Story of the month
As listed companies begin to publish their financial results for 2023, we want to take the pulse of the adoption of generative AI, as it appears in these announcements. To do this, we use three metrics: 1) revenues from the public cloud; 2) intentions to invest in GPUs (artificial intelligence chips); 3) innovative use cases enabled by generative AI, using a concrete example from the real (non-tech) economy: Mastercard.
PUBLIC CLOUD SPENDING: SIGNIFICANT ACCELERATION, DRIVEN BY GENERATIVE AI
Amazon, Google, and Microsoft outrageously dominate the public cloud business (the hosting of data generated by their corporate customers). However, in recent years, these 3 companies have experienced several consecutive quarters in which their growth in this area has slowed. This appears to have been a phase where customers were optimising their IT spending in order to reduce it, or at least control it. In the 2nd and 3rd quarters of 2023, these sales figures began to stabilise. For its part, Q4 2023 showed a clear re-acceleration of the public cloud business, with very strong growth in sales of very high-value cloud services in Q4 (compared with the same period in 2022) for each of the major players: Amazon Web Services +13%, Microsoft Azure +28% and Google Public Cloud +26%.
It is therefore clear that the corporate customers of these 3 data storage giants have suddenly moved from a cautious approach (optimising their IT spending) to a bolder one (building a genuine ‘data’ strategy). Such migration to the public cloud is the essential prerequisite for their future use of Generative AI, because AI needs accessible data to be used. Microsoft has also gone a step further in terms of transparency, confirming AI as a growth accelerator. The group revealed that, of its +28% growth in the cloud business seen in Q4, +6% came from its generative AI offering (essentially its range of “Co-Pilot” tools).
INVESTMENT IN AI SEMICONDUCTORS: CONTINUED HYPER-GROWTH
When publishing their 2023 financial results, Amazon, Google, Meta and Microsoft all announced increases in their investments (“capex”) into the infrastructure required for artificial intelligence. These additional investments build on the already very high levels in 2023. We believe that most of this investment will be in semiconductors suitable for training and using (“inferring”) large language models. These models, of which Chat GPT is the best known but not the only one, enable interaction with an AI. This family of semiconductors includes GPUs and accelerators (mainly supplied by Nvidia and, to a lesser extent, AMD), but also ASICs (Application-Specific Integrated Circuits), supplied by companies such as Broadcom or Marvell on behalf of customers such as Microsoft or Google. The Q4 figures therefore confirm that the semiconductor needs of the digital giants are both very substantial and growing almost exponentially. As a result, some major technology players are looking to develop their own accelerators, rather than relying on external suppliers. Amazon and Google are the two most advanced in this process.
AN EXAMPLE OF THE GROWING USE OF GENERATIVE AI: MASTERCARD
In marked contrast to 2022, we have to admit that companies are rather quick to communicate on their strategy for using Generative AI in their core business. Let’s take the example of Mastercard, which, at the presentation of its annual results in late January 2024, revealed three use cases for Generative AI that are both very concrete and very innovative:
1) Shopping Muse: this service aims to create an immersive experience for consumers online as if they were in a shop, with personalised recommendations at the end;
2) Small Business AI: this tool provides chat-based support to SME owners to help them tackle the challenges they may face in running their business;
3) Mastercard Access: this offering gives Mastercard customers a single point of access to all the Group’s digital, identity and AI services.
