Power-Hungry AI Data Centers: A Goldmine for Temporary, Mobile Power Solutions
Artificial Intelligence aspirations of major tech companies might boost shares in these energy sectors
The insatiable energy appetite of AI data centers in the U.S. could soon push the envelope, potentially surpassing the existing power capacity by an eye-watering 45 GW between 2025 and 2028, according to Morgan Stanley. To bridge this energy chasm, hyperscalers and data-center operators are tipped to slap together temporary, mobile power generation solutions.
This crunchtime calls for heroes, and a bunch of companies specializing in fuel cells, mobile natural gas turbines, and small modular nuclear reactors might just be America's saviors.
As the AI infrastructure expands, data centers are set to gobble up a whopping 65 gigawatts of power – a demand that's more Herculean than the existing system can manage. Morgan Stanley warns that all options – from fashioning crypto mining operations into data centers, constructing new power plants, and leveraging existing nuclear power plants – will need to be employed to keep the AI revolution alive.
But, it's not all smooth sailing. Sticker shock from bitcoin's price hikes might make miners hesitant to relinquish their mining facilities or sell excess power to data centers. Concerns about power grid stress might also prompt regulators to block new data centers from coming online until additional power sources join the party.
In the face of these challenges, Morgan Stanley foresees a middle ground: a bridge solution that deploys temporary, mobile generation to hurdle the regulatory and economic ruckus and quickly expand power capacity.
Embracing Nuclear and Natural Gas Generators
Small modular nuclear reactors represent a promising solution, as they offer the twin benefits of supply flexibility and emission-gutting, sweet music to Big Tech's emission-reduction ears. However, these cutting-edge reactors are still a work in progress, set to see the light of day sometime in the next decade.
Forward-thinking cloud hyperscalers like Microsoft, Alphabet, Amazon, and Meta Platforms have already started cozying up to existing nuclear infrastructure, leveraging its capacity during the past few years' AI buildout. Take Meta, for instance, which squared off with Constellation Energy, America's largest nuclear power provider, on Tuesday, signing a 20-year deal to power its AI operations.
With SMRs poised to make their debut in the next decade, interim data centers are likely to rely on small, agile natural gas generators from the likes of GE Vernova and Caterpillar. Fuel cell manufacturers like Bloom Energy also may find fertile ground, thanks to their low-carbon power servers that can convert natural gas or hydrogen into electricity. Analysts estimate Bloom Energy could swiftly ramp up manufacturing capacity to 3 GW per year, allowing the company to keep pace with the ever-growing demands of the global data center industry.
"We believe [Bloom Energy] could quickly increase manufacturing capacity to ~3 GW per year, with the potential for further increases in output if demand grows," the analysts wrote, acknowledging Bloom Energy as one of the undervalued beneficiaries of the AI data center power nightmare.
Worth a glance: Tipping off potential stakeholders that stand to gain from the AI infrastructure buildout's temporary power generation solutions: data centers and cloud providers, AI and technology companies, utilities and grid operators, investors, and environmental advocates.
- The financial sector, recognizing the increasing power demands of AI data centers, is exploring ways to fund temporary, mobile power solutions, such as fuel cells, mobile natural gas turbines, and small modular nuclear reactors.
- Cryptocurrency miners might serve a dual purpose in the near future, as their mining facilities could potentially be repurposed to supply power to data centers in an attempt to bridge the energy gap.
- Artificial Intelligence (AI) companies are collaborating with existing nuclear power providers to secure continuous power supply for their operations, given the emission-reduction advantages nuclear energy presents.
- As the demand for power surges due to AI infrastructure expansion, investing in companies that manufacture low-carbon power servers, like fuel cell manufacturers, offers a lucrative opportunity for investors looking to capitalize on the tech boom.
- The rising energy demands of the AI industry could spur advancements in technology, such as the increased production capacity of fuel cell manufacturers like Bloom Energy, which could help meet the ever-growing needs of data centers and the broader industry.
