Artificial intelligence (AI) is the simulation of human intelligence processes by machines. Human cognitive functions have machine versions: thinking is replicated by machine learning; vision by mechanical vision; language is equivalent to natural language processing. In order to create all these features, computers need data. Lots of data.
We find ourselves at the beginning of the data era. Research by Accenture and McKinsey (“Straight talk about big data”, October 2016) suggests that only 0.5% of all data is analysed, 80% of company-generated data is unstructured and 90% of the world’s data has been generated in the last two years. This booming environment will eventually involve the whole data supply chain. It starts with data’s most basic need: storage. Then, you need to organise it. Last, there are those companies which help to visualise and manage data. When a database reaches a certain size, there is the possibility of developing artificial intelligence. This is this whole algorithm supply chain that we would like to present to you here.
The value of the AI hardware sector is estimated to grow from USD 30 billion in 2018 to USD 110 billion in 2024, generating new growth for the semiconductor industry in particular, but also for manufacturers of hard discs and networking equipment. Altogether this would see significant growth in the whole semiconductor market, which is currently estimated to be worth around USD 400 billion. This part of the artificial intelligence industry is the one we can touch; most of the rest is either virtual or in the cloud.
In the future, most data will be stored in data centres instead of on local servers.
This transition is still in its very early stages: according to Citigroup (“Quantifying the Cloud”, April 2017), less than 10% of companies’ data are in the cloud. In June 2019, public cloud revenues grew by 54% (according to Goldman Sachs “Cloud Report”, June 2019). This massive growth shows how crucial this transition is. The first reason for this growth is that it is cheaper to host your data on the cloud (ten times cheaper according to our estimates for a USD 5 million project), but it is also more scalable and you immediately get the most innovative solutions. For comparison, in 2000 at the peak of the dotcom bubble, Oracle’s revenues were growing at just 15%. Here we can see how much faster the world is moving today. Amazon is the dominant player in this industry, followed by Microsoft’s Azure and Google Cloud. In China, Ali Baba is the number one host data centre operator.
We view the software sector – and most particularly software as a service (SaaS) companies – as those companies with the purest exposure to the artificial intelligence theme. Indeed, SaaS companies play a key role in the data supply chain, as they are the data entry gates. These companies store, manage and visualise their data through software.
SaaS companies are the beating heart of data.
In our view this is largely where the most sophisticated use of data and artificial intelligence will take place. Marc Andreessen wrote a now well-known article in 2011 entitled, “Why Software Is Eating the World”. At the time, there were just 64 software companies listed on the US stock market. Now, there are 185 and 2018 was a record year with twenty software IPO; thus far in 2019, eleven new companies have launched in the software space. We expect this trend to continue, as close to half of the so-called “unicorns” (private companies valued at more than USD 1 billion) are software companies. These are the IPOs of the future. The consensus expects revenues in the listed SaaS sector to grow by an average of 24% this year, which is higher than the average revenue growth over the last ten years (this came in at 17%), showing how much this growth has sped up over the last decade. Further, the overall sales revenues for this group has risen by more than 20% over the last twelve months, showing even faster software adoption by companies. Indeed, we see firms’ capex flowing into the software sector as a way of modernising, automating and rationalising any company.
This transformational trend is not without risk. On the business front, these architectures are increasingly reliant on cybersecurity. The IT security market is currently forecast to grow by 8.7% to USD 126 billion in 2019, while overall IT spending is expected to grow by 3.2%. It should be noted that the identity and access management segment is becoming critical, as is infrastructure and network security. Secondly, we are facing the emergence of dominant companies in the data management business in the same way that we saw the emergence of FANG for consumer data. Regulation might come to rule this new Eldorado, but we are still at a very early stage in the process, with the penetration of the cloud still estimated to be less than 10% of its potential.
Julien Leegenhoek
Equity Analyst & Advisor