The relentless appetite for power in AI has prompted technology firms to seek new energy options — a quest that has sparked rivalry and funding in both fusion and fission ventures.

For many, natural gas presents the simplest solution for continuous, baseload power. It’s proven, affordable, and readily accessible. However, the conflict in the Middle East highlighted its precarious supply chain after Iranian drone attacks damaged a significant amount of natural gas infrastructure in Qatar, a key exporter. Concurrently, a spike in demand has created an extensive waitlist for gas turbines, indicating that current orders likely won’t be fulfilled until the early 2030s.

These delays not only threaten tech companies but also the natural gas sector itself. 

In the United States, 40% of the natural gas used currently is for electricity generation. By the time the turbine shortage resolves, the industry may be inundated with a new wave of competitors. Both small modular nuclear reactor (SMR) startups and fusion energy companies are set to begin connecting their initial commercial power facilities to the grid within the next five to seven years, roughly the same time needed to procure parts for a new natural gas power plant.

Nuclear challenge

SMR startups might have the best chance of overtaking natural gas power facilities. In many cases, the technology adjusts the designs of current fission reactors, but the basic physics has been validated and extensively utilized for decades.

Several SMR firms are seeking to have reactors operational before the decade concludes. Kairos Power, which lists Google as a potential customer, is among them. The firm has received green light for its Hermes 2 demonstration reactor in 2024, and construction is progressing well. Oklo, which merged with Sam Altman’s SPAC in 2024, aims for 2028 for its first commercial operations, as stated in its annual report.

Others aspire to follow suit a few years later. X-energy, an investor-backed firm from Amazon, is targeting the early 2030s, while TerraPower, founded by Bill Gates and with a partnership with Meta, is planning to launch commercial operations in 2030.

To surpass natural gas as the preferred generating source, SMRs must rapidly scale, realizing the economies of scale their business models rely upon. This will prove challenging. However, tech firms seem sufficiently assured as they invest in startups or formalize agreements with them for substantial power quantities. 

Fusion’s schedule

Another technology gaining traction among companies is fusion power. While not as established as fission, nuclear fusion holds the promise of generating vast amounts of energy with little more than seawater as fuel. 

Fusion startups also have their sights set on the early 2030s — or possibly sooner — to roll out their initial reactors.

One leading candidate, Commonwealth Fusion Systems, is poised to activate its demonstration reactor next year. Its inaugural commercial reactor, the 400-megawatt Arc, is anticipated to begin power generation in Virginia in the early 2030s.

Another newcomer, Inertia Enterprises, aims to start building a grid-scale power plant in 2030, basing its technology on the reactor design utilized by the National Ignition Facility, which was the first to validate that controlled nuclear fusion reactions could generate more energy than they consume.

However, Helion may possess the most ambitious timeline among all. Backed by Sam Altman, the startup is racing to construct Orion, its first commercial-scale power facility, by 2028 to provide Microsoft with electricity. Additionally, reports suggest the company is in discussions with OpenAI to deliver up to 5 gigawatts by 2030 and 50 gigawatts by 2035. To achieve these goals, Helion will need to build 800 reactors by the decade’s end and another 7,200 in the subsequent five years. 

If successful in delivering energy at those volumes, it could significantly alter the energy landscape. Last year, the U.S. added 63 gigawatts of new generating capacity across all sources. If Helion manages to construct nearly 10 gigawatts of new capacity annually, it would contribute more power than the entire natural gas sector achieved last year.

The cost dilemma

The primary hurdle for these companies — including gas turbine manufacturers — is cost. 

SMR startups are banking on mass production to lower costs, but this assumption remains unverified. Currently, nuclear power stands as one of the costliest forms of new generating capacity at approximately $170 per megawatt-hour, as per Lazard. Fusion faces a similar scaling challenge, albeit with even more uncertainties. Some experts estimate that one megawatt-hour from a fusion facility could initially cost around $150. 

In contrast, new baseload natural gas power plants operate at about $107 per megawatt-hour, according to Lazard, although prices have been climbing in recent years, potentially creating friction with both new fission and fusion reactors.

Nonetheless, they might all be underpriced by renewable energy paired with battery systems. 

The expenses for wind and solar energy have plummeted dramatically over the last ten years. Wind energy seems to have hit a bit of a standstill recently, but solar prices continue to decrease, showing no signs of halting. Batteries have also become more affordable over the years, reaching the point where grids are integrating substantial amounts of them — 58 gigawatt-hours last year. Even without subsidies, solar combined with batteries ranges from $50 to $130 per megawatt-hour, intersecting with fusion, fission, and natural gas. 

These figures are based on current battery technologies derived from chemistries developed for electric vehicles. Newer designs targeting grid connections could further drive down prices. For instance, Form Energy recently entered into an agreement to provide Google with power from a 30 gigawatt-hour iron-air battery. Another firm, XL Batteries, repurposes old oil tanks to store its low-cost organic fluid — the size of the battery is only confined by the dimensions and quantity of the tanks.

Because these innovative batteries avoid using critical minerals like lithium, cobalt, or nickel, they are expected to significantly lessen the expenses of long-duration energy storage to a point where justifying alternatives becomes challenging.