What's up: How Iron Man’s Jarvis became the symbol of corporate America’s AI ambitions; Chinese hackers used Anthropic’s AI; Amazon, Microsoft back effort that would restrict Nvidia’s exports China exports.

Good morning. Overall corporate demand for AI is strong, but there’s at least one layer that’s driven by impulse. And that layer, difficult to quantify, might not be as sustainable as other streams of spending.

Here's how this market psychology might be at work, contributing to the pressure on tech valuations. A certain amount of corporate spending on AI companies might not be carefully thought through. It’s more performative, a demonstration that leadership is doing something. As that spending flows into AI public and private AI companies, it’s booked as revenue, where investors have a tendency to overvalue it.

This dynamic may have contributed to the froth in AI valuations, which is now being blown off. There are other factors, too. As the WSJ’s Christomer Mims and Nate Rattner write today, record capital expenditures and data-center planning are running up against the ground truths of physical infrastructure.

A correction doesn’t necessarily amount to catastrophe, either. As Richard Waters writes in the FT, “even if valuations are stretched, it doesn’t mean AI investment is facing the kind of general implosion that happens when bubbles collapse in on themselves.”

Where do AI valuations go from here? One catalyst is expected Wednesday when Nvidia is scheduled to release its next earnings report.

Relief over the end of the government shutdown has given way to concerns about lofty tech valuations and whether the Federal Reserve will slow its interest-rate cuts.

This week's tech selloff looks poised to continue early Friday, WSJ reports, after U.S. indexes posted their largest declines in a month. Tesla shares dropped 3% in premarket trading, while chip stocks Nvidia, Micron, Intel and AMD all fell at least 1%. 

Tech executives (who, frequently, are also sci-fi and fantasy stans) are leaning on one particular pop culture reference when describing their AI. And no, it's not Skynet.

WSJ Leadership Institute reporter Isabelle Bousquette keeps hearing one reference talking with tech leaders about their latest AI endeavor: Jarvis, or rather, J.A.R.V.I.S., which stands for Just Another Very Intelligent System.

Jarvis is the fictional AI from the Marvel films known for helping playboy turned superhero Tony Stark. The AI is responsible for everything from managing Stark’s household and business to helping him design and operate his Iron Man suit.

Oh yeah, and Jarvis helps save the world. What tech executive or entrepreneur wouldn’t want to align their AI agent with Jarvis?

Jarvis ticks all the boxes, Isabelle writes.

“One of the things Jarvis does is he puts a really nice face on agentic systems,” said Julian Chambliss, an English professor at Michigan State University, and the Val Berryman Curator of History at the MSU Museum.

Jarvis is moral and cognitively aware and dispels our anxieties that the technology could replace or even harm humans, said Chambliss, whose work explores the relationship between comics and culture.

“I use it a lot,” Robert Blumofe, chief technology officer at Akamai Technologies, tells Isabelle. “If you think about [artificial general intelligence] and what we’re trying to do, I do think that Jarvis is sort of the best version of that that we have.” Read the story.

AI can be used to supercharge cybersecurity defenses. But stronger AI systems also make for stronger attackers.

China’s state-sponsored hackers used Anthropic’s AI to automate 80% to 90% of a September hacking campaign targeting corporations and governments, the company said Thursday. Anthropic disrupted the campaigns and blocked the hackers’ accounts, but not before as many as four intrusions were successful, WSJ reports.

How the hackers did it. Hackers often use open-source AI tools which can be modified to remove restrictions against malicious activity. To use Anthropic’s Claude model, the China-linked hackers had to jailbreak its safeguards—in this case, telling Claude that they were conducting security audits on behalf of the targets.

Anthropic says that after the attacks, it updated the methods it uses to detect misuse.

More to come. Last week, Google reported that hackers linked to the Russian government attacked Ukraine using an AI model to generate customized malware instructions in real time.

A split between Amazon and Microsoft and their key supplier Nvidia underscores just how fierce the AI race has become.

Both Amazon.com and Microsoft support the Gain AI Act, legislation that requires chip firms to satisfy U.S. demand before sending products to China and other countries subject to arms embargoes, WSJ reports.

Opposing the measure is Nvidia, which controls roughly 80% of the market for AI processors, and calls the act an unnecessary intervention that could open the door to more export restrictions.

Nvidia has bulked up its lobbying. Nvidia Chief Executive Jensen Huang frequently talks to Trump about AI policy and the company has spent nearly $3.5 million on lobbying in the first three quarters of the year, up from $640,000 in all of 2024, according to data provider OpenSecrets.

Chinese tech giants continue to ramp up their chip-making efforts amid U.S. export restrictions on advanced semiconductors.

Baidu this week unveiled two AI chips, the M100 chip for large-scale inferencing and the M300 for model training. The chips are slated to launch in 2026 and 2027, WSJ reports.

Verizon joins a wave of companies slimming down, with some finding tech-driven ways of improving efficiencies. Corporate giants including Amazon.com, United Parcel Service and Target have all announced job cuts in recent weeks.

Verizon Communications is planning to cut roughly 15,000 jobs, the largest ever for the telecom, WSJ reports. The company had about 100,000 employees as of February, according to securities filings.

New CEO breaks the news. Last month, Verizon named its lead independent director Daniel Schulman as its new chief executive officer. Schulman, a former CEO of PayPal and Virgin Mobile USA, has said he would aggressively reduce the company’s entire cost base and take steps to reverse the customer losses.

Speaking at the WSJ Leadership Institute’s inaugural Board of Directors Council Summit this week, Schulman talked about how big organizations like Verizon need to start thinking of themselves as a startup. “We're going to go into a world of change right now, and we're going to have new competition coming who are not afraid,” he said. “The organization needs to become less bureaucratic and needs to be more nimble ... It needs to think of itself as a startup with a ton of great assets to leverage.”

Employers have a warning for the Class of 2026: Next spring’s graduate-hiring market is likely to be even worse than this year’s.

Come next Spring, graduates will not only be competing against each other for jobs, but also against those junior workers who have been recently laid off. More than half of 183 employers surveyed by the National Association of Colleges and Employers rate the job market for the Class of 2026 as poor or fair, the most pessimistic outlook since the first year of the pandemic, WSJ reports. 

Big Tech’s insatiable AI power demand is pulling Big Oil further into the electricity market.

Chevon said it is aiming to bring online by 2027 a power plant that would service an AI data center in West Texas, where it pumps natural gas. The project, in partnership with investment firm Engine No. 1 and GE Vernova, is expected to provide 2.5 gigawatts of off-grid power—enough to power almost 2 million homes—with the ability to expand to 5 gigawatts if demand warrants, WSJ reports.

The U.S. plans to eliminate tariffs on bananas, coffee, beef and certain apparel and textile products under framework agreements with four Latin American nations, a senior administration official told reporters Thursday. (WSJ)

Blue Origin took crucial steps toward its space-development goals on Thursday, launching its orbital rocket for the first time with cargo and landing the rocket’s booster. (WSJ)

Boeing’s unionized machinists ratified a new collective bargaining agreement Thursday, ending a nearly 15-week strike that took a toll on workers responsible for producing jet fighters. (WSJ)

The WSJ CIO Journal Team is Steven Rosenbush, Isabelle Bousquette and Belle Lin.

The editor, Tom Loftus, can be reached at thomas.loftus@wsj.com.