Brice Prunas, Oddo
291416.jpeg

Brice Prunas, manager of the ODDO BHF Artificial Intelligence, makes no secret of his enthusiasm for artificial intelligence as it quietly breaks through in all corners of the market. He not only invests in it but also uses AI to manage his fund. 

According to Prunas, artificial intelligence (AI) is a quiet revolution that will affect all sectors. “This is the first time that a company has been able to make a difference in the world. Today, forerunners are the software sector, transport and logistics, biotech and biopharma, internet and media. And video gaming also gets a mention because it has been given a new lease of life thanks to AI.”

He calls it quiet because it is little publicised while at the same time it is creeping into just about every company and business model. And a revolution because it will radically change the way we work. “The creation of the Internet was the starter but artificial intelligence will be the main dish”, said Prunas without hesitation.

“It is striking that the theme is not only driven by young, innovative companies but also by large companies with a long history. Disney is an excellent example, as it uses AI in all its activities, including optimising visitor traffic in its theme parks, so that visitors have to wait less and can consume more. Disney even uses AI and algorithms to make storylines in its productions for its Disney channel taste better to the audience!”  

According to the manager, the focus of AI is currently in the US and China, both of which have the most data. “Besides long-term and stable data, semiconductors are the second indispensable component for AI. And on this second point, the US has a big lead over China. Nvidia’s AI semiconductors are a good example of this. It will take China another 10 years to build a semiconductor industry. Clearly, we also invest in the semiconductor sector that helps make AI possible.”

Shortages?

When asked if the problems in the semiconductor supply chains have an impact on AI growth, Prunas nodded an emphatic yes. Prunas identified several pain points in the manufacturing process. “The first stage, chip design, happens in California at Nvidia and AMD, and then the semiconductors are produced in Taiwan. And in the third stage, Taiwan’s TSMC sends them to Malaysia for assembly. In the fourth stage, they are then taken to the US, Europe and Japan. Today, ship transport is limping along and there is a shortage of raw materials especially of substrates used for assembly.’”

According to the ODDO BHF AM fund manager, the shortages will drag on for a few more quarters but given AI’s strong secular growth trend, this will hardly cause an impact in the long run. “The pandemic, on the other hand, has provided an acceleration. A few months of lockdown have given digitalisation a huge growth spurt. Indeed, companies were inundated with data, the raw material for AI and its algorithms, during that period.”

Fund management based on AI

It is notable that the ODDO BHF Artificial Intelligence fund not only invests in AI but also uses it for its management. “We used AI right from the start of the fund in 2018,” explained Prunas. “We only do this in the first phase of the investment process through Natural Language Processing (NLP). ODDO BHF AM has concluded an agreement for this with a large European group specialising in data science. In other words, this first part is outsourced and thus a bridge is built between asset management and data science. Ultimately, this allows us to reduce our investment universe of 6,000 stocks, very broad given that AI themes can be found in all corners of the economy, to 200 names.” 

And what exactly does NLP do? First, a semantic analysis is done in which companies are selected based on a thousand AI keywords, such as “graphic processing unit (GPU)” and the number of times they are used in Bloomberg articles. At the same time, sentiment is analysed. The Bloomberg articles are screened by algorithms that interpret their underlying tenor. From the 200 remaining shares, a ranking with scores is made on this basis.

Thirdly, trend research is also carried out so that Nvidia and Google are not always at the top of the rankings and smaller stocks also get a chance. “In our fund, we have some smaller companies with a market capitalisation of EUR 1 billion. The most important contribution to our fund in 2021 will come from a small Canadian transport company, TFI International. We would not have found this company with only analysts.”

Human intervention

Brice Prunas pointed out that the 200 remaining shares is the yard in which they are looking to eventually arrive at a fund of about 40 names. “We use an in-house technology based on a multifactorial model. We look at different factors such as growth, valuation, momentum, etc. And then we get a new ranking that differs from our previous ranking. And then follows the fundamental analysis by the two managers.” Prunas adds that there is also a strong ESG constraint on his fund. “If a stock gets a too weak ESG note from MSCI then we have to sell our position. That has happened several times already.”

“Our preference today is for the software sector, which is our largest position in the fund at 30.1 per cent. It is striking that this segment is almost exclusively American. The semiconductor sector also holds a large position of 20.8 per cent, but we are rather reducing this interest. Healthcare (7.1 per cent), with biotech and biopharma as growth poles, can appeal to us, especially now that the sector is in the throes of digitalisation. The transport sector also still occupies an important position in the fund (7.5 percent) with Uber, which is entirely algorithmically driven, as an exponent. Media and internet including video gaming accounts for 12.8 per cent.”

Fund characteristics

Fund name: ODDO BHF Artificial Intelligence
ISIN code: LU19842267
Assets under management: EUR 288 million
Launch: 19/12/2018

Return (until 30 September 2021)
- Cumulative since inception: 58 percent vs. 69.7 percent benchmark MSCI World
- Annualised since inception: 23 percent vs. 35.3 percent benchmark MSCI World
 

 

Author(s)
Categories
Access
Limited
Article type
Article
FD Article
No