In a world where the geopolitical climate remains unstable, trust is an increasingly valuable commodity – and this rings true for the AI landscape too. More businesses are now buying from local AI firms with well-established and sovereign AI supply chains that have full visibility as standard in the offer package.
AI sovereignty is key: this refers to the ability for a country or company to independently control its AI technology, data or regulations.
“If a company in the public sector or a government uses an AI product, they need to be sure that the data used or the training data are sovereign – meaning they belong to them and adhere to the appropriate legislation when they’re operating,” explains Filippo Sanesi, global marketing and ops leader – startup program, at OVHcloud.
“When you go to the supermarket, you look at a product and check the ingredients so you know what’s in it and where it’s from – it's the same with AI today. More and more, there's scrutiny around where those models are coming from and which data they’re using, and this is where sovereignty comes into play.”
Sanesi says one key advantage of AI sovereignty could be the creation of better quality models. This is because it creates a better overview of the supply chain of those models and therefore has less potential risks down the line.
He believes as the industry matures, there will be more focus on these quality models too, meaning better quality of data and better understanding of the ethics implied in that model, which is where the AI sovereignty comes in.
“The AI sovereignty theme is not just about startups – it's about governments and how they can plan for the next generation to be trained. It’s also about how you can have a framework of data protection and data legislation that can guarantee the data stays in the country,” adds Sanesi.
Building trust
Axel Hellström is CMO and cofounder at Ebbot, which offers AI-powered customer service tools like chatbots. He says he sees being a generative AI solution that can offer a fully EU sovereign cloud setup for its clients as a competitive advantage, and has clients for whom this is a priority.
“We are a platform that offers various different setups, but our AI sovereign setup is really a unique selling point for us,” Hellström tells Maddyness UK.
“When it comes to AI, we work with open source models and tune them ourselves in-house to make sure that they perform the best they can. We run and host them with OVHcloud inside the EU, which is a requirement for a lot of our clients to fulfill not only GDPR, but also potentially the requirements of the AI Act.”
Ebbot’s clients expose AI towards their customers and end users via a conversational AI service in the form of chats and email. Hellström notes that within that, the end users tend to share a lot of personal data – where that data is being stored and processed is key for his clients, because there are risks with sending it abroad.
He believes it’s a big trust factor to be transparent on how data is being processed, where it's being processed and by whom.
“I believe working locally will enhance the capabilities of your local market,” says Hellström. “It will not only be a requirement to stay compliant, but it will also help to keep your local AI expertise developed.
“Between companies and providers, we will see a flow of knowledge between companies. Europe is in a global competitive market and I believe that that also makes the EU sovereign cloud important from a long term perspective.”
Sanesi agrees, and notes how companies that will embrace AI sovereignty will be seen by prospective clients as more serious contenders who have less risk, better quality and control of their supply chain, and be more capable of doing the work well.
“Is that product made under EU legislation respecting all the rules? As a startup or company, are you owning that product or did you use multiple sources from multiple legislation? These answers create a lot of questions on the ethics part of AI, like who trained your data and where you got that data from,” he adds.
Adopting AI sovereignty from the start
For up-and-coming startups, AI sovereignty is something that should be prioritised from the get-go – and this is the approach adopted by OVHcloud from its launch.
“Since the beginning, we've stood for innovation, freedom of moving your tech in and out, and security and privacy – that's been non-negotiable since day one,” explains Sanesi. “We’re a European company with presence across Europe and the UK, allowing organisations to choose whichever region is best for their needs and their data. Our value proposition is around the fact that OVHcloud can provide sovereignty, privacy and security, and we're not going to change that in the long term.”
Hellström says it's always harder to add things into a startup’s foundations at a later stage – you sign agreements with clients detailing how you're doing things, and then you need to update those agreements when big changes are made. As a startup, he says, you don't really have the structures in place to do that very efficiently – so it’s best to make those fundamental decisions as early as possible.
“Having AI sovereignty as a consideration from the start will make things a lot easier for startup leaders,” he explains. “Use that, as we do, as a unique selling point and leverage it directly – even if it costs more, it makes you a bit more premium that you can offer it and it's a competitive advantage.
“Even if you're a company outside of Europe, you can offer servers in Europe, but you can’t get around the fact that you're not a European company. Use that advantage – I believe it's a good strategy.”
If you're building a successful, sustainable, innovative startup or scale-up, you can benefit from up to €100K in cloud credits with OVHcloud. Learn more about the program here.
