This shouldn’t surprise anyone who studies technological evolution. Like electricity before it, AI is following a predictable path toward commodification. The story playing out now – from OpenAI’s complaints about DeepSeek using their material to train models, to the frantic attempts to create regulatory moats – reveals an uncomfortable truth: the billions being poured into frontier AI models may never generate the returns investors expect.

The reality is that AI, in its raw form, has nothing inherently defensible. It’s becoming the equivalent of electricity – ubiquitous and essential, but a terrible business on its own. When DeepSeek demonstrated that frontier models could be replicated more efficiently and at lower cost, they exposed the fundamental flaw in Silicon Valley’s AI investment thesis. The path from being first to being profitable isn’t always a straight one.

The pivot to national security arguments by some AI companies suggests how precarious they see their position. Because when competing on merit becomes difficult, regulation becomes your last resort for creating artificial moats. 

This dynamic is particularly visible in the diverging approaches to AI development. While some companies chase the chimera of artificial general intelligence (AGI) and try to insert AI into every human interaction, others appear more focused on efficiency and practical applications. The difference isn’t just technical—it reflects fundamentally different views about AI’s role in society.

So where does the real value lie? The answer isn’t necessarily in building bigger, more expensive models – it’s in vertical integration and specific applications. Companies succeeding in AI today aren’t trying to be everything to everyone. Instead, they’re treating AI as infrastructure and building specific solutions for well-defined problems. They understand that domain expertise and proprietary data create more defensible positions than general-purpose AI ever could.

That means businesses trying to navigate this landscape need to treat AI as the utility it’s becoming. Focus on creating value through vertical integration and specific applications. Understand that the real opportunities lie not in the AI models themselves but in how they’re applied to solve concrete problems. The companies that will thrive aren’t those building the most sophisticated AI – they’re the ones using AI to enhance their core competencies while maintaining the human connections that create genuine value.

The current AI bubble isn’t just about inflated valuations – it represents a fundamental misunderstanding of where value is created in technological evolution. As we’ve seen with every previous technological revolution, the most hyped capabilities rarely translate into sustainable businesses. The real winners are those who understand that technology is a means to an end, not an end in itself.

For investors, entrepreneurs, and business leaders, the message is clear: the future of AI isn’t in building bigger, more expensive models or in creating artificial regulatory moats. It’s in understanding how AI can be applied to create real value in specific contexts while preserving the human connections that drive genuine innovation and growth.

The sooner we recognize AI for what it is – a powerful but ultimately one-day commoditized utility – the sooner we can focus on building businesses that create lasting value rather than chasing speculative returns. The bubble may be bursting, but for those who understand these dynamics, the opportunities have never been clearer.