So it seems like because you have these speculators who don't really know what they're doing, they might be giving a bad name to some of these data center builders, adding to people's mistrust overall.

Not just the speculators.

I think we gotta be clear that some big, more experienced builders of data centers, specifically SpaceX AI, have done things like install many unlicensed small natural gas turbines, for example, in their Colossus data center campus in and around Memphis.

They have acknowledged that these are temporary turbines.

They say that they have begun shutting them down, and eventually they're gonna replace them with a one-point-two-gigawatt natural gas power plant, a more conventional one that's cleaner.

So in the race to stand up new compute as quickly as possible, you're seeing even experienced data center developers do some things which local communities object to, like setting up these temporary power plants, which are louder and more polluting than what's been used in the past.

How would that work?

These data center builders are already spending more money.

So we can help them by giving all of that while being neutral or agnostic on, okay, which of these hundred different capabilities do you want to use, and how do you do it in a way that's safe?

So let's take a, a step back and talk about some fundamentals because I think one of the challenges that a lot of analysts and journalists have with IBM is they don't actually, as far as I can tell, understand what your revenue mix is and how you make money [chuckles] quarter to quarter while everyone is building this AI future.

It's twenty twenty-six.

I think people don't appreciate that the core of your business is still this thing called the mainframe, and people don't even know what a mainframe is, right? It's a very peculiar beast, but it's in everything.

Like, if mainframes went away tomorrow, like that would be like an asteroid, like extinction-level event-

I acknowledge that.

Right.

And so you've also said in the past, I think your CFO reiterated this on your most recent call, for every dollar that somebody spends on IBM hardware, that's like two or three dollars that is gonna occur in terms of recurring software revenue, right?

How does that work?

So a lot of companies are claiming that through so-called water restoration projects, which could be anything, maybe they restore a wetland in order to retain more water, or they're paying to stop leaks in a municipal water system, that they are offsetting the water that they use directly in a given community.

And so they're claiming that by 2030, Google talks about this, Amazon talks about this, they're going to put back a gallon of water for every gallon that they directly use in their data centers.

But again, this doesn't account for indirect water use at the power plants powering these data centers.

So Christopher, what's the big takeaway here?

If you live in a water-stressed area and somebody proposes building a data center in that area, and they say, "Hey, don't worry, we're gonna limit the amount of water it uses," the next question should be, "Great.

Where does the power for that data center come from?" Because if it means that, let's say, a coal-fired power plant that would otherwise be retired is gonna stay online, if it means a bunch of new natural gas generation, which is often the case, the actual all-in water use of that data center could be 10 or 20 times greater than what is reported.

Can you walk us through the math? Is this thing saving you money?

So the math on this depends on two things.

One is your latitude and the exposure of the surface that you're putting the solar panels on, and then two, the thing honestly that it depends the most on is the cost of electricity where you are because for every kilowatt hour you produce with this system, that displaces however many cents you would have otherwise paid to the utility company.

For me, it's difficult to say without a year-long trial exactly how much these panels that I put on this part of my roof are gonna produce, but my guess is maybe $500 a year for the exact system that I was playing around with, so that's a pretty good payback.

But your mileage very much may vary, and for a lot of folks, to get the most out of a system like this, they're gonna need to connect their panels to some kind of backup battery, which will allow them to gather power during the day and then use it in the early evening, which is when home power use tends to spike.

Where do you think that these systems are heading?

I think potentially individual US states, New York, California, could end up being a gigantic market for these because these are places where people are already just paying eye-watering amounts for power.

You know, when people are faced with that amount of pain month after month from their utility company, there's a really powerful incentive to essentially prepay their utility bills for the next few years and kind of take things in their own hands.

Are there any benefits to accelerating code development at this kind of extreme pace?

There are really big benefits when AI agents are applied to coding in the right context.

So, for example, I talked to Anthropic and OpenAI recently, and they said there's a lot of legacy code that just needs to be refactored for newer systems, and it's just a huge amount of drudge work, and you can send AI in and kind of make a first pass and then even use AI to review that code to make sure that it's meeting best practices and to make sure it's secure.

The human still has to go in at the end of that and verify that everything has been done right, but it can just save many, many hours of work in this context.

The important thing, though, is that in that process, the human is very much in the loop.

It's more like when you turn on cruise control on your car as opposed to getting in a Waymo and trusting it to get you to your destination with no input.

But is there a danger in a generation that never learns how to code or do the stuff the, quote-unquote, "hard way"?

That is, I think, almost a trillion-dollar question, and I have heard so many different answers to it.

So generally speaking, what is , at this particular moment in time, what is AI good for, and what is it not good for?

Generally speaking, AI can be great at basic tasks like helping you edit a document, generate a presentation, right? Bouncing ideas around.

That's what these generative AIs, these chatbots that we use are great at.

There's, of course, a vast universe of, you know, what I call classic AI.

All the AI that came before the ChatGPT moment when it broke onto the scene in November of, what was that? Twenty twenty-two now.

But those tend to be the kind of things that get used by scientists or, you know, Meta or Google.

You know, these are the classic machine learning and deep learning systems that are populating our social media algorithms and giving us directions in our mapping app of choice.

So Christopher, as amazing as AI is for what it does well, what does it not do so well? What is it bad at?

He says these jamming devices are cheap and small and increasingly used in warfare and beyond.

