Is There Already Too Much AI Computing Power?
Showing mention at 7:15 — highlighted below
Transcript
39 segmentsOver 60 years, Medline has demonstrated a track record of delivering consistent growth. Today, Medline is proudly NASDAQ listed and is the largest provider of medical surgical products and supply chain solutions serving all points of care. See how Medline makes healthcare run better at medline.com. AI stocks sell off worldwide ahead of U.S. jobs data. Plus, we'll catch up with our advertising editor to hear about the technological tumult facing the industry and its workforce. You have the creative agencies thinking they're going to be disintermediated. Now you have the buying agencies that are all sitting around saying, oh, my God, you know, programmatic ad buying, we thought that was a hurdle. Now we're thinking machines are going to talk to machines and do all this work. And Russia strikes back against Ukraine with one of its largest missile. barrages since the start of the war. It's Thursday, July 2nd. I'm Luke Vargas for the Wall Street Journal, and here is the AM edition of What's News, the top headlines and business stories moving your world today.
U.S. stock futures are in the red as we near the end of a shortened trading week and ahead of a closely watched June jobs report due out at 8.30 a.m. Eastern. Investors are bracing for potentially strong hiring numbers that could raise the odds that the Fed may need to raise interest rates in the short term. Meanwhile, yesterday's tech stock slump carried over to Asia overnight. triggering a pullback in the Niki and leading regulators in Korea to once again step in to calm trading during another sell-off in the country's chip stocks. Korea's benchmark Kaspi index closed the day down almost 8%, while the tech-heavy COSDAQ, where trading was briefly halted, ended the session off around 7%. One trigger of the pullback, both in Korea and among some AI-focused cloud providers in the U.S. was a report that META planned to enter the cloud business to sell excess AI computing power. That caused meta stock to jump off hours, but reporter Sherry Queen in Singapore told me that it raised alarm elsewhere. So there have been a lot of fears about AI overspending by the Mac 7. and when these companies will produce a return on that investment. The AI rightly is really built on the belief that there is infinite demand and sustained supply shortage. This latest meta report just really raises the question that what if hyperscalers are building more compute than they can use profitably? One analyst just told me that hyper-skillers now may be preparing for a world where they have to monetize access capacity rather than just keep building and keep spending at any price. Another reason what has driven the heavy tech losses is because AI trade positioning is getting really crowded. Investors are pouring money into the exact same stocks, the Mac 7, the hyperscalers, the Asian semiconductor companies. And these concerns are especially amplified in North Asian markets, particularly South Korea. Just on the compute situation here, Sherry, it almost sounds like... There's like a secondhand market opening up for a product that previously people or corporations were only buying new at tag, shall we say. Yeah, I think your analogy of a secondhand market is correct. So at first people were really bullish on the AI computing demand, and with the AI infrastructure builders in Asia couldn't catch up with demand, they think we have to be more and more. And now the metanus just shows that maybe that's not the truth. Will the hypers still face supply constraints in the short term? They may not last forever.
And the AI infrastructure builders need to prepare for that. And investors are just a step ahead. They are trying to hedge that risk. That was the journal's Sherry Queen in Singapore. And adding a note of caution to the AI buildout this week is financial crisis hero Michael Barry of the Big Short fame, who's doubling down on bets that the hype around AI is souring. Since last year, he's taken bearish bets on the likes of Tesla. chipmaker applied materials and an ETF tracking semiconductor makers. And following Samsung and S.K. Hynix's announcement this week of plans to invest more than a half a trillion dollars into a new AI chip hub, he wrote on Substack that, quote, I see that as the beginning of an end. Barry has also added to a months long short of Nvidia. Barry and the companies, he said he's shorting, didn't return requests for comment. After years of battling with the European Union, Alphabet's Google has lost a bid to overturn its biggest ever EU antitrust fine. The EU's Court of Justice this morning upheld the more than $4.5 billion fine, which was levied after regulators alleged that Google unfairly used its dominance to make sure that traffic on Android devices went to its own search engine. Back in the U.S., Alibaba and a U.S.-based payments processor have agreed to pay $600 million to settle DOJ allegations that they allowed merchants to sell and import illegal pharmaceuticals and other restricted items into the U.S. The operator of Chinese e-commerce site, Alibaba.com, admitted that it failed to prevent around 80,000 sales for merchants between 2016 and 2024. Undercover law enforcement officials found they were able to buy contraband pharmaceuticals and counterfeiting equipment on dozens of occasions. Meanwhile, the DOJ and 17 states have reached settlements with three major egg providers to resolve allegations that they illegally colluded for years to inflate prices, including when eggs hit record highs last year. Neither Calmain Foods, Versova, or Hickman's Egg Ranch admitted wrongdoing, but under the settlements, they're on the hook for $3.3 million to be paid to the states, and 53 million eggs, which will be donated to food banks and nonprofits.
