How Startups Can Thrive Developing Sovereign AI

In today’s fast-evolving digital landscape, the concept of “sovereign AI” is gaining significant momentum. Especially in a changing political environment. As Europe looks to reduce its reliance on U.S.-based technology, the notion of sovereignty in data processing and artificial intelligence has become a key focus point. European tech firms are increasingly investing in local AI infrastructures with growing concerns over data privacy, security, and national competitiveness.

How can startups meet this demand?

Alexander Grau, Startup Program Manager at OVHcloud, a leading European cloud provider, is at the forefront of these discussions. In this interview, we explore his insights into the rising trend of AI sovereignty, its implications for the tech industry, and how OVHcloud is playing a pivotal role in supporting European startups in this transformative shift.

Alexander, in your opinion, what role does data sovereignty play in the broader conversation about AI sovereignty?

Alexander: Data sovereignty refers to the idea that people’s data should be stored on infrastructure within the country or continent they have their headquarters in.

At its core, data sovereignty is about ensuring that data is stored and processed within a particular jurisdiction, under the local laws and regulations. This is crucial because the quality, security, and ethical use of data directly influence the development and performance of AI systems.

In the context of AI, data sovereignty becomes particularly significant as it ensures that the data used to train AI models is subject to regional standards, such as the European Union’s GDPR. This gives individuals control over their data and protects them from potential misuse by companies that might operate outside of the region’s legal boundaries.

Could you explain what “sovereign AI” means and why it has become such a hot topic in Europe?

Alexander: “Sovereign AI” refers to the concept of building and deploying artificial intelligence systems that are based on locally stored data and developed within the legal and regulatory frameworks of a specific country or region.

This concept has gained significant attention in Europe for several reasons. First, Europe is very focused on protecting its citizens’ data privacy, and AI systems that process personal data must adhere to strict regulations like the GDPR. With many of the world’s leading AI technologies relying on data centers and models hosted in the U.S., there is a growing concern that Europe’s dependence on external infrastructure could compromise both data security and sovereignty.

Moreover, there is an increasing desire for Europe to foster its own technological independence and ensure that AI systems developed within the region are grounded in local languages and cultural nuances. This need for “localization” has become a matter of economic competitiveness as well as national security.

Can you tell us specifically from your experience how start-ups are implementing AI sovereignty?

Alexander: With pleasure. I’m seeing more and more startups attaching importance to IaaS, Saas and PaaS solutions that are covered by European data protection regulations. The startup moinAI, for example, offers a self-learning AI chatbot solution for digital customer communication for companies. moinAI’s customers also include insurance companies, where customers enter very sensitive data into the chat window that requires special protection. Startups fulfil their responsibility to meet the highest data protection requirements in their business model because they take a holistic view of their customers’ customers. Young companies in particular which are new to the market and deal with sensitive data need a certain amount of trust. Customers must quickly understand that even smaller startups have European data sovereignty fully on their radar and that no compromises are made here.

Thanks for this best practice example. In a nutshell, which exact benefits do sovereign AI models have for companies and startups?

Alexander: By having more sovereign AI models, Europe can ensure its tech ecosystem is more resilient, reduce reliance on foreign technology, and better serve its diverse populations with tailored solutions. This shift is crucial as AI becomes an integral part of industries like healthcare, finance, and government, and it’s driving the demand for more localized cloud infrastructures across the continent.

This trend is especially interesting for startups, as they can develop business models that are focused and tailored on those needs. A well-known example would be to create a chatbot with a large language model that is trained on smaller European languages and hosted according to GDPR in Europe at the same time. But as we all know – AI goes beyond Language Models and Chatbots and there are huge other potentials.

I’m thrilled to see more and more startups that meet this demand thriving!

As AI models, like large language models (LLMs), require massive amounts of data, how can European companies ensure that these models are both locally trained and aligned with local languages and cultures?

Alexander: European companies can ensure that large language models (LLMs) are locally trained and aligned with local languages and cultures by following several best practices:

• Multilingual models: Developing multilingual models that can handle multiple languages and cultures within a single model can be beneficial.
• Collaboration: European companies can collaborate with regional AI research institutions, universities, and local governments to share resources, data, and expertise. This can help in building more robust and locally adapted AI models that cater to the unique linguistic and cultural needs of the region. A good example is our OVHcloud Startup Program. We open our broad network for startups to collaborate with tech experts, use consulting and synergies.
• Transparency and explainability: Companies should be transparent about their AI models’ development, training data, and performance metrics. Providing explanations and justifications for model decisions helps build trust and ensures that the models align with local values, norms, and ethical standards.
• Compliance with regulations: European companies must comply with regional data protection regulations, such as the General Data Protection Regulation (GDPR), when developing and deploying AI models.

How the OVHcloud Startup program supports startups:

You are currently viewing a placeholder content from YouTube. To access the actual content, click the button below. Please note that doing so will share data with third-party providers.

More Information

Unblock content Accept required service and unblock content

What are some of the challenges that startups face when it comes to adopting sovereign AI?

Alexander: Developing AI technology, the biggest challenge will be the cost and resource requirements for startups. Startups often struggle with limited budgets and human resources, making it challenging to invest in the necessary infrastructure, technology, and talent to build sovereign AI solutions.

Another challenge is the access to diverse and high-quality data: Sovereign AI requires access to large and diverse datasets that represent the local languages, cultures, and contexts. However, data scarcity, poor data quality, or lack of representation in existing datasets can pose challenges for startups in developing accurate and effective AI models.

There are still some technological barriers: Developing sovereign AI solutions may require the use of advanced AI technologies, such as large language models or computer vision algorithms, that are proprietary or controlled by a small number of global technology companies

Looking ahead, how do you see OVHcloud’s role in helping European businesses and startups become more competitive in the global AI race while maintaining data sovereignty?

Alexander: OVHcloud can provide European businesses and startups with a robust and secure cloud infrastructure to support their AI development needs. By offering scalable computing resources, storage, and network services, OVHcloud can help businesses and startups build, test, and deploy AI models efficiently while ensuring data sovereignty.

OVHcloud is committed to ensuring that startups have 100% control over their applications and infrastructure, there is no vendor lock-in effect and anyone who joins us can leave at any time without getting into financial difficulties. For us, this great flexibility and independence is an essential part of data sovereignty.

We warmly welcome different types of startups and scaleups in our OVHcloud Startup Program. It provides startups and scaleups with free cloud credits and technical support to accelerate your journey onto our secured, reliable, and cost-efficient cloud.

Learn more about the OVHcloud Startup Program on our website. Feel free to reach out – I’m always keen to talk to innovators and inspiring people!