Friday, September 13, 2024

The Misguided Antitrust Investigations in AI

By Jennifer Huddleston of Cato.

"In early September, news of possible Department of Justice (DoJ) subpoenas indicated that the agency is escalating its investigation into antitrust claims against chipmaker Nvidia. Some critics, both left and right, have expressed concerns that there is already a “big tech” monopoly on artificial intelligence (AI). But as I’ve previously written, the criticism does not reflect the reality of the tech market or, more narrowly, the AI market in general.

Given these latest actions, an analysis of why these claims disregard the consumer welfare focus of antitrust law and the dynamics of an emerging market, such as AI, is needed.

Antitrust Turns its Investigations Towards AI

The DoJ alleged in its probe that Nvidia makes it harder to switch to other suppliers and penalizes buyers that don’t exclusively use its artificial intelligence chips. The investigation also focuses on the company’s acquisition of AI workload management firm RunAI for $700 million in April.

In response, Nvidia said in an emailed statement that the company “wins on merit, as reflected in our benchmark results and value to customers, who can choose whatever solution is best for them.” Also, a spokesperson encouraged regulators to approach with questions about their business practices. This comes after the Federal Trade Commission (FTC) announced in January the launch of an inquiry requiring six AI companies to provide information regarding “investments and partnerships involving generative AI companies and major cloud service providers.”

These investigations follow the antitrust cases brought against leading tech companies, including Google, Meta, Apple, and Amazon, indicating a heightened degree of scrutiny and enforcement in the technology sector. Additionally, US competition regulators like the FTC have been increasingly aggressive in merger and acquisition enforcement, both in the tech sector and more generally.

Despite a mounting record of losses, many of these cases shy away from the objective, economic-focused consumer welfare standard in favor of theories that are far more attenuated about potential harms and focus on competitors — not consumers.

Competition in AI Is Global

American companies were leaders during the internet era, in part due to our country’s light-touch approach to regulation. By allowing entrepreneurs and consumers to determine the best applications for this technology, both success and failure were determined by the market — not government bureaucrats. America’s leading tech companies became global household names.

More regulatory approaches, like that of the EU, made it difficult for innovative entrepreneurs to develop and launch products in those countries. However, one other side effect of the internet era is the global connectedness of people and markets. As my Cato colleague Scott Lincicome often points out, increased globalization has practical and moral benefits for societies.

Much like the internet and other communications technologies, AI is accelerating our ability to access information. Global competition for various elements of the AI stack is still developing. AI competition is diverse as many elements go into creating an AI model; however, competition exists throughout the various elements of AI development as well as the final AI products.

For example, while the US currently leads in much of the research and development of AI (or has the potential to do so), these companies face global competition. This includes UAE’s Falcon 2 open source model that has emerged as a significant competitor. While Europe has taken a more regulatory approach with the AI Act, there is still an unusual amount of activity around AI start-ups — particularly in France and the UK — that could prove competitive to US companies.

To look at the AI market only in terms of the United States misses much of the global nature of the current stage of development and competition where it is still unclear what the winning models, chips, and applications will be.

Competition Continues to Develop

Even if one were to look more narrowly at AI chips as one smaller element of a complicated market, there are competitors both internationally and domestically for Nvidia.

In many ways, the AI market is still in its early stages. Neither regulators nor the public should presume that the first mover or early successful companies will forever remain dominant. While Nvidia may currently be the leader in providing certain elements in the AI infrastructure, a growing number of competitors in the US are either offering products or looking to develop alternatives. In June 2024, CNBC reported on the rising competition in the AI chip market. This included AMD and Intel creating alternatives as well as companies like Amazon looking at if they could develop their own processors for their cloud and AI products.

If Nvidia wants to maintain its success in the market, it will have to do so by continuing to offer a superior product and adapting to demand. A first mover of a popular and successful product will face challenges from other innovators who see opportunities in the booming market, even if it momentarily appears to have few competitors. However, if it maintains that success through superior products, the response to consumer needs should be applauded and not presumed suspect.

The competition in the AI space is global and regulation or unnecessary enforcement might hamper America’s leading companies in the global market. Nvidia identified Chinese company Huawei as a competitor in 2024 filings. This raises the question of what alternatives allies might turn to if they seek to continue their AI development when an American company is bogged down in negotiations with regulators. Given the concerns expressed by many policymakers about Chinese technology, deterring an American company from further development and innovation by presuming its success is anticompetitive could create greater opportunities for its Chinese competitors.

Conclusion

The escalating investigations into some of America’s leading AI companies miss the competition at all levels of the AI ecosystem, as well as the global market these companies compete in. The eagerness of antitrust enforcement in such an emerging market is just the latest example of the “big is bad” mentality that drives much of the US’ current approach to competition policy. If antitrust loses its objective standard, it is consumers who will lose out on the most innovative products and the best service options in the market."

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.