Powering data centers is a herculean task

There are many ways to invest in the artificial intelligence (AI) boom. Some highly developed chips like Nvidia have taken the spotlight, but there are more ways to invest in a broad wave of AI adoption than just semiconductor chips. Chips are essential, but without the infrastructure to support them, they are useless. Investors should also consider the sector powering the data centers needed to advance artificial intelligence.

According to the Electric Power Research Institute, data centers are expected to consume up to 9% of U.S. electricity generated annually by 2030, up from 4% in 2023. That doesn’t sound like much, but between replacing an old power plant and regenerating new power tables the increase is significant. The reason is artificial intelligence. The energy requirements of AI computers are enormous. To put energy consumption into perspective, AI queries – just ask ChatGPT to clean up your emails to your boss – can require about ten times the electricity of traditional internet searches. Training new models and running more complex queries consumes even more energy.

The demand to produce more energy is so great that it is actually becoming a major obstacle for companies looking to innovate in artificial intelligence. The ideas are there, computing chips are expensive, but they are there, but you have to figure out how to meet the energy demand to keep the data center functioning.

As this space continues to evolve, we believe it is worth exploring opportunities in the power generation space where investors can benefit from partnerships between upstream and downstream players in the AI ​​market.

Data centers need electricity… and lots of it

To understand the first stage beneficiaries providing key contributions to the development of artificial intelligence, it is worth analyzing what data centers are used for when it comes to artificial intelligence. Data centers are the backbone of the AI ​​boom because they house the hardware necessary to train AI models. As high-tech companies get larger and the amount of hardware needed to maintain cloud operations and services skyrockets, the energy demands of hyperscalers follow suit.

Hyperscalers – companies that operate cloud computing infrastructure – provide the storage and computing power needed to train artificial intelligence and face the difficult task of deciding where to source the energy needed for everyday operations. Major players in this space include Amazon Web Services, Microsoft Azure, and Google Cloud, all of which are partnering with local utilities and energy providers to ensure their energy needs are met.

With the radical shift to cloud storage, electricity demand has become the biggest issue facing the industry. Hyperscalers providing infrastructure and platform services simply can’t get enough of it, and the world of institutional finance is seizing this opportunity. As we recently saw, Blackrock has partnered with Microsoft (MSFT) and MGX (UAE’s AI fund) to announce the launch of a $30 billion AI infrastructure investment fund focused on building data centers and energy infrastructure. While this may initially seem like a huge investment, it pales in comparison to the total AI investments of the top five hyperscalers over the next few years. For context, current projections indicate that investment levels will total $1 trillion in 2027, according to S&P Global Market Intelligence.

Straight to the Source

The way energy is produced and sold to large hyperscale consumers will change. While in the past many data centers could rely on local electrical grids, the future of data center electrification lies in the construction of dedicated power sources and long-term power purchase agreements (PPAs).

Hyperscalers prefer PPAs to spot market pricing because they are able to guarantee stable energy prices for a longer period, usually more than ten years. This level of certainty provides a hedge against undesirable price fluctuations that could impact the financial performance of hyperscale companies and put upward pressure on the already significant costs of model training. Perhaps more important, however, is the desirability of long-term contracts for developers and investors. PPAs with cash flow hyperscalers give developers the ability to secure financing for the rapid construction of new power generation facilities. Knowing that you have a reliable contractor allows you to better forecast your profits! Moreover, many power generation projects are dedicated solely to data centers and disconnected from the wider grid – a minimum rate of return agreement is essential for a project that takes years to break even and has only one customer.

Finally, it is worth noting that while renewable energy sources such as wind and solar are the preferred long-term energy source for data centers, demand is currently high enough that other options are also being considered. Nuclear power is back in the spotlight, with natural gas turbines enabling rapid deployment. Each power source has advantages and disadvantages – some are more flexible, while others are more difficult to adapt to the demands of the data center. There will be many different solutions to the power problem available over the next few years, and many solutions that started as off-grid or dedicated power supplies may eventually become part of the larger power system.

Opportunities in renewable energy sources

Long-term demand for electricity is an opportunity to accelerate the construction of clean energy solutions. Top hyperscale manufacturers have set ambitious goals to reduce their carbon footprint – renewable energy is a longer-term necessity for them. Leading companies include Google, which plans to operate its data centers based on emission-free energy by 2030, while Microsoft and Amazon hope to switch to 100% renewable energy by 2025.

This massive undertaking is already underway, as Microsoft (MSFT) and Brookfield Renewable Partners (BEP) signed the largest-ever clean energy deal for their data centers in the US and Europe this year, effectively adding 10,000 units of renewable energy capacity. 5 gigawatts from 2026 and is estimated to cost over $10 billion. Keep in mind that 10.5 gigawatts is 3 times the amount of electricity used by all of the data centers in Northern Virginia – widely known as the data center capital of the world due to favorable state tax incentives and best-in-class power access and connectivity Internet.

Given the continued growth in demand for renewable energy due to corporate responsibilities, there are several ways to invest. We have already touched on the issue of financing, where groups such as Blackrock and Brookfield raise capital to finance the construction of new energy projects. Then there’s the manufacturing side, companies that make solar panels or electrical components, or companies that build wind turbines like GE Vernova (GEV). Finally, there are utilities that want to own and operate power generation facilities for the next several decades, such as NextEra Energy (NEE) and Constellation Energy Corp. (CEG).

Finally, there are cutting-edge power innovations that are less proven. For example, innovative nuclear technology such as small modular reactors is developing rapidly and has received extensive support from the US Department of Energy. These new reactors, built by companies such as NuScale Power (SMR), are smaller than traditional reactors and do not have to be custom-built on site. The process of generating electricity from nuclear fission is the same, but the benefits include reduced construction times, improved passive safety features, greater deployment flexibility, and more efficient use of fuel. Further investment in capacity and regulatory compliance will be required to ensure widespread adoption in U.S. data centers, so patience is key.

Far from being finished

The rapid development of artificial intelligence and following energy demand create significant opportunities in the energy sector. As hyperscale players like Amazon, Google, and Microsoft expand their AI infrastructure, demand for massive amounts of electricity will drive investment in both traditional and renewable energy sources. While fossil fuels can provide immediate solutions through existing infrastructure, in the longer term, the focus is on renewable energy sources such as wind, solar and even nuclear power to achieve sustainable development goals.

Innovative, established companies in the energy sector, especially those forming strategic partnerships with hyperscalable players, can offer an alternative method of investing in the AI ​​market apart from directly buying chip or hyperscaling companies, even if the full social benefits of the investment will take years. play. Ongoing collaboration between energy and AI companies will be crucial to support AI adoption and address upcoming challenges. Regardless of any near-term uncertainties, unabated energy demand remains far from being met, and investors should continue to monitor the potential opportunity space.