Yazılar

Nvidia’s $100B OpenAI deal sparks antitrust scrutiny over AI dominance

/in /tarafından

Nvidia’s plan to invest up to $100 billion in OpenAI — while supplying the ChatGPT maker with millions of AI chips — is raising alarms among antitrust experts who warn the partnership could distort competition in a market already dominated by a handful of tech giants.

Nvidia controls more than half of the GPU market, the essential chips powering AI data centers. Experts caution that a financial tie to OpenAI could give Nvidia incentives to favor one customer over rivals through preferential pricing or faster delivery. “They’re financially interested in each other’s success. That creates an incentive for Nvidia to not sell chips to, or not sell chips on the same terms to, other competitors of OpenAI,” said Rebecca Haw Allensworth, a Vanderbilt Law School antitrust professor.

Andre Barlow, an antitrust lawyer, said the deal raises “significant antitrust concerns,” though the Trump administration’s pro-business stance complicates the outlook. President Donald Trump has emphasized both removing regulatory hurdles to accelerate AI growth and using antitrust enforcement to ensure long-term competition.

The scale of the deal highlights how expensive frontier AI has become. “The cost of chips, data centers and power has pushed the industry toward a handful of firms able to finance projects on that scale,” said Sarah Kreps, director of the Tech Policy Institute at Cornell University. Nvidia’s top two customers already account for nearly 40% of its revenue, underscoring its reliance on concentrated buyers.

Under President Biden, regulators had warned Big Tech could use scale to dominate AI. The DOJ and FTC pursued early inquiries into exclusionary conduct around AI resources. The Trump administration has kept many Big Tech cases alive, with DOJ antitrust head Gail Slater saying last week enforcement must focus on preventing bottlenecks: “The competitive dynamics of each layer of the AI stack and how they interrelate… are legitimate areas for antitrust inquiry.”

For now, Nvidia insists its investment won’t alter its sales practices: “We will continue to make every customer a top priority, with or without any equity stake,” a spokesperson said. OpenAI declined to comment.

CFPB ends Apple and U.S. Bank settlements early under Trump administration shift

/in /tarafından

The Consumer Financial Protection Bureau (CFPB) has ended oversight agreements with Apple and U.S. Bank years ahead of schedule, according to recent court filings. The move is part of President Donald Trump’s broader effort to reduce CFPB enforcement and roll back settlements imposed during the Biden administration.

Apple’s settlement stemmed from a 2024 CFPB action that found the company and Goldman Sachs violated consumer protection laws by mishandling disputes on the Apple Credit Card and misleading customers about interest-free transactions. The original agreement required five years of enhanced compliance, but has now been cut short. Apple paid a $25 million civil penalty, fulfilling its financial obligation.

U.S. Bank, meanwhile, faced a 2023 settlement over allegations it illegally blocked unemployed consumers from accessing pandemic-era benefits. The deal also required five years of compliance monitoring. The bank has since paid a $15 million penalty, made restitution payments, and pledged corrective measures, leading regulators to end oversight.

The filings also reveal the CFPB under Trump has dropped oversight for other firms, including Toyota and Bank of America, while halting nearly all enforcement actions still pending when Trump took office.

Critics say the changes mark a significant retreat from the agency’s consumer protection role, while supporters argue that excessive monitoring placed unnecessary burdens on businesses.

Meta’s Llama AI approved for use across U.S. government agencies

/in /tarafından

The U.S. General Services Administration (GSA) has approved Meta Platforms’ artificial intelligence system, Llama, for use by federal agencies, marking a milestone in the government’s adoption of commercial AI. The move comes as the Trump administration pushes to expand the integration of AI into federal operations.

Josh Gruenbaum, GSA’s procurement lead, confirmed that Llama will now be available as part of the agency’s catalog of authorized AI tools. Agencies can experiment with the model—free to use—with GSA’s assurance that it complies with legal and security requirements.

Llama, Meta’s large language model, can process multiple forms of data, including text, audio, video, and images. The approval gives government workers access to a tool that can assist in areas like accelerating contract reviews, handling IT troubleshooting, and managing vast amounts of information.

The GSA has also approved rival AI products from Amazon Web Services, Microsoft, Google, Anthropic, and OpenAI in recent months. Those firms agreed to offer their paid tools at steep discounts while meeting strict federal security standards.

Gruenbaum emphasized that the deals are not about political influence but about collaboration: “It’s about that recognition of how do we all lock in arms and make this country the best country it could possibly be.”

The approval reflects Washington’s accelerating interest in embedding AI across agencies to streamline operations, improve efficiency, and strengthen the government’s technological footing against global competitors.