Published: June 23, 2025   |  Reading Time: ~10 min | Auther : Sanjeet kumar

Technology Decides Power: A Historical Perspective

In 1519, Spanish military commander Hernán Cortés arrived at the Aztec Empire with only about 500 soldiers. The Aztec Empire, on the other hand, was a vast and prosperous civilization with millions of people. Its capital, Tenochtitlan, was one of the largest cities in the world at the time. At first glance, Cortés and his small force seemed certain to lose. Yet he had something the Aztecs had never seen before — technology.

Cortés’s soldiers carried steel swords, wore armor, carried guns, and rode fast horses. The Aztec warriors were brave, but their weapons were made of stone and obsidian. When they first witnessed soldiers on horseback and fire-breathing guns, the technological gap became unmistakably clear. Spain won that battle, and eventually the larger war, because of its technological superiority.

Technology has always determined a nation’s power and position in the world. Israel’s victories against multiple Arab nations in 1948, 1967, and 1973 are also attributed to its technological and strategic edge. When the Cold War space race began between the US and the Soviet Union, people wondered what the purpose was — and then GPS technology emerged, transforming everyday life forever. Today, when you order food with a single click, you are using the fruit of that technological race.

In today’s world, this same dynamic plays out between the United States and China. The US leads, but China has accelerated so dramatically over the past decade that many experts believe it may one day surpass America. Anthropic CEO Dario Amodei recently stated that China’s AI models are only 6 to 12 months behind America’s most advanced models. And on June 23, 2025, China proved it in a concrete way: its supercomputer Lianshine claimed the number one spot on the Top500 ranking — the definitive global list of the world’s most powerful supercomputers.

What Is a Supercomputer and What Does It Do?

A supercomputer is not a magical machine. It is, at its core, a regular computer — but its processing power is millions or even billions of times greater than a standard laptop or desktop. A task that would take your personal computer months to complete, a supercomputer can finish in hours or even minutes.

A Practical Example: Cyclone Prediction

Imagine a cyclone forming in the Arabian Sea. Meteorologists need to know: Will it hit Gujarat? Will it strike the coast of Maharashtra? Or will it weaken and dissipate in the ocean? Answering this requires analyzing billions of data points — wind direction, sea temperature, atmospheric pressure, cloud formation, and how all of these will change over the next hours and days.

Assigning this task to a standard laptop could take months. A supercomputer, using thousands of processors working simultaneously, can produce the same result in hours. This is why cyclone warnings are now issued days in advance — saving thousands of lives.

Other Key Uses of Supercomputers

China’s new Lianshine supercomputer is already being used to build a digital model of the Earth — simultaneously simulating weather, oceans, and polar ice caps — and to construct complex models of the human brain.

China’s Supercomputer Journey: From 1978 to Today

The Starting Point: Deng Xiaoping’s Vision (1978)

The United States built its first supercomputer, the CDC 6600, in 1964 — widely considered the world’s first true supercomputer. By the time China’s leadership woke up to the strategic importance of supercomputing in 1978, America was already 14 to 17 years ahead.

In May 1978, Chinese leader Deng Xiaoping made a statement that would change the country’s technological trajectory:

“China cannot achieve the Four Modernizations without supercomputers.”

The Four Modernizations referred to Agriculture, Industry, National Defense, and Science & Technology. This declaration launched Project 785, assigned to the National University of Defense Technology (NUDT) and led by chief scientist Ci Yungui. The goal was to build a supercomputer capable of more than 100 million calculations per second.

Yinhe-1: China’s First Supercomputer (1983)

After approximately five years of effort, China’s first supercomputer was completed in December 1983. Named Yinhe-1 (Galaxy-1), it achieved a speed of over 100 megaflops — more than 100 million calculations per second. This milestone made China one of only a handful of nations capable of independently developing supercomputers.

Steady Growth Through the 1980s and 1990s

China continued building on this foundation. The Yinhe-2 project began in 1988 and by 1992 could perform approximately 1 billion calculations per second. Yinhe-3 followed, and with each generation, China’s supercomputing capability grew substantially.

The 21st Century: China Enters the Global Race

As China’s economy boomed in the 2000s, government investment in supercomputing surged. China was no longer trying to narrow the gap with America — it was actively competing to take the top position.

The target was clear: dethrone the world’s fastest supercomputer, which at that time was Jaguar, housed at the US Oak Ridge National Laboratory, with a speed of 1.75 petaflops.

To build something faster, China sourced Intel Xeon processors and NVIDIA Tesla GPUs. The result, completed in 2010, was Tianhe-1A. When tested, it achieved 2.57 petaflops — approximately 47% faster than Jaguar. For the first time in decades, the world’s fastest supercomputer was neither American nor Japanese. China was number one.

