AI investment: A big race is currently going on between the big tech companies around the world regarding artificial intelligence i.e. AI. According to a report by UBS Global Research, the world’s six largest hyperscalers (big cloud and tech companies) are set to invest a whopping $7.7 trillion on AI between 2026 and 2030. About $6.8 trillion of this would be different from their normal spending. After such a huge investment, the market is now questioning whether these companies will be able to earn according to this cost. Recently, weakness in the bonds of some neocloud companies such as CoreWave has added to this concern. However, Sunil Thirumalai, Emerging Markets Strategist at UBS, says this pressure from the credit market should not be considered a failure of AI investment as a whole. Neocloud companies account for only 5 to 7 percent of total spending on AI data centers over the next four to five years. Therefore, they are not a major part of the global AI story.

A huge challenge of a trillion dollar profit

Until now, big tech companies are earning more than 40 percent return before tax from businesses like software, social media, e-commerce. These works did not require the installation of much physical property or machinery. But AI is changing this whole math. For AI, companies have to build large data centers, which is burdening them with large assets. UBS estimates that the change could reduce companies’ return on investment (ROIC) by 20 to 25 percent.

Even after this reduction, if these companies want to earn a 15 percent return on their investment by the year 2031, they will need to generate about $1 trillion in additional earnings before interest and taxes (EBIT). This amount is slightly less than double the current combined earnings of all these companies. According to current cloud margins, to achieve this target, companies will need to generate approximately $3.4 trillion in additional revenue. This simply means that the biggest question now is not the demand for AI, but the earning (monetization) from it.

Just increasing productivity is not enough

UBS says that this spending on AI can only pay off if two conditions are met. The first condition is that AI must demonstrate a real improvement in efficiency, i.e. productivity. Another condition is that companies creating AI must be able to convert that benefit into their earnings. Currently the total world economy is about 110 trillion dollars. If it continues to grow at the pace of the last decade, it will reach $136 trillion by 2031, even without any AI help. The report estimates that if AI boosts annual productivity growth by 0.5 percent, 1 percent, or 1.5 percent, global GDP could increase by $3 trillion, $7 trillion, or $1 trillion, respectively, by 2031.

But just because the economy benefits, it does not ensure that companies investing in infrastructure will also find money in their pockets. UBS compared it to the era of the old telecom, dotcom and cable industries. Those sectors increased the momentum of the overall economy, but were not able to profit accordingly. Internet, e-commerce, social media, streaming, smartphone companies took real advantage of the networks they created. However, the relief is that today’s hyperscalers have more ways to make money than the telecom companies of old.

There is good news about India

The recent sell-off in the market was linked to rising bond yields of tech companies. UBS clarified that the rise in hyperscalers’ bond yields is mainly due to the rise in US Treasury yields. The real pressure of debt is only on companies with weak balance sheets. However, analysts believe that there is a danger that the pace of expenses may increase too much relative to earnings. Apart from this, availability of electricity is also a major challenge for running data centers. UBS’s global semiconductor team believes that enough power can be obtained to run the data center as planned. But this requires that all work goes smoothly, laws are supportive, grid capacity is not delayed.

Amidst all these global ups and downs, there is good news about India. Despite major movements in emerging markets, Asia and global tech stocks, the Indian market has been very strong since October 5. This shows that the pressure on global AI spending has largely spared the Indian market its direct impact so far. Now investors are keeping an eye on who really profits in this game of AI, as there is little margin for error in this $7.7 trillion bet.