Massive Russian strikes have pounded Kiev and other Ukrainian cities, killing at least 13 people and wounding dozens more. As we've reported in recent days, the attack follows weeks of Ukrainian strikes on crucial oil refineries and military sites deep within Russian territory. Foreign correspondent Stephen Kalin says that Russia's retaliation is one of the largest since the war began and included striking Ukraine hard with long-range missiles. Missiles remain pretty difficult to intercept, and a large part of that is that Ukraine doesn't have enough of the air defense systems and particularly the interceptors used in those systems to strike down those missiles. They've been asking the U.S. and other allies around the world. to provide them with more, but they've not received enough. And Russia's been exploiting that weakness while it faces setbacks on the battlefield. It hasn't really been making many advances. And Ukraine has been more successful at striking more deeply into Russian territory. President Vladimir Putin recently reiterated demands tantamount to Ukraine's subjugation, giving no indication that he's prepared to end its invasion of Ukraine, which is now in its fifth year. Coming up with the world's top economists, sounding the alarm on artificial intelligence, we'll look at how the consulting and advertising industries are trying to adapt and make the most of AI in the process. That's after the break. If your organization isn't managing its data with Everpure, then it's likely scattered all over the place. My lord, three days have I ridden to bring thee warning. The records thy great AI machine requires are strewn across five kingdoms. Yeah. The Everpure data management platform unifies data so you can always find it. No messenger needed. No more running around? Get out of the data dark ages with Everpure, a new era in data management.
We turn now to Cintra in Portugal, where the world's top central bankers and economists have been gathering this week. And one question, dominating discussions there, is artificial intelligence a boon or threat to the global economy? While some are highlighting its productivity potential, many officials are warning that high debt stretched stock valuations and cyber vulnerabilities could tip economies into a recession. With almost every industry adopting AI, Policymakers warned that if the technology is widely successful, the resulting job displacement could spike unemployment and cripple consumer spending, as our Daniel Bach reports. One industry grappling with how AI will impact their bottom line is consulting. For years, consulting firms have relied on armies of junior analysts to do things like crunch data, build decks, and take care of a lot of tedious research. That bedrock work allowed senior partners to rake in the profits thanks to billable hours. AI is increasingly capable of automating a lot of this junior level work. It does it faster. It's more reliable. That's significantly cutting down on the hours it takes to perform some consulting projects. That's journal reporter Mark Maurer, who says the consulting industry has long been debating how to improve or jettison the hourly billing model, something AI is now accelerating. But Mark says with a shift in pay models, there are new risks. Firms have primarily focused on two alternatives to the hourly model. One is fixed fee. The other option has been outcome-based pricing. That generally means consultants get paid if they achieve certain metrics for the clients. And these are metrics that are mutually agreed upon between the two parties. But Mark says with a shift in pay models, there are new risks. There's the risk that the firm will have to absorb the costs of working for free if a project takes way longer than expected. Or if a project drags out for a while, the firm could face major delays in getting paid. GPT0, which is an AI detection platform, has uncovered some of these examples. I spoke to the CEO who noted that the shift toward fixed fee and value-based billing, as he puts it, can... considerable pressure on consultancies to produce more output. And if that comes at the expense of fact checking or certain things that they were lying on AI to do, that could result in reputational damage for the firms. Still, some firms are embracing the new pay models. Elevate is one example. The AI native accounting firm is offering financial incentives to junior employees to get creative and find ways to integrate AI into their workflows. And speaking of creatives, Madison Avenue is already all in on artificial intelligence. Journal Advertising reporter Suzanne Bernice recently returned from the Can Lion Ad Festival in the south of France and says that Rose-fuelled conversations between industry execs about the technology were filled with a mix of excitement, dread, and of course how to do more and pay less.
You can now see that AI is definitely here, and that is both good and bad for the business. There's lots of fear still about copyright infringements and different things. But clearly now we can see clients, marketers, agencies, this is go time, right? Like, creative is definitely being rewritten and using these tools. Pretty aggressively, CMOs tell me it's all about the pressure in the C-suite. How do I do more with less money? Most CMOs think is we're at this now tipping point, where it's not just about cost savings, but how do we get to a place where this improves the work and actually helps do things creatively that one could never have done without these tools? AI, of course, already plays a huge role in the ad business. Google, Meta, and Amazon currently control almost two-thirds of U.S. ad revenue, where agencies have already had to adapt, restructure, and invest in getting ads in front of app users. And where marketing meetings used to be about SEO on Google, Suzanne says brands and agencies are still trying to navigate chatbot-driven search and how to be included in AI-generated summaries. The other big area was the rise of agentic, right? Everybody thinks we're going to outsource all of our purchasing to these chatbots. And then the big question will be, after years of these advertising agencies figuring out, hey, how do I talk to humans? How do you now convince a chatbot to buy your product? The other side of this is this idea of agentic machine-to-machine ad buying. We are nowhere near this yet. And obviously, consumer behavior is really going to have to catch up now to the technology that's out there. But it's coming for sure. And it's something every brand and agency is trying to learn and figure out how to navigate. According to advertising giant WPP Media, global revenue from generative search ads are projected to hit about $5 billion this year and top. 100 billion by 2030. As to how the AI shift will impact jobs in the economy, WPP plans to cut hundreds of jobs in the coming months after slashing nearly 10,000 positions last year. Investors also aren't sure with shares in rival publicies down about 2% over the past year. And that's it for what's news for this Thursday morning. Today's show was produced by Hattie Moyer and Daniel Bach. Our supervising producer is Sandra Kilhoff, and I'm Luke Vargas for the Wall Street Journal. We will be back tonight with a new show. Until then, thanks for listening.
Most finance teams are spending on the wrong things. Expense reports that take longer than the trips they cover, spend policies living in PDFs no one reads, month and close that stretches into weeks. You become what you spend on, and right now that's maintenance instead of momentum. It's time to get Brex AF. Agentic finance that does your expenses automatically, enforces policies before spend happens, and closes your books in minutes. So your finance team stops managing tools and starts driving the business. Learn more at brex.com slash a.f.