the broader organizational setup at Tesonet, which has around 4,000 employees across its portfolio. This enabled Okmanas to quickly assemble a team of around 30 people who he knew and trusted to work on Nexos.ai full-time.We have these teams that can really join forces theyve been working together for so many years, theres no need to tell them whats what, Okmanas said. Well also be hiring from the outside, but that takes much more time.Nexos.ais platform is set to launch by the end of March, though Okmanas said it is already working with a bunch of beta customers and design partners.
