Huawei Stopped Whispering and Started Announcing

On September 18, 2025, Huawei Rotating Chairman Eric Xu stood in front of cameras at the company's annual Huawei Connect conference and laid out a detailed, public, multi-year battle plan to take Nvidia's crown in AI chips, according to Bloomberg.

Chinese tech giants rarely hand their competitors a written playbook. Xu did it anyway.

The SuperPod Play: Brute Force Over Brains

Xu revealed that Huawei's new SuperPod cluster architecture can link up to 15,488 Ascend neural processing units and run them as a single coherent system, according to Bloomberg.

Huawei isn't claiming its chips beat Nvidia's H100 or H200 in raw speed. Bloomberg reported directly that Huawei openly acknowledges its silicon can't match Nvidia chip-for-chip in raw power.

So the strategy is volume, networking, and policy. Flood the zone with enough interconnected chips that the cluster punches like a heavyweight even if each individual chip doesn't.

New generations of Ascend chips to power these SuperPod systems are scheduled to arrive starting next year, according to Bloomberg.

The Smartphone Front: 'LogicFolding' and a Big Claim

On the consumer side, Huawei's semiconductor division president Tingbo He unveiled a new chip engineering approach called "LogicFolding" at an industry conference in Shanghai on May 25, 2026, according to CNBC.

Using this stacking and folding design, Huawei says it can deliver chip density equivalent to 1.4-nanometer process technology by 2031. These chips will debut in Kirin smartphones this fall.

For context, TSMC — the world's leading chipmaker — has just begun volume production of 2-nanometer chips. If Huawei's projection is accurate, it would leapfrog the current frontier within five years.

Analysts Are Calling Out the Marketing

Not everyone believes it.

Paul Triolo, head of technology at DGA Group, was blunt to CNBC: "A stacked/folded design can produce effective density gains, but it does NOT mean Huawei has solved the full process, yield, power, thermal, and device-performance problems associated with true 1.4 nm-class manufacturing."

In plain English: density on paper doesn't equal performance in real-world conditions. Huawei is doing creative math. The 1.4-nanometer headline is marketing until proven otherwise.

Huawei is also seeking academic legitimacy for this work, dubbing it the "Law of Tau" or "τ scaling" — a would-be successor framework to Moore's Law. That's either genuine scientific contribution or branding exercise. Right now, it's unverified.

What Mainstream Coverage Is Getting Wrong

Most tech reporters are either dismissing Huawei's roadmap as propaganda or treating the 1.4 nm claim as a genuine benchmark achievement. Both approaches overlook the real story.

The hardware specs are secondary. Huawei's strategic posture has shifted. For years, operating under U.S. sanctions, the company improvised, substituted, and survived in silence. Now it's publishing timelines and holding press conferences. That signals a company that believes it's winning.

George Chen, partner at The Asia Group, told CNBC this trajectory "will likely heighten concerns in Washington, where Huawei remains emblematic of U.S. export restrictions."

Washington's export control strategy was built on the assumption that cutting off advanced chips would stall China's AI development. Huawei's September 2025 conference is the clearest sign yet that assumption is being stress-tested — and may be failing.

Nvidia's CEO Already Said It Out Loud

Last week, Nvidia CEO Jensen Huang told CNBC the company had "conceded" the Chinese market to Huawei. That's the CEO of the world's most valuable chip company saying he lost a market to a sanctioned competitor.

The Asia Group's George Chen put it plainly: "For Nvidia, this means the window to sell advanced chips such as the H200 into China is narrowing."

And Apple isn't safe either. Huawei's Mate 60, launched in 2023 with advanced 5G capabilities, already clawed back smartphone market share from Apple in China. Kirin chips running LogicFolding architecture this fall mean that fight isn't over.

What This Means for Regular Americans

If you own Nvidia stock, this is a direct threat to revenue in the world's second-largest economy — a market Jensen Huang just admitted he's already lost.

If you care about national security, a Chinese company publicly declaring a three-year timeline to dominate AI compute infrastructure is significant. AI chips aren't widgets. They run weapons systems, surveillance infrastructure, and military logistics.

And if you trust that U.S. export controls are actually working — Huawei just spent two days at a public conference explaining exactly how they're working around them.

Washington needs to answer whether the current sanctions strategy has any teeth left.