#82 Cloud Computing's Nuclear Option
Cloud Computing Ambitions Get a Boost from Nuclear Energy; Transparency in Generative AI - Exploring the Copyright Disclosure Act and Its Effects
Today, Saurabh Todi dissects the reasons why AI may be driving a renewed push towards nuclear energy adoption, and Rijesh Panicker shines a light on the US’ newly proposed “Generative AI Copyright Disclosure Act”.
Also,
We are hiring! If you are passionate about working on emerging areas of contention at the intersection of technology and international relations, check out the Staff Research Analyst position with Takshashila’s High-Tech Geopolitics programme here.
Cyberpolitik 1: Cloud Computing Ambitions Get a Boost from Nuclear Energy
— Saurabh Todi
Last month, Amazon Web Services (AWS) announced a $650 million deal for a data centre campus adjacent to a nuclear power plant in Pennsylvania to build several data centres, representing the first campus with direct nuclear power access. In January, Microsoft hired a director of nuclear technologies to lead a program to develop small-scale nuclear reactors to power data centres. The project includes liaising with engineering and design teams to ensure "technical feasibility and optimal integration of SMR and microreactor systems." This is quite an exciting development for the nuclear energy sector and technology companies focusing on AI.
Amazon and Microsoft, along with other major tech companies, are exploring nuclear energy to power their data centres for several reasons:
First, AI requires a lot of power, and the computing needs of AI, particularly Nvidia's graphics processing units, could substantially bump up the proportion of energy used. For example, Nvidia’s A100 AI chip has a constant power consumption of roughly 400W per chip, while that of its latest microchip, the H100, is nearly double that at 700W. So, if a full hyperscale data centre with an average of one million servers replaced its current CPU servers with these types of GPUs, the power needed would increase 4-5 times to around 1500MW – equivalent to a nuclear power station.
Second, Amazon and Microsoft have committed to reducing their carbon footprint and achieving sustainability goals. Nuclear energy is a low-carbon energy source, emitting minimal greenhouse gases compared to fossil fuels like coal or natural gas. By incorporating nuclear power into their energy mix, these companies aim to reduce their overall carbon emissions.
Third, nuclear power provides a stable and reliable source of energy. Unlike renewable energy sources such as solar or wind, which are intermittent and dependent on weather conditions, nuclear power can generate electricity consistently, ensuring the uninterrupted operation of data centres critical for online services. Furthermore, by diversifying their energy sources, technology companies like Amazon and Microsoft can enhance their energy security and reduce vulnerability to supply disruptions or price fluctuations in traditional energy markets.
Fourth, technology companies are also aiming to achieve cost efficiency. While nuclear power plants have high upfront construction costs, they can be cost-effective over the long term, especially considering fuel costs and stable energy prices. Emphasis on advanced technologies such as Small Modular Reactors, or SMRs, further underscores this point.
Apart from these factors, the interest in technology companies also pushes innovation in the nuclear energy sector. As Amazon and Microsoft are technology leaders, their interest in nuclear energy reflects a broader trend in technology companies seeking alternative energy sources. This increased interest would provide impetus to technological innovation in the energy sector. Advances in nuclear reactor design, such as small modular reactors (SMRs) or advanced reactor concepts, offer more efficient and safer nuclear power generation opportunities, aligning with the tech companies' focus on cutting-edge solutions.
Although these efforts are notable, experts don’t expect data centres powered by SMRs to come online for the next decade or so, given the emerging nature of the technology. Nevertheless, these developments add another factor to the recent global nuclear energy renaissance.
India is set to host the Quad Leaders' Summit in 2024. Subscribe to Takshashila's Quad Bulletin, a fortnightly newsletter that tracks the Quad's activities through the Indo-Pacific.
Your weekly dose of All Things China, with an upcoming particular focus on Chinese discourses on defence, foreign policy, tech, and India, awaits you in the Eye on China newsletter!
The Takshashila Geospatial Bulletin is a monthly dispatch of Geospatial insights for India’s strategic affairs. Subscribe now!
Cyberpolitik 2: Transparency in Generative AI - Exploring the Copyright Disclosure Act and Its Effects
— Rijesh Panicker
The “Generative AI Copyright Disclosure Act” proposal, unveiled this Tuesday by Representative Adam Schiff takes direct aim at the use of copyrighted data by large generative models, requiring companies to disclose the copyrighted material used to train their generative AI models.
The Act would mandate people creating training datasets or making significant changes to existing datasets to submit a notice to the Register of Copyrights with details of copyrighted work used and the URL for any publicly available data.
This notice, to be filed within 30 days of an AI model being available for public use, will also apply to existing AI models, which need to provide this information within 30 days of the bill going into effect. Financial penalties for violations begin at $5,000 (no maximum limits). The Register of Copyrights will publish and maintain an online database with all notices available to the public.
This act is similar to provisions in the EU AI act, agreed upon last year, which specifies that general-purpose AI models (a superset of generative AI models) must comply with EU copyright law and publish detailed summaries of the content used for training.
Industry, meanwhile has moved forward. Since facing multiple lawsuits last year, OpenAI has licensed content from Axel Springer and Associated Press and is negotiating with several other publishers. On the other hand, large producers of data like Reddit have also looked to monetise their content.
With the incentives this act seeks to put in place, one should expect to see a deeper market develop to create and sell datasets. This will ensure that even as artists and publishers get paid for the use of their data, they are always subject to the discipline of competition. Of course, it benefits large owners of data such as Google and Microsoft, who are less susceptible to market pressures.
There could also be some unintended consequences. As observers like Gary Marcus have pointed out, the newer generations of generative AI models are clustering together in terms of performance. If this new law comes into being, this is likely to become even more accelerated as the datasets being used will become public and any performance advantage due to data will be rapidly narrowed down. In the long run, this should lead to an increased focus on technical breakthroughs rather than just scaling data.
This proposal could be a good first step that drives transparency in how generative AI models are trained. Lawyers for OpenAI have argued previously it is not possible to create large models without copyrighted data and that using such data for training falls under fair use. Meanwhile, creators and owners of such data argue that the way large generative AI models are using this data for inference amounts to theft. The truth lies somewhere in the middle and publishing detailed listings of training sets is a first step to discovery.
Admissions are now open for Takshashila’s Postgraduate Programme in Public Policy. This comprehensive, multidisciplinary, and entirely online 48-week program is tailored for those aiming to delve deep into both the theoretical and practical aspects of public policy. Apply by 22nd April to avail yourself of the Early Bird Deadline (10% scholarship)!
What We're Reading (or Listening to)
[Opinion] Warmer Moscow-Beijing ties are a chilling signal for New Delhi, by Manoj Kewalramani and Amit Kumar
[Takshashila Blog] An Invitation to Inquiry: Rekindling Flame of the Mahabharata, by Wg Cdr Amit Gaur
[Article] How A.I. Tools Could Change India’s Elections, by Suhasini Raj