It’s often the most creative and fantastical applications of artificial intelligence (AI) that make the headlines. Think about the potential of Generative AI to create personalised movies, with the ability for audiences to pick their own actors or even insert themselves into the storyline[1].
Or how about AI powered robotic bees that could support pollination and crop development to help stem biodiversity loss against a backdrop of declining honey bee populations[2].
Yet behind the headlines, AI is being put to work in areas such as drug development and chip design. Take Google DeepMind’s AI technology AlphaFold: in November 2020, it was recognised by a community forum of scientists as having found a solution to a 50-year old biology problem[3].
For decades scientists had been trying to devise a method to reliably determine how the attraction and repulsion of amino assets leads to seemingly spontaneous protein structures with intricate curls and loops, called protein folds. The AI model that DeepMind developed, using the same foundational techniques that powered AlphaGo to victory against the legendary Go player Lee Sedol in 2016, is able to accurately predict the 3D structure of protein folds. AlphaFold’s updated database, that today contains over 200 million structures was made freely available to the world in July of 20223.
Thanks to AI innovations like AlphaFold, companies are now able to sift through billions of different molecules to simulate and understand how various drugs interact with the human body on a molecular, cellular or even genetic scale. This has the potential to significantly reduce drug development timeframes, improve success rates and ultimately reduce costs.
Moreover, AI is also being used by the same chip designers, such as Nvidia, AMD and even Alphabet, whose processors are the workhorses of Large Language Models and Machine Learning systems, be it graphic processing units (GPUs). These are particularly good at parallel computing, data processing units (DPUs), which manage data transfer, encryption and compression in data centres or Application Specific Integrated Circuits (ASICs), that are complex processors designed for specific workloads.
A single printed circuit includes a raft of very small and very complex components, with many thermal and electromagnetic exchanges between them. Optimizing their placement involves simulating their interactions, which is a very iterative process.
For example, physical implementation and verification which were long and delicate processes performed by humans for months, can now be accomplished by AI in a matter of hours. What’s more, AI models learn and build upon all previous designs and make rapid simulations to come up with new chip architectures that are even more efficient and can be tweaked along the way by engineers.
US-based Cadence Design Systems’ latest innovation for example has helped increase a chips power, performance area (PPA) by as much as 20% using generative AI design[4].
Indeed, during a conference call in June, Cadence’s General Manager of Strategy and New Ventures claimed: “Our customers are realizing transformative results with our broad Generative AI portfolio […] for instance, customers have seen up to 60% improvement in timing [and…] up to 10x increased engineering productivity.”
Karen Kharmandarian and Alexandre Zilliox are portfolio managers at Thematics AM, an affiliate of Natixis Investment Managers
The provision of this material and/or reference to specific securities, sectors, or markets within this material does not constitute investment advice, or a recommendation or an offer to buy or to sell any security, or an offer of any regulated financial activity. Investors should consider the investment objectives, risks and expenses of any investment carefully before investing. The analyses, opinions, and certain of the investment themes and processes referenced herein represent the views of the portfolio manager(s) as of the date indicated. These, as well as the portfolio holdings and characteristics shown, are subject to change. There can be no assurance that developments will transpire as may be forecasted in this material. The analyses and opinions expressed by external third parties are independent and does not necessarily reflect those of Natixis Investment Managers. Past performance information presented is not indicative of future performance.
Additional notes
This material is provided by Natixis Investment Managers UK Limited (the ‘Firm’) which is authorised and regulated by the UK Financial Conduct Authority (register no. 190258). Registered Office: Natixis Investment Managers UK Limited, One Carter Lane, London, EC4V 5ER.
When permitted, the distribution of this material is intended to be made to persons as described below:
In the United Kingdom: this material is intended to be communicated to and/or directed at investment professionals and professional investors only. In Ireland: this material is intended to be communicated to and/or directed at professional investors only. In Guernsey: this material is intended to be communicated to and/or directed at only financial services providers which hold a license from the Guernsey Financial Services Commission. In Jersey: this material is intended to be communicated to and/or directed at professional investors only. In the Isle of Man: this material is intended to be communicated to and/or directed at only financial services providers which hold a license from the Isle of Man Financial Services Authority or insurers authorised under section 8 of the Insurance Act 2008.
To the extent that this material is issued by Natixis Investment Managers UK Limited, the fund, services or opinions referred to in this material are only available to the intended recipients and this material must not be relied nor acted upon by any other persons. This material is provided to the intended recipients for information purposes only. This material does not constitute an offer to the public.
It is the responsibility of each investment service provider to ensure that the offering or sale of fund shares or third party investment services to its clients complies with the relevant national law.
The above referenced entity is a business development unit of Natixis Investment Managers, the holding company of a diverse line-up of specialised investment management and distribution entities worldwide. The investment management and distribution subsidiaries of Natixis Investment Managers conduct any regulated activities only in and from the jurisdictions in which they are licensed or authorized. Their services and the products they manage are not available to all investors in all jurisdictions.
Although Natixis Investment Managers believes the information provided in this material to be reliable, including that from third party sources, it does not guarantee the accuracy, adequacy, or completeness of such information.
The provision of this material and/or reference to specific securities, sectors, or markets within this material does not constitute investment advice, or a recommendation or an offer to buy or to sell any security, or an offer of services. Investors should consider the investment objectives, risks and expenses of any investment carefully before investing. The analyses, opinions, and certain of the investment themes and processes referenced herein represent the views of the portfolio manager(s) as of the date indicated. These, as well as the portfolio holdings and characteristics shown, are subject to change. There can be no assurance that developments will transpire as may be forecasted in this material.
This material may not be distributed, published, or reproduced, in whole or in part. All amounts shown are expressed in USD unless otherwise indicated.
[1] Source: https://www.forbes.com/sites/kalevleetaru/2019/04/23/could-ai-replace-hollywood-with-personalized-movies/
[2] Source: https://isr.umd.edu/news/story/autonomous-drones-based-on-bees-use-ai-to-work-together
[3] Source: https://www.deepmind.com/research/highlighted-research/alphafold
[4] Source: https://www.forbes.com/sites/karlfreund/2023/06/13/cadence-ai-can-increase-chip-designer-productivity-by-over-10x
