Generative artificial intelligence (Gen AI) has become a big focus. Investors wonder if this technology could usher in a broad wave of disruptive innovation, fuel productivity and investment, boost climate-saving solutions, and spur economic growth. With generative AI still in its infancy, is the elation premature? Derek Glynn outlines where Gen AI fits in the tech scene, what makes it attractive, and which risks investors should consider.
Setting the scene
Tech stocks have been attracting particular investor interest recently. In one area – cloud computing – we see the growth in expenditure bottoming out in the second half of this year and potentially reaccelerating in the fourth quarter. The growth of spending in this area has been slowing for many quarters, to around 20% year-on-year in the second quarter of 2023 from over 40% five to six years ago.
Cloud spending fits in with a longer-term digital transformation trend driving tech stocks. Shifting resources and IT workloads to the cloud gives companies cost efficiencies and the benefit of being able to dynamically adjust computing as demand changes. Given the scale and importance that cloud computing providers have in the tech ecosystem, a higher cloud spend is a welcome development for the sector.
Funding growth, though, has generally become more challenging with interest rates at around 5%, making debt and equity capital more expensive. As a result, many tech companies are now striving for profitability (if they were previously incurring losses), and those that were already profitable are focused on expanding margins. These efforts are likely to continue as long as interest rates are high, and investors should clearly welcome a focus on profitability and margins.
Gen AI – Advanced reasoning and breadth of use
One potential driver of increased cloud spending is the rapid progress of generative AI (which would also benefit graphics processing units (‘chips’), data storage, software and equipment).
Generative AI is a technology that can generate content from a text-based prompt using a large language model trained on a vast amount of data (also see our infographic). Originally designed to predict the next word in a sequence, it can now reach advanced levels of “reasoning” with access to data and the right type of training.
This technology is still in its early stages, but there’s already a breadth of use cases, from writing essays to summarising data or even assisting software developers with generating code. It’s conversational in the sense that AI responses incorporate previous human interactions, so the dialogue with an AI chatbot can flow in a natural way.
Gen AI is rapidly evolving into more multi-modal formats. From text-based prompts and output initially, the technology is now being used to generate images, video and audio input. If these capabilities continue to expand, we could imagine a world where we have access to specialised virtual assistants via a smartphone or an augmented reality device. For example,
- An AI virtual shopping assistant who can recommend clothes based on our sizes, preferences and budget, and then source those clothes from local online merchants
- A virtual health assistant who creates a meal plan and workout routine that’s customised to our schedules and preferences based on data it collects from synching with our wrist monitor
- An entertainment assistant that generates customised video clips based on the genre of movies we’re interested in.
Gen AI can unlock new avenues of growth, and companies could do more with less. Developers can write more code per hour; customer service helpdesks can be AI-assisted and handle more questions without the intervention of a human. This drives cost efficiency.
The businesses best positioned to benefit would be those that develop the foundational models themselves; those providing the computing and (cloud) storage resources to enable AI; and the hardware or semiconductor supply chains.
There are also companies with proprietary data that can find ways to weave this technology into their existing offering to develop new products or improve existing ones to better retain customers.
…and the risks
Whenever a new technology emerges and gains rapid adoption, some companies will be disrupted or disintermediated if they don’t adapt quickly. As gen AI matures, it could weaken the economic “moats” that now protect industry leaders or call into question the durability of some businesses’ growth.
Software businesses without a strong value proposition or without proprietary data could face more competition. Content itself may become more commoditised because its generation could be almost free and consumers could personalise it themselves. Gen AI could also be a competitor to online education companies as it provides quality tutoring in certain subjects.
So, is generative AI being overhyped?
There has certainly been an explosion of interest from investors and the media. Consumers are trying it out and management teams are testing it to see how it could impact their business.
It’s early days. Companies need to be confident that their own and their customers’ data will be secure, that there are strong cybersecurity defences, that the AI output is accurate, that any bias in the AI models is minimised or eliminated, and that they are not violating copyright laws. Ironing out these and other issues could prolong the period before companies adopt it and realise the benefits.
We have already seen an expansion in market valuations for certain tech stocks exposed to the gen AI theme. It may seem overhyped in the short term but given how transformational it can be we believe many observers are still underestimating its longer-term impact (also see Whither tech stocks? Should we ask ChatGPT?).
This technology ultimately looks set to impact all of us at some level, just as the internet and smartphones did, so investors should pay close attention to the development of generative AI.