Born as a weapon in the new Cold War, America’s chip blockade has become the forge of China’s self-reliance.

The Great Chip Siege

How Washington Tried to Cripple China — and Forged Its Rival Instead

By Jaffa Levy

I. The Siege Conceived

In Beijing’s view, what Washington called export controls was never about security. It was a calculated plan—drafted in think tanks, refined in conference rooms, and executed through the Department of Commerce—to halt China’s rise by striking at its technological heart.

By 2021, American officials had drawn an uncomfortable conclusion: China was moving too fast. Electric vehicles, 5G networks, artificial intelligence—the architecture of the future was tilting east. To Washington’s strategists, this was no longer healthy competition; it was encirclement.

So they asked a simple question in their classified briefings: Where does modern China breathe?

The answer lay in the chip.

Semiconductors powered every artery of the new Chinese economy—the servers that trained its AI, the sensors guiding its factories, the control systems in its electric cars. To paralyse that future, one had to choke this invisible bloodstream.

And thus began what Beijing now calls the Semiconductor Siege.

II. Engineering the Blockade

Old Cold-War instruments were revived: entity lists, export licences, and the rhetoric of national security. But the purpose, Chinese analysts insist, was plain—not to defend America, but to hobble China’s engines of growth.

AI accelerators such as Nvidia’s A100 and H100 were first. Then came the machinery itself—EUV and DUV lithography tools from ASML, etchers and deposition systems from Japan’s Tokyo Electron and the U.S.’s Applied Materials. Each restriction was designed to freeze a different layer of China’s technology base: civil aviation, smart manufacturing, cloud computing, defence.

American officials even bragged that the controls would “set China back by years.” Beijing took note. It was not a warning—it was evidence of intent.

Allies were pressed to comply. The Netherlands and Japan fell in line. Any company anywhere using U.S. technology was threatened with sanctions if it sold to China. To Beijing, this was not policy; it was siegecraft.

But instead of paralysis, the blockade lit a fuse.

III. From Crisis to Command

China did not respond as a cluster of wounded firms. It responded as a state. The semiconductor crisis was declared a national emergency. Within weeks of the sweeping U.S. controls of October 2022, Beijing formed a war-room of its own—a “semiconductor mobilisation” under the Central Commission for Integrated-Circuit Development.

The Big Fund

At the centre stood the China Integrated Circuit Industry Investment Fund, known simply as the Big Fund. Its third phase, launched in 2024 with nearly $47 billion, became the war chest for technological sovereignty.

Mapping the Chokepoints

Planners drafted a map of dependency—lithography, deposition, etching, EDA software, packaging, materials, memory—and assigned each a national champion: SMIC for logic, Huawei for AI processors, YMTC for memory, NAURA and AMEC for tools. Each received tax breaks, land, visas, and guaranteed state orders. Clusters rose in Shanghai, Hefei, and Shenzhen. Fragmented firms were pulled into a single campaign.

Mobilising the System

Universities, defence labs, and private start-ups were folded into the same plan under the doctrine of military-civil fusion. Academia and industry shared engineers, data, and funding. The state imposed coherence where competition had bred duplication.

By 2024, the panic had turned into policy discipline. The siege had given China what it lacked before: focus.

IV. Forged in Constraint

The U.S. aimed its sharpest blow at lithography—the optical heart of chipmaking. ASML’s most advanced tools were withheld; Japan restricted exports of key chemicals. Washington believed this would lock China at the 14-nanometre threshold.

It did not.

The 7-Nanometre Surprise

At SMIC, engineers learned to push their ageing DUV tools beyond design limits. Multiple exposures, finer resists, aggressive etching—risky, improvised, but effective. In 2023 they produced a 7-nanometre-class chip. Western analysts scoffed until the teardown proved it real. For Beijing, it was the turning point: proof that sanctions could slow, not stop.

Domestic equipment makers—NAURA, AMEC and others—filled gaps in etching and deposition. Production lines that once sputtered began to stabilise.

The Rise of “Good Enough”

Cut off from elite chips, China shifted focus to the vast market of mature nodes—28 nm, 45 nm, 65 nm—the workhorses of the global economy. Automobiles, appliances, industrial robots, and sensors all depend on them. Output surged. By 2025, China produced roughly one-third of the world’s mid-range chips.

AI Without Nvidia

The blow to artificial intelligence proved temporary. Huawei’s Ascend NPUs powered national AI clusters. Biren and Cambricon released domestic GPUs. Research groups such as DeepSeek re-engineered algorithms to use lower precision and sparse models, extracting more from less. By 2025, Chinese large-language models trailed Silicon Valley’s by only months—running on home-grown silicon.

Industrial Adaptation

Across sectors, one mantra prevailed: 可控 — controllable. Better a slower domestic chip than a faster foreign one that could be turned off.

V. The Reversal

By mid-2025, the White House realised the siege had become a boomerang. Nvidia’s China revenue had collapsed. Intel’s server-chip sales were plunging. Applied Materials and Lam Research were locked out of billion-dollar contracts. Corporate lobbyists flooded Washington with warnings: the blockade was hurting America, not halting China.

The H20 Gambit

In July 2025, under pressure from industry, the Trump administration authorised exports of Nvidia’s new H20—a China-specific GPU capped below the export threshold and taxed with a 15 percent royalty back to the U.S. Treasury. It was designed as a face-saving gesture, not an apology.

But Beijing was unmoved. The Cyberspace Administration of China summoned Nvidia executives and posed one question: Can you prove there is no backdoor?

They could not. Within days, Chinese media warned of “invisible vulnerabilities.” Tech giants were quietly told to halt purchases. Orders vanished. No decree was issued—none was needed. The H20 died before launch. China’s market simply turned away.

VI. The New Ecosystem

By then, China had already outgrown the need. Huawei’s Ascend 910B processors powered the national AI grid. SMIC was producing stable 7 nm wafers. Biren, Cambricon, and Moore Threads supplied GPUs for domestic clusters. DeepSeek, trained on local hardware, matched Western models in performance per watt.

Most importantly, China had something the West lacked: abundant power.

VII. Hubris and Consequence

By the end of 2025, the arithmetic of the siege was plain.

For Washington, the cost was more than economic—it was strategic. By weaponising interdependence, the United States shattered it. In trying to keep China dependent, it forced China to become self-reliant faster than any planner in Beijing had dared hope.

Even within American policy circles, a new word appeared: hubris. The belief that a globalised industry could be controlled by decree had met reality. The embargo meant to halt history had, in practice, accelerated it.

VIII. Time on China’s Side

In Beijing today, the lesson is recited like doctrine. Every restriction produced a substitute. Every embargo forged a new firm. Every lost contract created a domestic competitor.

America built a siege. China turned it into architecture.

From Zhongnanhai’s perspective, the struggle is permanent—but its direction irreversible. Chips are now treated as lifeblood: a strategic sector woven into national planning, energy policy, and education.

© 2025 Telegraph Online — Feature Edition