Nvidia’s Strategy to Make Sure the AI Boom Doesn’t Go Bust

[00:00:03] Welcome to WSJ Tech News Briefing. It's Tuesday, August 25th. I'm Isabel Busquets, a reporter for the Wall Street Journal Leadership Institute. Companies are increasingly using AI monitoring tools to track employee productivity. We're diving into how they work and strategies for outsmarting them.

[00:00:23] Then, Wall Street will be watching as NVIDIA reports earnings this Wednesday. And while the chip giant has smashed estimates for the last 14 quarters straight, it's now having to step up its strategy to patch some weak points in the market. But first, ever feel like your boss has a way of monitoring your every keystroke? In the age of AI, it's an increasingly valid concern.

[00:00:50] Tools for tracking employee productivity have only gotten more sophisticated since their debut in the work-from-home era. And thanks to an AI-enabled boost, your traditional mouse jiggler might not cut it anymore. So how do you know if you're being monitored? And is there anything you can do to game the system? WSJ's Callum Borgers joins us to share his tips on outsmarting the trackers. Callum, can you give us a common example of how employees' time is being tracked?

[00:01:20] One thing that I think is actually a welcome measure or can be for employees is a calendar integration. Let's say your company uses Slack or Outlook, and you've got the little green dot that shows when you're active, and then it goes to yellow or red if you're away and you're logged off. Well, the objection from a lot of employees was, well, if I make an old-fashioned phone call or I go to an in-person meeting, I might look like I'm away. And in fact, I'm working. I'm very much active.

[00:01:43] So one of my takeaways was be meticulous about maintaining your calendar because some of these apps now will integrate your away or present status with your calendar. So if you have that phone call or meeting blocked, the software is smart enough to know, hey, you're not just slacking off. You actually have something else going on at that time. But if you don't have it on your calendar, it might look like you're just taking a long break. So if you are taking a coffee break, would you advise putting coffee break on your calendar? That's a bit overkill.

[00:02:11] And this is where one of the things I was really trying to drill down on on this story was like, so what is the expectation here, right? Because we can get in our own heads and think, geez, well, can I ever just chit-chat in the break room? Can I go to the bathroom? You know, you don't want to become paranoid. And so, you know, I asked the chief executive of one of these companies called Insightful that makes some of the software that businesses are using. And I said, all right, what are your clients actually looking for?

[00:02:33] And I was struck when he said, actually, most of my customers are looking for between 60% and 80% activity level during the day, which was lower than I might have expected. But it's because they understand that people take some time away from their desks. So you don't have to put every single coffee break or bathroom break on your calendar. And you don't even have to necessarily sweat watching the occasional YouTube at your desk. Most reasonable employers understand that even their best, most productive workers do need to take some mental breaks during the day.

[00:03:02] What they're really trying to guard against is the people who are walking away for hours at a time and not staying on task at all. You wrote about mouse jigglers or these mechanical systems that make it look like a user's mouse is constantly in motion. Are these a solution to some of the monitoring systems? They could be, but be careful. You know, the mouse jiggler is sort of the poster device for these kinds of shenanigans. The sort of classic version is a software app that you download and it makes your mouse appear to continue moving.

[00:03:30] Or maybe you get a small device that you could plug into a USB port, say, where the mouse would typically go. And it does the same thing. I've learned that more and more companies' cybersecurity tools will block those things from running in the first place. And if they don't block it, they claim anyway that they can detect it. So what some people have done is revert to an old-fashioned mechanical solution. I spoke with a company called Tech8USA that makes a, just what it sounds like, a physical mouse jiggler.

[00:03:57] It's a dock that you place your mouse on top of at your desk and it just moves a little bit to keep the thing moving. They say it was primarily originally designed for video gamers who were worried about losing their progress in a game that wouldn't save. And then in more recent years, they found that a lot of their buyers were actually remote or hybrid workers. Another thing companies have been tracking besides worker productivity is AI usage. Companies want to see the employees using AI tools, but not too much.

[00:04:26] What's behind that and what's the right amount of AI usage? The key thing to know right now is that the more is better is already kind of an outdated strategy. We were reporting here at the Journal as recently as the early spring about so-called token maxing, right, where you're trying to use AI for everything as much as possible in the workplace to show, hey, I'm experimenting with these new tools.

[00:04:50] You know, I spoke with a company called Vizier, another one of these makers of AI-powered software that monitors employees' work habits and helps companies analyze them. And they surveyed 1,000 employees and almost half of them admitted to artificially inflating their AI usage at work, most likely driven by this desire to say, hey, look, I'm an adopter. I'm using the technology. Well, just in recent months, you've seen more and more companies saying, hold on, this AI token spending is getting pretty darn expensive.