US Sanctions and China’s Unexpected Response

The Obama-Era Entity List (2015)

In 2013, China began upgrading to Tianhe-2, which became the world’s fastest supercomputer. This alarmed the United States. A country with the world’s fastest supercomputer could realistically simulate nuclear weapons, hypersonic missiles, and other advanced military systems — all without conducting physical tests.

In February 2015, the Obama administration placed four Chinese entities on its Entity List, effectively barring them from purchasing advanced US processors without a special (and likely unobtainable) license:

• National University of Defense Technology (NUDT)
• National Supercomputing Center Changsha
• National Supercomputer Center Guangzhou
• National Supercomputing Center Tianjin

China criticized the move, arguing that the rules were being changed precisely when China began successfully competing with the US. As Chris Miller documents in his book Chip War, the US viewed advanced supercomputers as a direct national security threat.

China’s Counter-Move: Domestic Chip Development

Rather than being stopped, the US sanctions accelerated China’s push for technological self-reliance. Within a few years, Chinese processors — notably the Feiteng and Shenwei series — developed rapidly. In 2016, China unveiled Sunway TaihuLight, a supercomputer built entirely with Chinese-made processors. It became the world’s fastest supercomputer without relying on a single American chip.

“The very sanctions meant to slow China down pushed it toward technological independence.”

Biden-Era GPU Restrictions

Under President Biden, the US dramatically escalated chip restrictions — specifically targeting NVIDIA’s advanced GPUs (Graphics Processing Units), which are the primary hardware used for both AI training and supercomputer acceleration. The US also pressured the Netherlands to prevent its company ASML — the sole producer of EUV lithography machines needed to manufacture advanced chips — from selling equipment to China. You can read more about ASML’s role in global chip manufacturing at ASML’s official technology page.

China’s response was characteristically pragmatic: it sourced second-hand lithography machines through gray markets, reverse-engineered components with the help of Chinese engineers who had worked at major US tech companies, and began building domestic prototypes. Chinese officials have stated that within one to two years, China expects to manufacture its own advanced chip-production machines — eliminating its dependence on foreign suppliers entirely.

The Lianshine Breakthrough: World’s Fastest Supercomputer (June 2025)

On June 23, 2025, China’s Lianshine supercomputer was officially declared the world’s fastest machine on the Top500 June 2025 list. This is the first time in approximately nine years that a Chinese system has held the number one position. The previous record holder, the US El Capitan at Lawrence Livermore National Laboratory, has been surpassed — Lianshine is estimated to be around 20% faster.

What Makes Lianshine Truly Special: The CPU-Only Architecture

Most modern supercomputers rely on a combination of CPUs (Central Processing Units) for general calculations and GPUs (Graphics Processing Units) for massively parallel tasks. NVIDIA’s GPUs are the global standard for high-performance computing and AI workloads. But NVIDIA’s most powerful chips are precisely what the US has banned from export to China.

China’s Lianshine engineers chose a completely different path. They built Lianshine entirely on CPUs — no GPUs at all. The machine contains approximately 14 million computing cores, housed across 90 large hardware cabinets, with chip designs developed entirely within China.

Current Applications of Lianshine

Scientists are already using Lianshine for:

• Digital Earth modeling — simultaneously simulating weather systems, ocean currents, and polar ice caps
• Human brain modeling — building complex neural simulations to advance neuroscience research
• Climate research — long-range climate change projections

What Does This Mean for the US-China Tech Race?

American experts are now openly debating what Lianshine reveals about the effectiveness of US tech sanctions. Several analysts, speaking to The New York Times, have suggested that China’s ability to build the world’s fastest supercomputer without advanced GPUs shows that it can continue to find workarounds to US export controls.

In the AI race, Anthropic CEO Dario Amodei has estimated that China’s AI capabilities are only 6 to 12 months behind the US — a gap that has been narrowing steadily. To understand how AI models are evaluated globally, you can refer to Chatbot Arena’s AI leaderboard, which tracks the performance of AI systems worldwide.

Meanwhile, the geopolitical dimension is sharpening. When Donald Trump visited China and met with Xi Jinping, the two sides reportedly reached an informal understanding: the US would ease some chip-related restrictions in exchange for China moderating its stance on Taiwan and providing access to rare earth minerals — critical materials that China dominates globally and that are essential for chip manufacturing. China had leverage precisely because of its technological progress. Nations that hold strong cards negotiate from strength.

The story of Lianshine is ultimately a story about strategic patience, self-reliance, and the unintended consequences of sanctions. The US sought to contain China’s technological rise. Instead, it pushed China to develop capabilities it might never have built otherwise. Whether this changes the balance of power in AI, defense, and computing over the next decade remains to be seen — but the race is far closer than it was even five years ago.