Far Point Global

Opening statements in the criminal prosecution of Huawei Technologies began this week in the Eastern District of New York, before Judge Ann Donnelly, eight years after charges were first brought. Twelve counts are before the jury — racketeering, bank fraud conducted through a Hong Kong entity called Skycom, export-control violations under IEEPA, and trade-secret misappropriation from five American technology companies. Huawei has pleaded not guilty. The trial is expected to run approximately three months.

The commentary has organised itself around the question of whether the government will secure a conviction. I would suggest that is the less interesting question, and that the risk in this proceeding is distributed almost entirely opposite to how it is being described.

Consider what a conviction would actually deliver. The policy objective this prosecution was constructed to serve — the removal of Huawei equipment from American and allied telecommunications infrastructure — was achieved years ago through entirely different instruments: Entity List designation, Federal Communications Commission equipment authorisation bans, and federally funded rip-and-replace programmes. Those are administrative and regulatory mechanisms. None of them required a criminal verdict, and none of them is contingent on one. A guilty verdict in December would not exclude Huawei from any market from which it is not already excluded.

Nor would it automatically trigger further consequence. A conviction does not place a company on the Specially Designated Nationals list, does not sever access to correspondent banking, and does not activate secondary sanctions. Each of those remains a discretionary executive-branch decision available today without a jury.

The financial remedy is similarly constrained. Statutory maximums permit $500,000 per count or twice the gross gain, whichever is greater, alongside criminal forfeiture. Analysts have estimated exposure in the hundreds of millions to billions of dollars, though no figure has been publicly specified by the government and quantifying benefit would itself require substantial post-verdict litigation. More materially: Huawei's assets are overwhelmingly located in China. There is no extradition mechanism for a corporate defendant and no significant domestic seizure target in a company already barred from the US market.

Set against that limited upside is a genuine and asymmetric downside. The Department of Justice's two most recent attempts to convict a large Chinese corporate defendant at trial both failed. Fujian Jinhua was acquitted outright by Judge Maxine Chesney in the Northern District of California in February 2024 — the flagship trade-secrets prosecution of the same China Initiative that produced this indictment. In July 2026, the government dismissed its economic-espionage case against the Pangang Group with prejudice, days into trial, ending a fifteen-year prosecution.

On Monday the US Treasury Secretary unveiled Operation Economic Outcast, a substantial expansion of the sanctions architecture around Iran. Most coverage has focused on the scale and the rhetoric. I'd point at two details in the announcement itself that are more informative than either.

The first is the sector list. The operation adds five newly sanctionable categories: digital assets, technology, gold, aviation and shipping. Consider what those five have in common. They are the settlement and transfer channels that exist precisely because conventional banking is unavailable. Gold and digital assets are how value moves when correspondent banking is closed; aviation and shipping are how physical goods move when documented trade is closed.

Extending sanctions to cover them is therefore not simply an escalation. It is an implicit acknowledgment that the existing dollar-based dragnet leaks, and that enforcement is now directed at the workarounds rather than at the primary target.

The second detail is the geography. The designated network is described as operating across the United Arab Emirates, Hong Kong, China, Singapore, Switzerland and Europe. These are not Iranian institutions. They are the intermediation layer of the global financial system — the jurisdictions through which most cross-border commerce is arranged.

That reframes what the policy is. Operation Economic Outcast functions less as an Iran policy than as a compliance obligation imposed on third countries and their financial institutions, whose practical demand is that they select a side. The Treasury Secretary said the campaign will not end until the regime stands alone, that he expects a major financial institution to be designated by the end of this week, and — asked specifically about Chinese banks financing Iranian crude — that no one is above the reach of US sanctions. He did not exclude them, which in this context is a substantive position rather than an evasion.

There is a sequencing question worth putting on your risk register. If a major bank designation lands this week, it arrives approximately eleven weeks before 10 November 2026 — the date on which Beijing's one-year suspension of its October 2025 rare-earth export-control package expires, and, on the same day, the date the US–China tariff truce lapses. A designation involving a Chinese institution in August would provide a well-formed rationale for Beijing's November decision. For any firm whose bill of materials includes rare-earth content, those two files are no longer separable.

One final observation. The campaign carries no deadline. That is unusual, and it matters: an open-ended pressure campaign cannot be judged to have failed, only to be continuing. Six months into a war, that is often what policy looks like when it is standing in for a settlement.

For clarity: no institution has been named and no nationality specified, and Beijing's November decision is not yet known. The point is the structure of the exposure, not a prediction.

On Monday the 30-year Treasury yield touched 5.31% — the highest since June 2007. On Wednesday the Treasury announced it will at least double the size of its long-end liquidity support buybacks, from $2 billion per operation to at least $4 billion, beginning September 9 and running through the November 4 refunding. Yields fell: the 30-year down 9 basis points, the 10-year down 6. On Thursday Scott Bessent told CNBC the number could go higher than $4 billion. Federal debt crossed $40.05 trillion the same week.

America Built China's Chip Industry (By Accident)

The most effective industrial policy Washington ran this decade is China's chip industry.

Bloomberg Businessweek just declared China's semiconductor breakout official: Huawei's AI chips are training LLMs and steering humanoid robots, and memory maker CXMT — profitable this year after a decade of losses — surged 466% after July's IPO to become China's most valuable listed company. In today's Daily Insight, Far Point Global breaks down why it happened: every breakout chipmaker in history needed a committed state plus an anchor customer, and US export controls delivered the ingredient Beijing's $150B could never buy. We cover the option nobody discusses (Nvidia's case for keeping China addicted to American silicon), why denial is a depreciating lever while dependence was a renewable one, FPG's call for the September summit, and the single tell to watch: CXMT shipping high-bandwidth memory at commercial scale.

Chapters (approximate): 0:00 The conclusion the market just reached 0:20 The scoreboard: CXMT +466%, Huawei's AI chips 0:50 The anchor-customer formula (TI, Intel, TSMC) 1:25 The wall became a matchmaker 1:45 The option Nvidia begged for 2:10 Denial vs. dependence — the unsettled bet 2:35 FPG's call: cash the card at the summit

Far Point Global's Daily Insight — one thesis a day at the intersection of markets and geopolitics. Subscribe for tomorrow's edition.

This content is for informational purposes only and is not investment advice.