[00:05:16] And so you can't just assume that more is better. Where that sweet spot is, I can't answer for you without knowing the details of your role or your employer. But I think that's something to keep in mind, too, is just to be careful and not go overboard about puffing up your AI usage. That was our columnist, Callum Borgers What do you think about employee monitoring tools? If you're a listener on Spotify, leave us a comment with your take. Coming up.

[00:05:43] Once again, analysts are looking to NVIDIA earnings as a bellwether of where the AI industry is headed. And the chip giant is doing what it can to make sure the AI boom stays just that. That's after the break.

[00:06:06] The world is waiting with bated breath to hear what Jensen Huang has to say on NVIDIA's earnings call on Wednesday. An increasingly global, interconnected industry is once again depending on the chip giant to prove out the AI investment thesis. NVIDIA has continued to outperform market expectations quarter after quarter. But this time, some challenges lie ahead.

[00:06:30] Including the fact that some of NVIDIA's key customers, once-cash-printing machines, are relying more on debt. WSJ's David Uberti joins us to talk about NVIDIA's plans to shore up some of those weak points itself, as well as the risks of that circular strategy. Dave, analysts and media often talk about NVIDIA stock as a bellwether for the whole AI industry. Walk us through why their earnings are so significant.

[00:06:58] Well, this isn't just a tech story at this point. It's increasingly a financial story. So, like you said, NVIDIA has been the bellwether for chip demand for some time now. And for 14 straight quarters, it has beat Wall Street's expectations in terms of its profits. But what's different about this time around is that NVIDIA is increasingly shoring up weak points in the market for AI.

[00:07:23] It is investing in startups such as Poolside, which is aimed at developing a powerful AI model. It also took a stake last week in Cloverleaf Infrastructure. And earlier this month, it put out this plan with six really big Wall Street firms to finance something like $500 billion worth of loans in which NVIDIA will provide something of a backstop for potential customers going forward. So it's not just this crucial chip maker at this point.

[00:07:52] It's also increasingly the financial center of this entire thing that we're talking about. And so you mentioned weak points for the AI market. What are these weak points and how concerned should we be about them? I mean, one of its weak points was OpenAI. In recent quarters, it reported pretty tepid sales growth, 18 percent quarter over quarter earlier this year.

[00:08:14] And there's also this sort of looming question as to how some of the other major companies in this space are actually going to turn this investment boom into profits. It does feel like there has been this constant ever-present fear that this is a boom and this boom could go bust. How real is that fear? How concerned are analysts actually about that happening? I mean, they're totally concerned about it because the market is so top-heavy right now. If you look at the S&P 500, it's weighted so much toward many of these hyperscalers.

[00:08:43] SpaceX is another addition to that group of companies as well. And those companies as well as all the chip suppliers that are feeding into that demand for those pieces of AI infrastructure, they have really driven the stock market higher in recent years. And for the broader U.S. economy, so much of consumer spending now is driven by the upper echelon of people in this country who have a lot of wealth in the stock market.

[00:09:08] If NVIDIA were to report a disappointing quarter or if other major chip makers were to show some signs of weakening demand for the so-called picks and shovels of the AI trade, that could have a cascading effect not just insofar as the stock market could stop going higher, but also it could ripple through down to consumer spending and have a larger impact for the U.S. and global economy. So let's talk now about what actually might be reported on Wednesday.

[00:09:35] You mentioned that the company has smashed analysts' earnings estimates for 14 quarters. Should we be expecting to see something similar on Wednesday? When I checked FactSet recently, the current projection is annual earnings growth of 95% this quarter. So all they have to do in order to beat expectations is post $51.5 billion worth of net income,

[00:10:00] which for any other company in the stock market would just be insane to think about. But the funny thing about NVIDIA is that traders typically sell off NVIDIA even after it beats earnings expectations. So in at least the last four quarters, NVIDIA stock has fallen after it reports these sort of blockbuster profits. So it's not just that NVIDIA has to beat expectations.

[00:10:23] It also has to beat the whisper number, the sort of expectations beyond the expectations, because so much is tethered to its growth as a company. We spoke to one guy. He said its earnings report is not just the Super Bowl. It's the World Cup final. The point that he was trying to make is that this is not just an America story at this point. Because the AI trade and AI boom has gone global, we're talking not only about all these American companies, but also chip fabs in Taiwan.

[00:10:51] We're talking about sovereign wealth funds in the Middle East. We're talking about potential data centers in Europe as well. So all of these sort of demand questions that people have can kind of be traced back to NVIDIA. So this is a very crucial sort of snapshot into how that's going and how it might look one quarter from now, one year from now, and five years from now. That was WSJ reporter David Uberti. And that's it for Tech News Briefing. If you're a listener on Spotify, be sure to leave us a comment.

[00:11:21] Today's show was produced by Julie Chang with supervising producer Katie Ferguson. I'm Isabel Busquets, a reporter for the Wall Street Journal Leadership Institute. We'll be back later this morning with TNB Tech Minute. Thanks for listening.