Nvidia’s dominance: Stay alert to the risk of an AI monopoly
Summary
- Antitrust scrutiny of Big Tech has mostly been late, but AI chipmaker Nvidia is under the shadow of a US scanner already. The US Department of Justice is reportedly probing its practices. This field is too risk-laden to afford any regulatory failure.
For most of the Information Age, antitrust regulation has had the effect of bolting stable doors after the horses fled. Industrial Age models of judging abuse of monopoly in the US were to blame for it. Although digital platforms operated in winner-takes-all markets, they could not be hauled up for extortionist pricing, as many of their services were free.
Also read: Mint Quick Edit | AI chip-maker Nvidia’s share crash is a reality check
Yet, a late recognition of other forms of harm caused in the pursuit of profit, with the management’s interest found to lie in looking away, has forced a big rethink in America on the scope of its antitrust law. The power of Big Tech firms, in particular, has raised official eyebrows.
Given this policy shift in the world’s foremost free-market economy, it should not be a surprise if chip major Nvidia officially finds itself under scrutiny. Last week, its share price slid sharply on word of a likely probe by the US Justice Department. Is there a high-profile case brewing?
The value of Nvidia’s hold over the market for chips that run computers used for artificial intelligence, be it AI training or applications, is not in doubt. Its share is estimated at over 80%, which partly explains why its market valuation has soared amid a frenzy of investors betting on all things AI. Added up, all its shares were worth some $350 million at the start of 2023.
By mid-2024, that sum had shot up to about $3 trillion, briefly making Nvidia the world’s most valuable business, before AI stocks came off their highs. Now it is worth $2.5 trillion or so. While a market held under a company’s sway may well be a mark of its success, the misuse of this power is illegal, as judged by any harm caused to customers or competitors.
So, why should Nvidia be under watch? According to a Wall Street Journal report, the company has not yet got a subpoena (as misreported elsewhere) asking for its internal documents, but the US Justice Department has enquired about the terms of its contracts and partnerships. In August, Nvidia said various regulators had queried its operations.
Also read: Nvidia share price tanks over 9%, suffers record $279 billion rout in market value
How it allots its chips, for example. Its supplies have not been able to keep up with a surge in demand, leaving it in control of who gets what. The impact of this strength seems to be the focus of antitrust concern. Some rival chip-makers complain that buyers are reluctant to buy chips from them in fear of supply denial by Nvidia for disloyalty, but the company has denied that it insists on exclusivity.
Unless it’s proven that it uses its clout to punish clients and hurt rival players, it is not a market bully. To check, regulators may need to delve deeper. Separately, Nvidia’s move to buy Run:ai, a startup that offers software to optimize chip clusters, is getting a look-in for the extra heft it may acquire.
The biggest reason that Nvidia is best kept under an antitrust lens, however, has to do with the field that its chips underpin: AI. Here, large sunk investments have raised AI entry barriers, even as AI apps begin to exhibit ‘network effects’ that could unfold into a winner-takes-all scenario.
Instead of waiting for market concentration, regulators may be able to pre-empt it, and if a pre-emptive policy finds so many takers, it’s because AI is too powerful a technology to risk erring on. The fact that an AI takeover of the world is no longer deemed impossible should keep us alert.
Also read: Nvidia draws antitrust scrutiny as enforcers signal early interest in AI
That’s why AI software majors are under the scanner too. And if no AI app maker should get to wield outsize influence over us, let alone hold us in its thrall, the same applies to what underpins AI.