Talking confidently about AI is not evidence of being able to lead with it.

For founders, the focus should be on what a candidate has tried, learnt and understood about applying AI within their function. For investors, it is whether portfolio leaders understand the technology well enough to help their teams use it effectively.

At the EUVC Summit & Awards Show in April, Chris Preston, CEO at ZEREN, and Rishabh Kaul, Venture Partner at Hoxton Ventures, joined us to discuss how to assess that capability.

Chris described a client that asks candidates to prepare a paper on AI in their function, then debate it with stakeholders for two hours. The exercise tests how they think and whether they bring original ideas.

Rishabh looks for hands-on experimentation: even an imperfect tool can reveal what AI can and cannot do. But some roles still require deep experience, including regulatory knowledge in fintech. Curiosity should reinforce expertise, not replace it.

That gives the next interview or leadership review three telling questions: What have you tried? What did you learn about its limitations? Which parts of this role still depend on specialist judgement?

When the role itself remains unclear, Chris suggests using interim expertise to help define it before committing to a permanent hire.

Start with the conversation below, then take the accompanying reading into your next hiring or portfolio review. Beyond this week’s focus, a few other developments have also made the cut.

With 💖
David & Andreas

Assess recent learning, distinctive strengths and essential experience against the work ahead. These pieces help you weigh early-stage potential against functional depth and recognise when growth calls for different capabilities.

Look for changes in how portfolio leaders work, then scrutinise how your own firm evaluates talent. ZEREN’s approach to leadership capability reviews and TDK Ventures’ independent interview assessments provide practical starting points.

AI compute is getting cheaper faster than it is becoming more energy-efficient. New research from Redstone models the global compute buildout through 2050 and estimates that, in its model, the same dollar could buy 143× more compute, while the same watt powers 59× more.

Two figures put the scale in context:

Alongside the research, Redstone has built a startup simulator that lets founders test how improvements in efficiency or cost, or added capacity, combined with industry adoption, could change global AI compute by 2050 under its Base case.

Callosum has raised a $100 million seed round, led by Atomico with participation from AlbionVC and several other investors, to scale its software layer for heterogeneous intelligence, which co-optimises AI models and chips.

The Cambridge-founded company orchestrates workloads across the hardware best suited to run them, aiming to improve the cost, speed and capability of AI inference.

Its tailored-inference platform is already in production, while the funding will also support deeper partnerships across the silicon ecosystem and team expansion.

Europe continues to produce world-class founders and technical talent, but keeping more of that value in the region will depend on capital, policy and technological sovereignty.

In this episode of This Week in European Tech, Dan Bowyer and Priyanka Savjani of SuperSeed, alongside Andrew J Scott of 7percent Ventures, discuss EU Inc., AI sovereignty, Europe’s funding gap and what it will take to build more global technology leaders from here.

The strongest responsible products do not ask customers to compromise. In this talk, Rokas Peciulaitis, Founder and Managing Partner at Contrarian Ventures, uses Vinted, Patagonia and Mako to show how durability, reuse and smarter resource use can make a product more compelling.