Nvidia, AI and the Dot-Com Echo

These days often remind me of dot-com boom.

Last week I read a statistic that simply blew my mind - Nvidia's market cap represents 3.8% of global GDP and 14.7% of US GDP. While market capitalization and economic output aren't directly comparable, the sheer scale raises an important question – what does this valuation actually reflect?

Yes, the current AI infrastructure demands seem real and immediate, unlike some dot-com era promises. Companies like Nvidia are generating substantial profits today. Their latest results are proof of that. But even accepting "this time is different" argument, are these valuations justified? I don’t think so.

Here is another warning sign – many highly valued AI startups depend on cheap API access to frontier models. These model companies in turn have raised money at lofty valuations and using that funding to subsidise pricing while making huge infrastructure investments. What happens when VC funding tap closes, and they are forced to raise prices? Won’t it destroy the P&L of downstream startups? Will consumers pay 3-4x more for chatbot subscriptions?

At the same time, I can’t imagine a scenario where we go back to a world without these tools. Like the early days of the internet, many business models look flaky at best, but the underlying business value of AI seems strong enough to survive such setbacks in the long term.

So even if the prevailing narrative, "AI impact is like nothing we've seen before" is true, big change doesn't happen overnight, despite what catchy headlines suggest. Changing workflows/processes takes finite time. The recent MIT study shows 95% of Enterprise GenAI pilots delivering no business return due to several reasons, one being broken workflows.

Models are tools. Tools are a means to an end, not an end in itself. Better models don't automatically mean higher adoption.

History has taught us that whenever new technology or idea captures our collective imagination, one thing remains constant - human behaviour patterns repeat. Greed and fear make markets move between euphoria and reality, often with painful corrections in between.

As they say, "History doesn't repeat itself, but it often rhymes." Nobody can predict the future, but we can bet on one thing - human nature staying unchanged.