Why Silicon Valley isn't in charge of the AI boom

In a recent blog post, technology writer Ed Zitron outlines the argument that artificial intelligence (AI) is creating a bubble. The ethical problem with this spending, according to Zitron, is that it is drawing money away from other parts of the economy where it could be better used. His blame is specifically given to “dimwits” like Microsoft’s Satya Nadella, Google’s Sundar Pichai and AirBnB’s Brian Chesky. “The people propping this bubble up no longer experience human problems, and thus can no longer be trusted to solve them,” wrote Zitron.

It is possible that AI valuations have extended too far, and these companies’ massive investments will not pay off as expected. Investors’ Chronicle made this exact argument a few months ago. The point then was that the increasing capital intensity of these businesses would create profit headwinds and would lead to valuation contraction. However, this contraction has come even quicker than expected, with Microsoft’s forward PE ratio falling from 34 to 30 in the last month.  

The point Zitron’s article misses is that these Silicon Valley chief executives are almost solely responsible for leading consumers down this wasteful path. That view gives these technology leaders too much agency, when they are simply the most important part of profit-making machines.

In its modern form, AI has existed for almost a decade without much money being directed towards it. Investment only ramped up in the past two years because of the successful launch of ChatGPT as a consumer product. It took just two months to reach 100mn users, four times the pace of TikTok and 15 times the pace of Instagram.

Innovation is not just driven by a top-down agenda but is a response to the environment. Bell Labs was AT&T’s research lab during the 20th Century. Before AT&T was broken up by the regulator, the telecoms monopoly would fund Bell Labs to do research into speculative technology. Its success was remarkable, inventing the semiconductor, the first solar panel, and the laser. Since then, no one has been able to replicate Bell Labs' innovation track record.

Engineer Brian Potter laid out why Bell Labs was so unique. One of the reasons was it was part of a sprawling monopoly, which gave it a low bar for what constituted a valuable technology. “Even a tiny improvement that saved a few cents on a component or service would be large when multiplied by the enormous scale of the Bell System,” he wrote.

In other words, a large internal market created substantial economic benefits for AT&T to innovate. This effect was even more exaggerated because there were no barriers between Bell Labs and the rest of AT&T’s business. The technology could be easily disseminated and quickly lead to productivity improvements and earnings growth. There weren’t even any of the marketing costs that a consumer company faces when trying to sell a new product.

In its recent earnings, Microsoft reported it increased its capex spending by 78 per cent year-on-year to $19bn when leases are included. This was ahead of analyst forecasts and led to concerns spending was becoming irresponsible. However, chief executive Satya Nadella promised this was merely a response to the popularity of its AI servers, saying they “will only be bought for inferencing and everything else if there is a demand signal”.

This isn’t to say that companies always accurately predict consumer demand. Mark Zuckerberg’s Meta has lost over $40bn in the last four years investing in virtual reality technology. It is yet to make a profit, and last quarter it made a loss of $4.5bn. 

Zuckerberg even renamed his company Meta as part of the PR push to promote the technology. There have been examples of companies using smart advertising to encourage consumers to buy products that are bad for them and their health. However, so far customers have shown little interest in adopting the headsets. And if the virtual reality division continues to lose money, Meta’s share price will fall and eventually Zuckerberg will have to cut the investment.

Often people are displayed as powerless, forced into a world created by Silicon Valley. However, these technology chief executives are also powerless, forced to respond to the demand signals of their customers. There might be an AI bubble, but it hasn’t come out of nowhere. This spending is happening because people are using the technology and as soon as they stop using it, the spending will stop. They might be victims, but it is of the system, rather than a tyrannical techno-maniac. And in the long run, the system usually points us roughly in the right direction.