AI hype vs reality
GICP expert view / 4 September, 2024
The report ‘Technology: AI hype vs reality’ from CreditSights provides a comprehensive analysis of the current landscape of artificial intelligence, focusing on the gap between expectations and actual performance. You can read the full report here. Note that a CreditSights account may be required to access the report.
For credit analysts, understanding the financial and strategic implications of AI investments is crucial. Analyzing a company's AI strategy can provide insights into its long-term viability and growth prospects. Credit analysts should assess the financial health, strategic focus, and potential risks associated with AI investments, including technological obsolescence and regulatory challenges. By doing so, analysts can better evaluate the creditworthiness and future stability of companies navigating the AI landscape.
Key findings:
- Innovation and adoption of generative AI is high. However, the revenue opportunity does not seem to justify the near-term investments. Issues around AI to be addressed include hallucinations, guardrails to protect the user experience, copyright infringements, security and privacy concerns. Cost is another aspect - although that is expected to come down over time driven by improved GPU efficiency and availability, amongst other factors.
- In the race to beat the competition, budgets have opened up indiscriminate spending towards gen-AI. For most companies, it will be hard to justify the ROIC, not least for those whose internal development efforts will not lead to mass adoption.
- Tech companies are racing to have the most advanced LLMs (large language models), whilst other companies are racing to adopt AI to stay ahead of the competition – or through fear of falling behind the competition. Capex as a percentage of revenue was more than 20% for Microsoft and just shy of 20% for Meta in 1Q24, with the expectation for capex in the top hyperscalers (Amazon, Microsoft, Google and Meta) to surpass $200bn in 2024.
- AI hype has translated in higher market caps for the big tech companies, since early 2023. Earnings growth (some driven by gen-AI) and multiple expansion has driven this, along with earnings growth from other, non-AI related factors. Nvidia, where earnings growth has been driven particularly by gen-AI, now has a market cap nearly as large as Microsoft and Apple.
- Enterprise software has had limited success with monetization associated with gen-AI. Long term the preferred approach appears to be combining proprietary data with third party LLMs in order to create unique and differentiated services, although the article authors do not expect a step-function change in growth that justifies the infrastructure buildouts taking place in 2024.
If you are interested in learning more about AI, you may like the following webinars:
Staying competitive with the latest AI applications