If you were looking for a single barometer of the state of the world – or of a particular copper-rich sector – it would be hard to go past the copper price. Since late October, it has slumped 7.3 per cent.
Much of the bullishness that pushed the price to record levels in September related to the extraordinary sums being spent on artificial intelligence and the infrastructure to host it.
While the initial sell-off from that peak related to Donald Trump’s tariffs and the deferment of the foreshadowed tariff on US copper imports that copper bulls had anticipated, the rest of the decline would appear to be linked heavily to a shift in perceptions about the future of AI and, most recently and specifically, in perceptions of the future of Nvidia, the company at the epicentre of the AI boom and its financing.
On October 28, the New York Times first reported the relationship between Nvidia’s chief executive, Jensen Huang, and a suspected Chinese spy, Alice Huang (no relation). The copper price fell 2.84 per cent.
On November 3, in response to the “blue tsunami” in the midterms by Democrats who made opposition to the massive buildout of data centres a central issue in their campaigns, it fell another 2.8 per cent.
On November 8, when Jensen Huang admitted to the affair, the price fell 1.7 per cent. Last week, when he resigned/was forced out by his board it was down another 1.8 per cent.
From its peak, copper is now down more than 10 per cent. If “Doctor Copper,” as the metal is often labelled, has diagnosed the patient, AI has some significant health problems.
Nvidia itself has lost nearly $US2 trillion ($2.9 trillion), or about a third of its market capitalisation, since its October peak. The recent falls illustrate how central Huang is seen to be to the company, in which he has no obvious successor, and to the sector, where his personal indiscretions overlay mounting concerns about the ever-increasing cost of the AI build and the ways bin which it is being financed.
The concerns about Nvidia are two-fold. One is the view that Huang is irreplaceable, which explains why there is lobbying for his reinstatement. The other is that the model he built, where Nvidia essentially has been funding demand for its chips and underwriting a significant slice of the industry, will implode if he isn’t at the company’s helm.
Nvidia’s chief financial officer, Colette Kress, who is now acting as CEO, is known to be sceptical about the ways in which Nvidia has been providing neoclouds and other data centre developers with chips, finance and guarantees that they couldn’t access elsewhere.
Nvidia has purchasing commitments, financial and contractual obligations, including contingent guarantees, of more than $US530 billion.
The $US279 billion of purchasing commitments to its own suppliers are one side of a coin, with the various other real and contingent liabilities supporting the appetite of the smaller AI labs and the construction of the data centres that will provide the demand for its chips representing the coin’s other face.
There is a very real question raised by Huang’s exit from Nvidia as to whether a different leadership might pull the rug from under the weaker – the non-hyperscaler – players in the sector.
Combined with the likely slower roll-out of new data centres – only those already built or approved and under construction are likely to be unaffected by the Democrat landslide at both the federal and state government levels – the turmoil within Nvidia has the potential to blow up a substantial slab of the sector and Nvidia’s revenues and earnings in the process.
One certainty emanating from the controversies around Huang is that the US Congress will tighten up restrictions on the sale of America’s most advanced chips (which means Nvidia’s Blackwell chips and their successors) to China.
It’s not that Nvidia sells many of its chips directly to China. The Biden administration significantly tightened the restrictions and, while Donald Trump approved the sale of Nvidia’s second-tier H200 chips to China (with a 25 per cent commission for his Treasury), the official sales are modest.
However – and this is where Alice Huang entered the picture – a host of data centres have sprung up in South-East Asia, hosting the Blackwell chips and renting their capacity to Chinese companies to train their AI models or, it is suspected, simply smuggling those chips to China.
Alice Huang helped set up a company, Megaspeed, in Singapore in 2023. Between early 2024 and June 2025, according to the Wall Street Journal, Megaspeed smuggled nearly $US12 billion of Nvidia’s most advanced chips to China.
There’s the problem for Jensen Huang and those advocating his return. How could he not be aware that his mistress’ company was buying the sort of volumes that could only be explained if there were Chinese end-buyers?
Nvidia claims to be able to track its chips. It should have (must have?) known where they were ending up.
Jensen Huang is inevitably going to be grilled by congressional committees about his knowledge and actions.
Congress has before it three bi-partisan bills forcing location verification and reporting for advanced chips, a prohibition on the sale of chip-making equipment to China and whistleblower protections for those reporting suspected export control violations.
Those bills, and the more intense scrutiny of where its chips end up, appear a threat to the Nvidia business model.
Nvidia’s $US279 billion of purchasing commitments represent agreements to buy chips from its manufacturer, Taiwan’s TSMC, and other components and services from third parties.
It needed a market for all the chips it has in the pipeline, which explains why Huang has been investing in, supplying finance to, guaranteeing data centre finance and capacity and providing guarantees that it will buy back unused chips from its customers.
The need to help his customer base expand has been accentuated by Nvidia’s reliance on OpenAI and Anthropic and the hyperscalers.
With Google, Amazon, Meta and Microsoft developing their own chips, the frontier labs rapidly diversifying their supply bases and competitors like AMD, Broadcom and a hoist of start-ups closing the gap on the performance of its chips, Nvidia – with a market share of around 90 per cent – has a lot to lose.
Should its “circular finance” model start to unwind, there’d be neocloud and data centre builders in immediate strife and Nvidia would be left with a lot of chips and a big hole in the bullish assumptions of its continued revenue and earnings growth.
The fallout from Huang’s sexual adventures – there are also reports of affairs with at least one, and possibly more, Nvidia executives – is not just a threat to the company and the segment of the AI sector that is fed by it. It’s also a threat to the White House and China.
The White House is caught up in it because, Trump, despite being warned of the relationship between the two Huang and the possibility that Alice Huang was a Chinese spy, is close to Huang.
Huang – a donor of $US25 million o Trump’s ballroom – has played an influential role in Trump’s decision to leave the US AI sector largely unregulated. Trump has played down the seriousness of the scandal and been very supportive of Huang as it has developed.
There’s an internecine battle developing within the administration, with the Treasury Secretary, Scott Bessent, blaming the Commerce Department, and its secretary, Howard Lutnick, for resisting calls (after Alice Huang was identified as a potential Chinese agent) for Megaspeed to be placed on America’s “entity list,” which would have prevented Nvidia from dealing with it.
There’s also the potential for a new confrontation with China. If the US does enact new and tougher legislation that chokes off any material volume of its most advanced chips being used to train Chinese models – or being acquired by China – China has vowed to retaliate.
It does have its dominance of rare earths, which it has successfully used as leverage in the trade war with the US to force Trump to back down, with which to buy time while its own AI sector intensifies its efforts to develop chips that equal or better the performance of Nvidia’s.
There are a lot of significant and diverse implications that flow from the controversies that have enveloped Nvidia. The easiest way to monitor the net sum of their potential impacts on the near term future of AI is to observe Dr Copper’s diagnosis. At present, it is ominous.
The Market Recap newsletter is a wrap of the day’s trading. Get it each weekday afternoon.
From our partners
Disclaimer : This story is auto aggregated by a computer programme and has not been created or edited by DOWNTHENEWS. Publisher: www.smh.com.au






