The AI Race
Why it’s not 2000 again

It’s easy to draw the connection between the early days of the internet and the emergence of AI. The ‘growth before profit’ mentality is here again as well as market volatility and huge amounts of hot money.

In 100, finance professors at Purdue University published a study analysing the 95 stocks that added an internet suffix to their name during the dotcom bubble. On average, the stocks rose by 74%. During the bubble the director of the SEC’s investor-education office told the WSJ ‘Don’t invest by just a name. That’s asking for losses.’ Similarly, AI firms have secured enormous valuations, the suffix ‘AI’ is being added to the names of only tangentially related business and the term itself is being wilfully misinterpreted to appeal to investors.

When a new technology emerges, investors want to pick winners. With the emergence of the internet in the 90s you would have to have a crystal ball to see that the winning trades would be a small online bookstore, an online DVD rental service, or a video sharing platform with offices above a pizzeria in San Mateo.

The internet’s key innovation was the rapid dispersion of information and ideas of how to monetise this came gradually and from all quarters.

The AI space is different, the clear winners in 10 years will not be startups such as Arize AI or Toggle AI, it will be one of the Large Language Model (LLM) developers.

The winning business plan in the AI space is already clear, creating and selling artificial general intelligence to service-based businesses. In fact, it is OpenAI’s stated objective. It defines AGI as "highly autonomous systems that outperform humans at most economically valuable work."

When one of the LLM developers brings a credible, commercially viable AGI to market, the profits will be enormous. Chasing cost-cutting, much of the workforce of almost all services-based businesses will be supplemented with and gradually replaced with AGI. Instead of large wage bills, firms will pay millions in subscriptions for virtual workforces.

The sheer scale of the development costs of this product have already run into the billions of dollars, the only players with the capital available to pull it off are:

1. OpenAI (ChatGPT)

2. Anthropic (Claude)

3. Meta (Meta AI)

4. Microsoft (Copilot)

All of these are controlled, totally or partially, by the most valuable tech companies in the world.

As a result, magnificent seven in ten years’ time will not be comprised of AI companies with rags-to-riches growth fuelled by visionary understanding of the new technology, but instead the key business will be controlled by already dominant players in tech.