Artificial intelligence (AI) has been the buzzword for some time, with rarely a day going by without talk of how the technology sector is being transformed by the latest AI-related innovations.
Yet a recent research trip to Silicon Valley and Boston – the world’s leading tech and innovation hubs – highlighted three significant developments that look set to have far-reaching impacts for industry, businesses, consumers and investors.
AI is transforming the software industry: In 2017, Jensen Huang famously predicted on LinkedIn that ‘AI will eat software’. While it is still early days, this vision is beginning to come true as AI brings new, more functional, and cost-effective software to market.
Unlike traditional Software-as-a-Service (SaaS) offerings, these AI-driven solutions automate, enhance productivity and offer a solution at a fraction of the price. If AI drives the cost of coding software to zero, the question becomes, why buy software versus building your own?
We believe that AI's potential to permeate every aspect of the economy is immense, akin to the transformative impact of electricity. It is poised to touch every pound of GDP, reshape industries, and drive substantial productivity gains.
A prime example of this shift is the emergence of companies such as Harvey.ai, a legal AI assistant built on OpenAI’s GPT platform, which exemplifies how AI can revolutionise traditional sectors.
By automating tasks such as legal research, document drafting, and contract analysis, Harvey.ai enhances efficiency and accuracy for legal professionals.
Harvey.ai has demonstrated its ability to reduce task times by 20-50%, significantly cutting costs and increasing capacity for handling clients. Its custom-trained models tailored to legal needs have seen widespread adoption among top-tier legal professionals and firms including Allen & Overy, PwC, Wolters Kluwer, and Baker McKenzie.
Baker McKenzie utilised Harvey.ai to analyse and compare clauses in commercial real estate leases, saving over $1m in costs. Wolters Kluwer reduced the average time for drafting legal memos from 4.5 hours to just 1.5 hours with Harvey.ai.
As AI continues to evolve, its impact on software and various industries is expected to be profound, validating Jensen Huang's prediction.
Large language model (LLM) training clusters are set to expand significantly
In the past seven years, a critical realisation has emerged in AI development – the scale of compute matters enormously. By increasing both compute power and data, we can train better models, thereby broadening the application of AI and unlocking substantial revenue opportunities across the economy.
This trend, evident since 2017, is expected to continue for at least another five years, strengthening the fundamentals and stock prices of companies that supply GPUs and networking equipment for AI data centres, as well as those that manufacture the equipment for these components.
The underlying assumption is that more compute leads to more advanced LLMs, which opens up new use cases for AI. This has driven the strategic decisions of major corporations, including Microsoft, Alphabet, and Goldman Sachs.
Demand for these clusters is evident across major technology companies and even sovereign AI initiatives such as Qatar. As AI technology progresses, each new LLM vintage introduces new sources of demand, driving further growth in compute scale. This underscores the transformative potential of compute power in shaping the future of AI.
Apple is poised to lead us into a super cycle of edge computing upgrades
After a challenging couple of years for edge computing, a significant upgrade cycle is on the horizon for devices such as smartphones, PCs, and laptops. This anticipated surge in upgrades is driven primarily by the need to accommodate new operating systems like iOS 18 and Windows 12, both set to launch later this year or early next year.
These new operating systems are designed to enable consumers to utilise AI functionalities directly on their devices, necessitating substantial increases in memory capacity.
The upcoming Windows 12, for instance, will demand a hefty 16 GB of RAM, a significant leap from the 4 GB required by Windows 11 and the mere 1 GB needed for Windows 10.
Similarly, iOS 18 will also push the memory requirements of smartphones, necessitating upgrades to ensure optimal performance and to leverage the advanced AI capabilities embedded within the new operating system.
This shift marks a substantial escalation in hardware requirements, reflecting the growing complexity and capability of contemporary software.
This upgrade cycle is poised to benefit not only major companies such as Apple and Arm but also smaller firms such as Onto Innovation and Ultra Clean technologies.
These companies are integral to the supply chain, providing essential components and technologies that support the enhanced memory and processing needs of modern devices. As a result, they are likely to see increased demand and potential growth opportunities.
However, this transition is not without its challenges. The entire value chain is currently experiencing significant bottlenecks, particularly evident in memory inventory levels.
These constraints highlight the strain on supply chains as manufacturers scramble to meet the heightened demand for upgraded components. This scarcity underscores the urgency for strategic planning and investment in production capacities to mitigate the risks associated with supply shortages.
The advancements we are witnessing today in innovation across the technology sector fill us with optimism for the future, and we eagerly anticipate the transformative developments the next decade will bring.
Clare Pleydell-Bouverie & Storm Uru are part of the Liontrust Global Innovation team. The views expressed above should not be taken as investment advice.