
Generative AI: the opportunities and risks for investors
In recent months, there’s been no avoiding mention of ChatGPT in the media. It’s the foremost example of generative artificial intelligence (AI), a form of technology that creates new content, such as writing, audio or art. How it will impact the way we live is an open question – and one that investors need to address.
According to analysts, new technologies often generate initial excitement before falling short of expectations. Thereafter, people start to find practical applications and the technology increases productivity at a steady level. This can create opportunities, as well as risks for investment. The research team at Investec Wealth & Investment is currently exploring the outlook for generative AI.
What is the impact of generative AI?
ChatGPT first entered the public consciousness as Microsoft unveiled its use in its Bing search engine and Alphabet (the owner of Google) responded with its much-maligned Bard chatbot.
Internet search is a vast and lucrative market, estimated to be worth $200 billion a year with a gross margin of around 80%. So it’s a key battleground for the tech giants. However, the impact of this new technology may be limited.
“So far, generative AI is not well suited to all types of search,” says Simon Lapthorne, Senior Research Analyst at Investec Wealth & Investment. “And that’s because of its tenuous relationship with the truth”.
With large language models (LLMs) like ChatGPT and its rivals, a clear problem is their tendency to ‘hallucinate’, or provide incorrect information in a convincing format. This appears to be inherent. Stephan Pretorius, the CTO of advertising agency WPP, recently noted that LLMs are designed to produce “plausible next words” rather than facts. In a similar vein, AI expert Gary Marcus describes LLMs as “autocomplete on steroids”. 1

It could make more sense to put money behind the people who sell the prospectors their picks and shovels – in other words, those who will prosper.
Does its cost inhibit potential growth?
Another consideration is cost: generative AI costs approximately ten times more than traditional algorithms to perform a search and, at present, is slower.2
Then there’s the incumbency effect. Bing + ChatGPT will have to be demonstrably superior to Google + Bard to have much chance of gaining recognition and market share. As Simon notes: “Google is a verb; bing isn’t”.
One of the barriers to penetration is smartphone use. As much as three-quarters of our internet searches are performed on our smartphones, where Apple holds around a two-thirds market share. This means that roughly half of the entire search market flows through Apple – which is why Alphabet reportedly pays Apple $15bn per year to keep Google as the default search engine on Apple devices.
How might generative AI gain traction – and where might the opportunities be?
At Investec, the research team sees a useful analogy in the gold rushes of the nineteenth century. “Rather than backing individual prospectors to strike gold,” says Simon, “we think that it could make more sense to put money behind the people who sell the prospectors their picks and shovels – in other words, those who will prosper”.
The team has been monitoring ASML, the Dutch semiconductor giant, and Nvidia, which makes graphics processing units (GPUs) and other AI-related hardware and software. Together, these companies will provide much of the immense computing resources needed to power generative AI – regardless of who wins the battle for search.
Nvidia is evolving into a ‘full-stack’ provider, covering the entire AI value chain. And it’s also creating an AI platform analogous to the iOS and Android operating systems in mobile phones.
Who loses out?
If ASML and Nvidia could prosper from the AI boom, could any sectors or businesses be at risk? “The simple answer,” says Simon, “is that we don’t yet know”. A number of industries have come under the spotlight, but at the moment it looks as though those who rely on proprietary data are likely to be fairly well insulated and, indeed, advantaged by AI. However, there’s a good chance that copyright protections could be severely tested in some sectors.
In creative industries such as advertising and graphic design, generative AI could have significant implications. “However, for now we see it as more of a co-pilot than a substitute in these sectors,” says Simon. Of course, much depends on how quickly these technologies evolve.
From history, we know that new technologies do eliminate tasks. But technologies can also increase productivity by freeing up employees to undertake higher-value work. “New technologies may open up possibilities of greater scope and scale for businesses,” says Simon. “While, there are risks to all investments, this growth is what drives more jobs and opportunities.”
Please get in touch if you’d like to speak to one of our team today.
More investing insights from Investec
Browse articles in