AI visionaries set out MWC25 gameplan
By Kavit Majithia
With AI set to dominate the keynote agenda and discussions at MWC25, big hitters from the ecosystem will convene in Barcelona to discuss the evolution of the technology and how it is beginning to impact the world as we know it.
AI-powered demos and trials will no doubt blanket the show floor from the get-go, but it will really be in the first keynote on day two, dubbed Tech Game Changers, when the established and emerging players in the sector enter the fray.
Setting the scene from a European point of view will be Arthur Mensch, an entrepreneur who has been hailed as one of the industry’s brightest visionaries.
Mensch is best known for co-founding and heading up France-based Mistral AI, which made waves when it outlined its vision to build and deploy AI systems much more efficiently than bigger competitors.
After reportedly achieving a valuation of more than €6 billion in 2024 and securing backing from the likes of Nvidia and IBM, Mistral AI has been touted as one of Europe’s big hopes to compete with rivals in the US and contenders emerging in China.
Joining Mensch in the keynote is a technology stalwart. Peggy Johnson, the former CEO of AR venture Magic Leap, was appointed as CEO of Agility Robotics in March 2024 and has been charged with leading the next phase of growth at the company.
It is pushing its vision of manufacturing robots that can empower companies to take control of their processes. Johnson comes to MWC25 to discuss how the company is developing its hardware with AI and automation software to solve industry- specific problems, as well as opening-up on Digit, the company’s “industry-leading mobile manipulation robot”.
Staying with an AI-packed Tuesday, Mensch and Johnson will be followed on stage by Peter Sarlin, the co-founder and CVP of AMD Silo AI.
Sarlin is one of the people behind what was formerly Silo AI, another company that can be considered a massive European success story.
It was one of the largest privately held AI startups in Europe, until AMD acquired it for more than $650 million last year, as part of the US chipmaker’s push to develop its AI play and take on the likes of Nvidia and Intel.
Based in Finland, AMD Silo AI specialises in developing AI solutions which help customers integrate the technology into their products, in addition to open source multilingual large language models (LLMs).
Ahead of their MWC appearances, Insight sat down with the trio of executives to discuss their plans for the show and beyond.
Q1
What are the key messages you plan to discuss at MWC25 and how important will AI prove as a theme at this year’s event?
We are very happy to participate in MWC25 and contribute to the global conversation on the future of mobile technology.
Since the 2024 edition we have seen remarkable progress, especially in making AI more accessible to users across all mobile devices. Additionally, network operators and chipmakers have demonstrated the immense value and innovative ideas that generative AI can bring. And this is just the beginning.
2025 will demonstrate the mobile industry is uniquely positioned to drive the widespread adoption of AI, making it accessible to a broader audience anywhere.
Q2
Can you provide an overview of Mistral AI’s core focus areas within AI development and the key challenges?
Our mission is to make genAI accessible to everyone and we currently focus on three things: keep releasing state-of-the-art models; helping the developer community to use our technology; and improving our AI assistant Le Chat
For example, we recently released our mobile app and we have announced Small 3, our most efficient and versatile model yet – available under Apache 2.0 licence.
We are excited to share more about our work at MWC and contribute to the transformation of the mobile industry.
Q3
How does Mistral AI envision the integration of AI into society and its potential transformative effects on global industries?
The integration of a multi-faceted technology such as genAI will likely be non-linear. However, we are confident that global industries will be the first to harness its significant value. We also think that open models are instrumental to ensure general integration.
In the mobile industry specifically, we already see genAI significantly transforming network infrastructure, customer service and many other things.
We support companies from the mobile industry to enhance maintenance teams, alert clients during network issues and improve 5G-enabled robots. We are also exploring the enhancement of the network itself with dedicated genAI models.
Q1
What are your key discussion points at MWC25?
Agility is testing reinforcement-learning and genAI models to enhance the capabilities of our humanoid robot, Digit, enabling it to acquire and hone skills over time. The integration of LLMs for natural language commands allows for a natural human-robot interaction.
Soon, instead of deploying engineers to write new and custom algorithms for Digit, we’ll have expert LLMs trained on data unique to each market we enter.
Q2
What specific industries or use cases are driving Agility Robotics’ efforts and how are these influencing your R&D priorities?
We are focused on our customer needs right now in the warehouse, logistics and manufacturing industries.
One of our primary goals is full collaborative safety – where Digit can safely work side by side with people on the production floor. No humanoid robot has the ability at this time and Agility will be the first to bring a collaboratively safe robot to market.
We’re also working on expanding how much weight the robot can carry and extending battery life so Digit can work up to 20 hours a day with only two hours of charging. In the future, we see Digit moving from industrial spaces such as manufacturing and logistics to light industrial spaces and eventually into the home.
Q3
What is Agility Robotics’ long-term vision for how humanoid robots will coexist with humans in society?
Digit takes on the tasks that are difficult for businesses to recruit humans for – highly repetitive, physically demanding and injury prone tasks.
Manufacturing and logistics, and other similar industries are initial target markets for Digit due to their familiarity with conventional automation and severe labour shortages.
However, Digit’s design allows for expansion into additional markets like healthcare and elder care with continued evolution of its design and safety capabilities.
The question of when Digit can coexist with humans in everyday settings depends a lot on regulatory and societal norms. It may take some time to figure out how to safeguard and accept a humanoid robot as a consumer product.
Q1
What are the key talking points at MWC25 and how important will AI prove as a theme at this year’s event?
AI has become a cornerstone technology that’s elevating digital products and services across industries.
At MWC25, we’re seeing this manifest particularly in the mobile sector, where AI integration is driving innovation in everything from network optimisation to user experiences. As AI becomes more pervasive, we’re witnessing innovation in making it more efficient throughout the value chain – from model architectures to training techniques.
Companies like DeepSeek have shown how innovations can dramatically improve efficiency. There’s still massive potential for innovation in areas like specialised hardware, efficient training methods and optimised inference – all crucial for the mobile sector’s AI future.
Q2
Given Silo AI’s strong presence in Europe, how will the acquisition by AMD influence the European AI landscape?
Silo AI and AMD have an aligned mission to create a global leader that will shape the future of computing, while preserving Silo AI’s essence as a leading AI lab with European roots. By joining forces with AMD as a compute platform leader, we can now work with more advanced and larger AI initiatives.
We maintain our headquarters in Finland and continue building state-of- the-art AI for customers, now obviously on top of AMD compute platforms. We continue training performant multilingual open source models that particularly cover European languages.
This combination strengthens Europe’s position in the global AI landscape. Europe has an opportunity to lead in open source AI development.
Q3
As AI becomes more pervasive, how are Silo AI and AMD addressing concerns related to ethical AI practices, considering Europe’s stringent data protection regulations?
Our approach to ethical AI and data privacy is deeply rooted in European values and regulatory frameworks.
We focus on developing AI that is both powerful and responsible, ensuring compliance with Europe’s data protection standards and general best practices. This manifests in our work on open source models and transparent development practices.
By maintaining our European base and values while accessing global capabilities through AMD, we’re well-positioned to continue developing AI solutions that prioritise both performance and responsibility.