National Grid: How AI is Impacting the Energy Grid

With AI-powered data centres expected to only become more in demand, sustainability and energy efficiency are dominating the global conversation in the data centre industry.

By the end of the decade, data centres are expected to be responsible for consuming 9% of total electricity generation in the US - a figure that surges from the current 4%.

With the current rate of expansion, data centres in the US alone could demand the levels of electricity that is needed to power 40 million homes by 2030.

However, whilst AI is being touted as a challenge for operators, it can also be used for good to combat climate change.

The technology has the capacity to analyse energy consumption patterns, enhance the accuracy of demand forecasts and optimise renewable energy production, which makes it a useful resource for businesses seeking to decarbonise.

Industry research also suggests that AI could reduce global greenhouse gas emissions by 4% by 2030, which would be a 20% reduction from 2024.

With this in mind, Pradeep Tagare, VP of Investments at National Grid Partners — the CVC arm of major utility National Grid — explains how businesses can balance the innovative potential of AI with managing energy demands.

Please introduce yourself and your role.

I’m Head of Investments at National Grid Partners, leading investment in innovative startups that can help build the utility of tomorrow. 

That means helping to achieve ambitious net zero goals and improving our grid to meet the surging demand for energy.

I focus on a range of technologies, with an emphasis on AI to accelerate critical innovations.

In my tenure, we’ve made 50 investments totalling US$450m in investment across sectors ranging from battery technology to AI. 

Notable portfolio companies include carbon capture startup Captura, satellite-powered vegetation management AiDash, low-emission hydrogen production company Modern Hydrogen, optimised charging solution ev.energy, transmission line expander LineVision, high-performance conductor producer TS Conductor and more. 

I’m active with the Next Grid Alliance, an organisation that helps more than 120 utilities embrace innovation. 

I’m fortunate to spend time talking to ambitious founders and working with National Grid’s business units to better understand how new technology can transform utilities.

What is National Grid Partners and how does it contribute to both the energy and tech/AI landscapes?

National Grid Partners is the corporate venture arm of National Grid, a global utility serving over 28 million customers worldwide. 

By providing corporate venture capital, business development counsel and direct integration with National Grid's innovation team, National Grid Partners helps startups break out of ‘pilot hell’ to accelerate the energy transition and help innovators scale.

We focus our investments on four pillars derived from conversations with startup founders, utilities and other innovation leaders to identify the most pressing challenges in the clean energy transition.

• Future electric networks: Upgrading and rebuilding the grid to boost renewable energy, increase reliability and meet new electric demand while reaching net-zero
• Customer first: Delivering solutions to help customers manage a fair and affordable energy transition
• Efficiency through innovation: Deploying artificial intelligence, cloud and edge computing and other emerging technologies to improve operational performance
• Decarbonising gas: Delivering innovative technologies to reduce leakage and decarbonize our networks.

Given the significant energy demands of AI, what specific AI applications do you see as having immediate ROI for utility and cleantech companies?

We’re seeing and investing in many startups that are using a wide array of AI applications to deliver immediate ROI for utilities and the broader energy sector. 

For instance, portfolio company Exodigo uses AI to improve underground mapping to avoid costly missteps when expanding critical energy networks. 

LineVision is using AI to move more electricity via existing transmission lines — vital to enable more renewables and meet soaring demand to power electric cars and heat.

Cogniac uses AI image analysis to reduce the cost of field engineering inspections for utility companies. 

And Sensat uses AI to help owners of critical infrastructure visualise and collaborate via digital twins, resulting in fast project completion and lower costs. 

In a world where AI value isn’t always clear, these payoffs are substantial and directly drive decarbonisation, digitisation and lower costs.

How can utility and cleantech companies effectively integrate AI into their operations to maximise benefits while maintaining progress towards net zero goals? How does energy consumption factor in here?

Utility and cleantech companies can maximise the benefits of AI by using it to monitor fluctuations in renewable energy sources, predict supply and demand patterns with the next wave of smart meter tech and adjust power flows and manage congestion.

For example, Urbint — a National Grid Partners portfolio company — develops a risk management platform using AI to predict and prevent threats to critical infrastructure and worker safety, helping utilities and infrastructure operators make informed decisions. 

New AI platforms fueled by a broad range of data promise to help utilities overcome energy consumption challenges, improve operations and achieve net zero — while optimising their own energy consumption.

What strategies and technologies are available to reduce the energy consumption of AI? Can you provide examples of how training models are working more efficiently?

While electric cars and electrification of heat will be the primary demand drivers in the foreseeable future, data centre energy needs will add a significant load factor, thereby placing even greater stress on both energy generation and transmission/distribution.

Startups, as well as large technology companies, are racing to develop solutions to reduce AI data centre energy demand. 

Examples include DC cooling infrastructure (heat recycling), power-efficient computing platforms and servers and renewables plus storage.

While LLMs are the first generation of mass-scale AI technology, next generations will focus on more efficient algorithms, leading to a much better energy demand/computation curve. We are already starting to experience this with Small Language Models being deployed in enterprises.

With AI's potential to significantly reduce global greenhouse gas emissions, what role do you see AI playing in the broader context of decarbonisation efforts?

To achieve net zero goals, we need to measure, monitor and reduce the carbon footprint of large industries and countries, especially in the hard-to-abate category. 

AI can be used in all three steps to effectively manage and reach net zero goals. 

For example, AI is being used to design new materials and manufacturing techniques to significantly reduce the carbon footprint in areas such as cement and steel. 

AI tools are also being used to track and report Scope 1, 2 and 3 emissions, thus helping decarbonise the global supply chain.

Explore the latest edition of Data Centre Magazine and be part of the conversation at our global conference series, Tech & AI LIVE and Data Centre LIVE.

Discover all our upcoming events and secure your tickets today.

Data Centre Magazine is a BizClik brand