The large-scale investment cycle in data centres currently offers the best investment opportunities for AI investors, said AI fund manager Tobias Rommel in an interview. «Nvidia is the largest holding, but there are more companies benefiting,» said the fund manager, whose top-10 positions had on average tripled in five years.
Artificial intelligence is still in its infancy, said Rommel, who manages the DWS Invest Artificial Intelligence fund. The German fund house started this fund in late 2018, well before the outbreak of the AI craze. First-time investors had more than doubled their investments. «We launched the fund after a visit to Silicon Valley. US tech companies were investing heavily in AI research at the time and told us that it is going to give a big positive boost to sales and profits,» said Rommel.
«Amazon Alexa and other smart speakers quickly became popular with consumers at that time, which gave an idea about the potential impact of AI on society. Recent innovations like ChatGPT will further accelerate the adoption of AI by the masses.»
‘Nvidia is not a bubble’
AI stock prices had already risen sharply, raising the question of whether AI was just hype. According to Rommel, it is not. «Investors compare the enthusiasm around AI to the internet bubble at the beginning of this century, but we are still far away from this. Valuations are generally reasonable and certainly do not indicate a bubble. Our portfolio is trading at 25 times expected earnings. While that is more expensive than the broad market, a year ago the price-to-earnings ratio was also at 25.»
True, prices had risen sharply, but so have profits. Especially at the biggest positions, including Nvidia, Alphabet, Microsoft, and TSMC, profits were rising significantly. «Over the past five years, on average, the top-10 positions have more than tripled their profits. And growth prospects remain strong. We are looking for companies that can double their revenue and earnings in the next five to ten years and preferably show double-digit annual earnings per share growth.»
Much of the focus was on chipmaker Nvidia, which provided the key component for AI infrastructure. Investors were concerned about the valuation. According to Rommel, however, the risen share price was keeping pace with profits. «Sales have almost quadrupled in a year to $26 billion, so the P/E ratio has not changed. At some point, the chip market is saturated, but the cycle is only a year old, so we are far from that yet.»
Artificial intelligence was still in the early stages of a long-term mega-trend, Rommel believes. «Many companies tell us that AI is still in the testing phase and that it will be years before the technology is put to use.»
Data is a huge growth market
From the start, the fund focused on three categories: companies that provide the computing power to enable AI, companies that collect and provide data for AI, and companies that have a competitive advantage through the use of AI. «Computational power and data have been key drivers of AI innovations from the beginning. In the last decade, training AI models has become a million times faster thanks to increased computing power.»
Rommel sees the best investment opportunities in the large-scale data centre investment cycle right now. «Through manufacturers of semiconductors and networking equipment, we are capitalising on this.» Nvidia was the largest holding, but more companies were benefiting. «The big hyperscalers are developing their own, cheaper semiconductors, to reduce dependence on the market leader. Of interest are companies that help the hyperscalers do this, with software or chip production, for example. But also think of networking companies that ensure the chips can communicate with each other faster.»
In terms of data, these are companies that help collect data or have data themselves. «This is also a huge growth market because data is the lubricant of AI.» If companies were active in all three categories, they got a relatively high weighting in the portfolio. These included Alphabet, Microsoft, and Amazon, for example.
Key risks
Increasing regulation posed a threat to the sector, but Rommel does not expect things to get that bad. «It will not be easy, because governments are dealing with very large companies and in fact a global coordinated approach is needed. In a few years there will undoubtedly be more laws on AI, but this is expected to have little impact on our portfolio.»
He pointed to European privacy laws. This had benefited large companies precisely because it was difficult for smaller players to comply with it. The biggest risk for the industry was that the rapid progress of recent years would not be sustained, said Rommel.
«We are all now waiting for ChatGPT 5. If this new version is not a substantial improvement, the market may take this negatively and cause share price pressure in the sector. AI is not a new technology and development is going by fits and starts. In the long term, we expect excellent returns for our fund.»
This article was originally published in Dutch on InvestmentOfficer.be.