Rapid advances in low-cost AI development highlighted the growing commoditisation of artificial intelligence and reinforced the importance of cybersecurity and proprietary data.
Last week saw the announcement of yet another new AI model. This one, developed by researchers at Stanford and Washington universities and named S1, is designed to compete with OpenAI’s o1 model. It was reported that the S1 model was trained for less than $50 in cloud computing resources and completed its training in under 30 minutes using 16 Nvidia H100 GPUs – the very units Microsoft and Meta are estimated to have acquired over 150,000 of.
The trend in AI appears to be one of increasing commoditisation. With widely accessible and affordable models becoming available, simply having AI technology may no longer provide a competitive edge. As a result, the focus is likely to shift to those who have unique applications for the technology and can protect themselves through other means, such as holding a proprietary dataset.
Securing valuable datasets is crucial, and companies such as Broadcom, CrowdStrike, and Palo Alto Networks, which provide assurances to their customers, are included in the First Trust Nasdaq Cybersecurity UCITS ETF (held in the T. Bailey Multi-Asset Growth and T. Bailey Global Thematic Equity Funds). In the meantime, innovation will continue and we will see new applications and novel use cases across a wide range of industries. The Polar Capital Artificial Intelligence Fund, held across all three T. Bailey funds of funds portfolios, aims to capitalise on the expanding AI sector. These funds have achieved year-to-date gains of 10.0% and 6.5%, respectively.
