Aschenbrenner Shorts $8.46 Billion in Chips Before Huawei Drops the Tau Bomb

Leopold Aschenbrenner short semiconductor, Huawei Tau Scaling Law, Situational Awareness LP 13F, NVIDIA put options 2026, Huawei Ascend chip roadmap, Moore’s Law alternative China, SMH short position,

On May 18, 2026, Wall Street looked inside Leopold Aschenbrenner’s hedge fund. His firm had grown from $255 million to $13.7 billion in under two years. People naturally expected a highly optimistic bet on artificial intelligence infrastructure. They found the exact opposite. The 24-year-old former OpenAI researcher had built an $8.46 billion short position against semiconductor stocks. He targeted NVIDIA, Broadcom, AMD, TSMC, ASML, Micron, Oracle, Intel, and the wider sector.

Seven days later, Huawei stepped onto a stage in Shanghai. Their engineering team presented a technical roadmap that gave the Leopold Aschenbrenner short semiconductor trade serious scientific backing.

Decoding the $8.46 Billion Semiconductor Short

When the Situational Awareness LP 13F hit the public record, the numbers caught immediate attention. The fund bought $2.04 billion in put options against the VanEck Semiconductor ETF, creating a massive SMH short position. Individual targets followed closely. The data revealed a massive position in NVIDIA put options 2026, totaling $1.57 billion. Oracle took $1.07 billion in puts. Broadcom saw $1.01 billion, and AMD hit $969 million. TSMC drew $535 million. Micron and Intel received $584 million and $159 million respectively.

Put options made up 66% of the disclosed portfolio. At the exact same time, the fund bought stock in Bitcoin miners like CleanSpark, Riot Platforms, IREN, and Applied Digital. They also went long on data center infrastructure.

Traders call this approach infrastructure arbitrage. The theory suggests physical resources will dominate the next phase of the computing race. Power, land, cooling, and network connections win out over chip volume. The data center owners collect the real profits while hardware sellers peak early. A second argument hides inside the numbers. United States semiconductor dominance relies on a fragile foundation. On May 25, Huawei showed exactly how China plans to build parallel computing power.

Huawei Unveils the Tau Scaling Law

He Tingbo leads Huawei’s Scientist Committee and semiconductor division. She delivered a keynote at the IEEE International Symposium on Circuits and Systems in Shanghai. Her presentation introduced a Moore’s Law alternative China can control. She called it the Huawei Tau Scaling Law.

Traditional chip design follows Moore’s Law. Companies try to double the number of transistors in a given space roughly every two years. Shrinking transistors creates physical walls below 3 nanometers. Heat builds up fast. Quantum tunneling ruins efficiency. The extreme ultraviolet lithography machines required to print these chips cost billions. No single company has completely solved these barriers.

Huawei proposes a completely different path. The Tau framework measures time instead of physical size. Engineers focus on how fast signals travel through a chip and across a computing cluster. The goal is a shorter signal path rather than a smaller transistor. He Tingbo outlined the strategy directly. Her engineers will use math to fix physical limits. They will lean heavily on cluster computing to bypass single-chip boundaries.

The keynote outlined a real chip design breakthrough China is already utilizing. Huawei has mass-produced 381 chips using Tau principles over the last six years. A new Kirin smartphone chip arrives later in 2026. It features LogicFolding, a specific architecture that physically shortens wiring to cut latency. SMIC shares increased 7.6% after the news broke.

The engineering roadmap sets a clear target. Huawei expects chips built with the Tau framework to reach 1.4nm-class density by 2031. They plan to achieve this through architecture tweaks instead of manufacturing smaller nodes. China cannot buy extreme ultraviolet machines from ASML. The Tau system gives them a highly engineered way around the blockade.

How the Ascend Roadmap Validates the Bear Case

Market optimism for Western chipmakers depends on a massive lead. Building a top-tier AI model currently requires NVIDIA chips made by TSMC using ASML machines. The Tau Scaling Law attempts to dry up that moat.

Huawei plans to hit 1.4nm equivalent density without extreme ultraviolet machines. SMIC currently prints 7nm chips and works toward 5nm. If cluster routing and LogicFolding work as advertised, those older manufacturing methods become highly useful for AI training.

Western semiconductor companies trade at massive premiums. Investors assume these firms will hold an AI compute monopoly for years. China is actively engineering a detour.

The Huawei Ascend chip roadmap backs up this timeline. The Ascend 950 series rolls out in 2026. The 960 and 970 versions follow in 2027 and 2028. The DeepSeek V4 model already uses the Ascend 950 as its main compute engine. If the Tau architecture improves power efficiency, Western chipmakers lose access to a massive market. Those high earnings projections suddenly look out of reach.

The Geopolitical Shift in Computing Power

The Shanghai keynote and the hedge fund filing happened in the exact same news cycle. Beijing continues funding SMIC and expanding the Ascend program. The new design philosophy completely shifts global competition. China wants the world to know they do not need to match TSMC on every manufacturing node. They have a different map.

Investors holding semiconductor stocks at high multiples face a serious question. The geographic center of hardware supply is splitting. A massive fund placed billions on the belief that current stock prices ignore this shift. The Tau framework provides technical proof that hardware monopolies remain highly vulnerable to innovation.

Disclaimer:
All views expressed are my own and are provided solely for informational and educational purposes. This is not investment, legal, tax, or accounting advice, nor a recommendation to buy or sell any security. While I aim for accuracy, I cannot guarantee completeness or timeliness of information. The strategies and securities discussed may not suit every investor; past performance does not predict future results, and all investments carry risk, including loss of principal.

I may hold, or have held, positions in any mentioned securities. Opinions herein are subject to change without notice. This material reflects my personal views and does not represent those of any employer or affiliated organization. Please conduct your own research and consult a licensed professional before making any investment decisions.