We are compelled to wager on AI

According to Terry Tanaka, it's impossible to predict whether AI will change the world just yet, but failure would undoubtedly be extremely expensive

I sometimes question whether AI is truly desired by people outside of Silicon Valley. Indeed, the advancements in technology in this field are astounding. Observe how AI tools and agents analyze data, write text, create photorealistic images and videos, andpossibly most unsettling of alldirectly control other computer programs through human-designed text and visual interfaces. You can't help but be impressed. Simultaneously, a lot of people claim that AI is being imposed upon them (noting the abundance of AI features that cannot be disabled in every piece of software we use), they don't feel the need or desire to experiment with AI, and they believe that concerns about the risks and social repercussions are being disregarded.

All of this does not imply that AI will not be beneficial or flourish, which are two different things. In the early days, there were a lot of people who were skeptical of the internet's economic potential; today, there aren't as many. Concerns regarding the negative effects of social media have grown since it spread like wildfire. But it seems true that big tech is pressuring us to use AI more. This illustrates how much money these businesses are investing in it and how much they depend on it to be profitable.

Substantial AI investment.

US private investment is changing.

Other categories include software and equipment for information processing.

At an event last week, Julian Bishop and James Ashworth of Brunner Investment Trust (LSE: BUT) succinctly outlined the reasons to exercise caution. Take a look at the amount of money invested this yearbetween £400 and £450 billion. Remember that businesses claim they intend to continue doing this until 2030. The capital expenditure (capex) is estimated to be £3 trillion. As a partial proxy for core AI-based revenues, consider OpenAI and Anthropic's reported revenues, which could reach £20 billion this year. This is expanding quickly, but much more growth is required to generate a respectable return on capital expenditures of this magnitude. This has consequences for free cash flow because it contrasts sharply with how capex-light the tech giants used to be. Take note of how this capital expenditure moves through the market, from data centers being built by real estate developers to energy infrastructure. Bishop says, "A lot of the market's continued strength depends on AI working out." (See also Joel Suss's Financial Times chart above, which suggests that US investment is now dominated by AI spending. it).

Even if investors lose out, a capital expenditure bubble may have long-term advantages. Consider railroads and canals. However, Ashworth points out that the funds are being used for transient assets. Nvidia's expensive chips might only be in use for five years or less.

Positively, the US tech companies are very profitable. In the long run, it might not be unreasonable for them to spend a lot of money now to protect their positions from possible dangers. Bishop and Ashworth claim that private markets, which own some of the mega caps, have the most froth. Nevertheless, they are obviously watching for indications that the cycle is changing, and this seems appropriate. The possibility of an AI revolution cannot be disregarded, but if it disappears, there will be significant consequences because the tech companies account for 35% of the SandP 500.