From: allin

The Energy Race: US vs. China

The United States is currently in an “energy race” with China, a reframing of the “green energy race” by new Treasury Secretary Scott Besson [18:11:00]. This competition is crucial given the foundational role of electricity in powering modern advancements like AI and automation, which are critical factors for future economic growth [57:42:00].

Current Disadvantages and Challenges

The United States energy supply chain challenges currently faces significant disadvantages in this race:

  • Cost of Electricity Americans pay roughly 1.5 to 3 times more per kilowatt-hour for electricity than China [57:12:00].
  • Production Capacity The U.S. has approximately half of China’s electricity production capacity [57:25:00].
  • Cost to Add Capacity The cost to add new gigawatts of production capacity in China is significantly lower, ranging from 1/10th to 1/2 of the cost in the U.S. [57:30:00].

Historically, the U.S. disinvested in its power grid for a long period, especially since 2000, due to de-industrialization and the outsourcing of manufacturing, leading to a decreased need for additional electricity [1:26:00]. Conversely, China has ramped up its capacity significantly [1:27:10]. For example, since 1985, the net additions in electricity capacity in China are almost 7x that of the U.S. [1:27:17]. In 2000, both countries had roughly 1,000 terawatt-hours of capacity; today, the U.S. has 1,600 terawatt-hours, while China boasts 9,000 terawatt-hours [1:27:29].

The Role of Nuclear Energy

To catch up and ensure energy security, particularly for burgeoning AI infrastructure like the Stargate project (a new company intending to invest $500 billion over four years in AI infrastructure) [1:10:00], a substantial increase in nuclear energy is deemed essential [1:27:00].

Nuclear Capacity and Technology

The U.S. nuclear capacity has remained relatively flat for over 25 years [1:30:05]. In contrast, China has significantly increased its nuclear capacity over the same period [1:30:27]. France, for example, generates over 70% of its electricity from nuclear power [1:32:02].

New, safer nuclear technologies, such as Gen 4 reactors and pebble-bed reactors, are now available and in production [1:30:57]. These designs are touted as “meltdown-proof,” eliminating the risk of runaway heat and radioactive material release seen in older designs like Chernobyl [1:30:57]. China is rapidly deploying dozens or hundreds of these new systems, while the U.S. is deploying zero [1:31:28].

Policies and Solutions

The shift in energy policy under the new administration aims to accelerate power plant permitting on federal lands to meet AI data center demands [1:28:01]. This includes a broad approach to energy generation, rather than restrictive conditions based on source type [1:28:13].

The primary obstacle to increasing U.S. nuclear capacity is regulatory challenges and roadblocks [1:32:56]. Experts suggest that emergency actions may be necessary to remove these regulatory impediments, considering the issue a national security threat [1:33:35]. There is sufficient intellectual capital in the U.S. eager to advance nuclear buildout, but they are hindered by these regulations [1:33:16].

A potential downside to rapid deregulation and construction is the time required to build these facilities safely and properly, as the U.S. lacks recent practice in large-scale, accelerated construction of power plants [1:35:01].

Companies like Constellation Energy (CEG) are already signing direct deals with large tech companies (e.g., Amazon, Microsoft) and even federal agencies to provide direct energy access, indicating a growing acknowledgment that nuclear power is the best path forward for the country [1:28:56]. This trend is expected to continue [1:29:48].

The ability to build a robust power base is critical for national security and to ensure the U.S. remains competitive in areas like AI [1:36:11]. Without sufficient power capacity, China’s larger industrial AI capacity could eventually overshadow U.S. advantages in advanced GPU technology [1:35:54].