In 2026, during the Spring Festival, while the world was still awestruck by OpenAI’s latest model parameters, China showcased another face of AI—embodied intelligence made real through a Spring Festival Gala.
Looking at the 2026 CCTV Spring Festival Gala program, we witness an unprecedented "AI parade"—far beyond the simple robotic dance performances of a few years ago, this marks a concentrated breakthrough in China’s robotics industry, showcasing multiple companies, multiple robot models, and full-scenario applications.
- Magic Atom’s full-stack cluster enabled robots to become the ultimate “atmosphere group,” dancing alongside Chen Xiaochun and Yi Yangqianxi in “Intelligent Future,” with movements so coordinated they’re indistinguishable from reality.
- Unitree's G1 and H2 robots demonstrated astonishing motion control capabilities in "Wu BOT"—autonomously maintaining balance through on-device computing, without real-time remote control. When the H2 danced with a sword draped in a red robe, it proved that China’s robotic motor coordination has matured.
- In the skit "Grandma's Favorite," Song Yan Power enabled the robot to take on the comedic roles of delivering punchlines and responding to jokes, successfully transitioning from a "prop" to an "actor."
- Galbot G1 by Galaxy General performs the "cracking walnuts" maneuver in the short film, a seemingly simple action that showcases the pinnacle of dexterous hands and haptic feedback technology.
This Spring Festival Gala sent a clear message: China's AI is no longer confined to servers—it has grown limbs and stepped into the real world.
Yet, even as we cheered for the robots, a silent panic gripped Wall Street across the ocean—they realized that the “blood” powering these AIs—electricity—was running out. When we shift our gaze from the Spring Festival Gala stage to the data centers of Silicon Valley, we confront the elephant in the room: electricity.
By early 2026, electricity prices for U.S. residents had surged 36% to $0.18 per kWh. But this is merely surface-level; the core crisis lies in the collapse of supply. Training a GPT-4-level model consumes as much electricity as the combined annual usage of 100,000 households. By 2028, annual electricity consumption by U.S. data centers is projected to skyrocket to 600,000 GWh.
The U.S. power grid is facing a dual crisis—like a heart attack and a blood clot—with 5% of its electricity relying on aging fossil fuel and nuclear plants that are now undergoing widespread retirement. The grid is fragmented into three isolated regions: the Eastern Interconnection, the Western Interconnection, and Texas’ ERCOT, with extremely poor interconnectivity. Approving a single interstate transmission line can take up to 15 years, preventing wind power from the Midwest from reaching data centers on the East Coast.
As Sam Altman said, “Energy is currency.” In today’s Silicon Valley, the question troubling CEOs is no longer about chip allocations—but where there’s enough electricity to power these chips.
If computing power is the engine of AI, then electricity is its fuel. In this energy competition, China has built a strategic moat that the United States cannot replicate, thanks to a decade-long head start.
By 2025, China had built 45 ultra-high-voltage (UHV) projects, with the total length of UHV direct current transmission lines exceeding 40,000 kilometers. This "electricity superhighway" can transmit abundant clean energy from the west to data centers in the east at millisecond speeds, or directly support the "East Data, West Computing" hubs. China possesses 35 of the world’s 37 largest high-voltage direct current cable systems—a generational infrastructure advantage that the United States cannot bridge in the short term.
The high energy consumption of AI inherently demands clean energy. In 2025, China’s share of renewable energy capacity surpassed 60% for the first time in history, with new wind and solar installations exceeding 430 million kilowatts. Nearly 4 out of every 10 kilowatt-hours consumed nationwide came from green sources. While the U.S. continues to grapple with nuclear power plant delays, China has already achieved grid parity for solar and wind power, providing AI data centers with affordable and sustainable energy solutions.
China is the world’s center for transformer manufacturing, accounting for over 60% of global production. Meanwhile, the biggest challenge facing U.S. grid upgrades is transformer shortages, with delivery times now stretching to three to four years. Whether through transshipment via Mexico or direct procurement, the U.S. grid’s operation heavily relies on Chinese-made transformers. While U.S. data centers halt operations due to transformer shortages, Chinese power equipment manufacturers are operating at full capacity, supporting the rapid expansion of domestic computing infrastructure.
The 2026 Spring Festival Gala is not just a celebration of robots, but also a snapshot of China's industrial strength.
When we see Unitree’s robot dog rolling or Galaxia General’s robots working on screen, don’t forget: behind every fluid movement lies not only advanced algorithms, but also stable current transmitted over thousands of kilometers via ultra-high-voltage lines, supported by a powerful power grid.
In the second half of the AI revolution, the marginal cost of computing power will no longer depend on the nanometer scale of chips, but on the cost of acquiring joules. The United States possesses the most advanced algorithm design, while China has the most powerful energy conversion and transmission systems.
For investors, the logic is clear: in this gold rush, if NVIDIA is selling shovels, then China’s infrastructure builders (ultra-high-voltage, power equipment, green energy) control the actual water source.