If you haven’t already, subscribe and join our community in receiving weekly AI insights, updates and interviews with industry experts straight to your feed.

DeepDive 

Your weekly immersion in AI 

A founder steps onto a stage and says “We’re building an electricity startup.” 

It would throw you a little, wouldn’t it? Electricity is infrastructure, not a startup category. It powers everything else. 

Right now, AI startups are everywhere. But in just a few years time, the idea of calling yourself an AI startup could be obsolete. 

Increasingly, AI isn’t a vertical. It’s a layer – woven into logistics platforms, fintech apps, biotech labs, defence systems and creative software. It’s becoming the operating system of the modern economy.

And when infrastructure becomes ubiquitous, you can’t use it as your startup differentiator. 

AI stops being the story

In the past 24 months, ‘AI startup’ has been shorthand for ambition, funding and technical edge. But diffusion is accelerating fast.

According to the Stanford AI Index, enterprise adoption of AI continues to rise sharply year on year, with generative AI deployment spreading across sectors rather than remaining concentrated in specialist firms. 

We’ve seen this pattern before. You’re unlikely to hear someone pitch a cloud startup anymore, or a mobile startup – and definitely not a web startup. Because those technologies diffused too. 

Economists have long described AI as a general-purpose technology (similar to electricity or the steam engine) meaning its impact compounds as it becomes embedded across industries.

So that same transition appears to be underway. In the near future, it will be assumed that your startup uses AI. 

Where will differentiation move? 

If AI becomes infrastructure, founders will need to compete on different terrain. We’re already seeing these four shifts: 

1. Proprietary data
   Models are powerful, but data is leverage. Startups with access to unique, high-quality, domain-specific datasets will hold defensible positions. In healthcare, that might mean longitudinal clinical data; in climate tech, granular sensor networks. The value moves from generic intelligence to contextual intelligence.
   
2. Distribution and trust
   When we interviewed Abdalla Kablan (AI Investor and Entrepreneur and Chief Strategist at AIDEN) for the newsletter, he argued that an AI idea’s worth “hinges on its scalability, applicability, and ability to address a genuine pain point” – not just technical sophistication. He stressed the importance of validation and practical utility over buzzwords, and that insight becomes even more important in an infrastructure era. When everyone has access to similar models, distribution becomes oxygen, and trust becomes a growth strategy. In another DeepDive, Jordi van den Bussche (Founder and CEO at JVDB Studios) highlighted the evolving collaboration between AI developers and trusted communicators – arguing that demystifying AI for broader audiences is essential to mainstream adoption. In other words, infrastructure still needs human interfaces.
   
3. Regulation and governance
   As AI embeds itself into critical systems, regulatory scrutiny intensifies. The EU AI Act has now entered its implementation phase, marking a shift from principle to enforceable compliance obligations. Globally, governments are moving from strategy papers to standards. In this environment, governance literacy becomes a competitive advantage. The startups that thrive will demonstrate auditability, explainability and compliance readiness from day one.
   
4. Hardware and compute efficiency
   The International Energy Agency has warned that data centre electricity demand could more than double by 2026 (compared with 2022), driven in part by AI workloads. Compute is now an economic and environmental infrastructure. Startups that can deliver high performance with lower compute footprints will be attractive partners in a cost and carbon-conscious world. Infrastructure thinking demands systems thinking.

From AI-native to AI-ambient 

There’s also a cultural shift underway.

We recently wrote about research from MIT, Stanford and Salesforce AI which found that workers are selective about what they delegate to AI systems. Even when AI is technically capable, people prefer to retain control over high-stakes, creative or ethical tasks.

That tells us that as AI becomes ambient (embedded into workflows and systems), human judgement can become more valuable. So the startups that succeed will be the ones that design for collaboration and human agency. 

AI itself will fade into the background; supporting, and increasingly taken for granted. 

The question for founders in 2026

If you’re building today, here’s the energising question:

If every competitor had access to the same frontier models tomorrow, would your business still work?

If the honest answer is no, you need to double down on: 

• Exclusive data partnerships
• Deep vertical expertise
• Regulatory fluency
• Community and brand trust
• Operational excellence

AI will still transform industries. But we’ll all stop thinking about it so much – and focus instead on the businesses and societies built on top of it. 

We want to know what you think 

If AI is becoming infrastructure, where should founders focus their edge?

Open this newsletter on LinkedIn and tell us what you think.