GPS jammers really took off with the conflict in Ukraine because Ukraine was relying on GPS for two things.

One was to help guide their, uh, drones, and then also because there are a number of so-called smart bombs, smart munitions that were relying on GPS in order to find their target.

And Russia figured this out pretty quick, and according to some public testimony, the effectiveness of one GPS-guided bomb that the US used to use a lot went from seventy percent hit rate to six percent hit rate in the span-

The Strait of Hormuz is basically a GPS dead zone?

Yeah.

It's worse than that.

There's a ton of spoofing happening in the Strait of Hormuz, which is a little more sophisticated.

What-- but was there a specific breakthrough with AI, some specific tool that was-- that changed your mind and turned you into a convert?

Yeah, I think it was NotebookLM, and it was just kind of beautiful how they had...

You know, it was the, it-- in the same way that Slack made, you know, ICQ, uh, accessible, right? Or, or Reddit, uh, made, you know, Usenet, brought it to the masses or something.

There's that moment where you take something and you turn it into a really friendly, really easy-to-use product.

You really make hard design decisions about how you're gonna limit its functionality, and you s-- and you really kind of handhold a little bit, and you're like, "This is what it's for," right? I mean, NotebookLM was, was s- sort of co-designed internally, right, by a famous author, I think it was Steven Johnson.

Was at Google for a couple of years, or maybe still there.

And he was like: "I want to create the, uh, ambient brain that I've always wanted when writing my books," his very brainy, deep books that are full of facts about science and technology.

Yeah.

Christopher, what is happening in the realm of neurotechnology?

In terms of human computer or human AI interfaces, lots of interesting things are coming.

They're all, like, wearables.

So Meta, uh, released a, uh, wristband that just by moving your hand it reads the neural impulses in your nerves going into your hand and allows you to control an interface inside a, a VR headset or smart glasses.

But there are more coming that are sort of more medical in their application.

There's a company called Cognition.

They combine a headset and a skull cap full of electrodes that reads your brain, and they claim that they're going to be able to help you diagnose early more than 400 different brain disorders and potential neurological disorders.

Christopher, what's your take on where satellite technology is heading?

Christopher, what is happening in the realm of neurotechnology?

In terms of human computer or human AI interfaces, lots of interesting things are coming.

They're all like wearables.

So Meta had released a, uh, wristband that just by moving your hand it reads the neural impulses in your nerves going into your hand and allows you to control an interface inside of a VR headset or smart glasses.

But there are more coming that are sort of more medical in their application.

There's a company called Cognition.

They combine a headset and a skull cap full of electrodes that reads your brain, and they claim that they're going to be able to help you diagnose early more than 400 different brain disorders and potential neurological disorders.

Christopher, what's your take on where satellite technology is heading?

Christopher, why do you make the argument that today's tech giants are becoming more like the robber barons of yore, or in other words, spending vast amount of money on actual stuff rather than code?

So there's been this big change.

When I started covering tech more than a decade ago, you would look at the quarterly filings for a company like Meta, and what felt like the big liability was the exploding cost in terms of payroll.

They were just hiring engineers so quickly.

You look at their balance sheet now, or any of the big tech companies, Amazon, Alphabet, etc., and the big expenditure now is on stuff.

It's real estate.

It's data centers.

And it's not just the U.S., right? Who are the other big tech spenders of actual stuff across the globe?

Why do we need better solutions for all of AI's huge projected power demands?

The main reason we need new ways to do AI and rethink how we're building it from the ground up is that the projected growth in the amount of power that AI requires is pretty unsustainable.

So one estimate has it that 50% per year increase in the amount of electricity required for AI is a reasonable scenario through 2030, and we're already at the point where companies like Meta and Google and others are building multi-gigawatt data centers, and one gigawatt of power, that's like a pretty decent sized coal-fired power plant.

So META has said, "Hey, we're gonna build this data center.

It's gonna be as big as Manhattan, and someday it's gonna take five of those power plants." And Google has said, "We're gonna need nuclear power.

We're investing in fusion.

We're investing in exotic ways to store energy so we can make this work in Europe where energy is even more expensive."

Tell us about these chips.

What is it doing that could be a potential game-changer in this solid state battery space?

Ion came up with a new kind of solid state battery.

They borrowed technology from fuel cells, like hydrogen fuel cells, and they said, "What if instead of the usual kind of crystal, chunky granola matrix that's inside a solid state battery, we put this elegant ceramic, like the ceramic in your coffee mug, and it's full of tiny holes where the lithium ions can hang out as they're moving back and forth between the two poles of a battery?" And what do you know? This battery does not expand or contract the way that a normal solid state battery does.

It's very stable, and it can be manufactured in factories that already make regular lithium ion batteries.

So very importantly, it doesn't require any kind of new gigafactories to be built.

So for consumer electronics, what does this mean? In what ways could it revolutionize the way that we use everyday technology, and how soon could that happen?

If Ion's promises bear out, and keep in mind that, you know, it's rare that we get a fundamentally new battery technology, so no one should be surprised if they completely flame out.

But if what they claim proves true in the real world, you could have your smartphone last 50% longer, or you could have things like smart glasses which last all day long.

Christopher, what is this not-so-new technology for getting clean water, and why is it getting fresh attention now?

So this desalination technology was invented in the 1960s, but it wasn't viable then because in order to make it work you have to put the desalination plants at the bottom of the ocean, 400 meters down.

And the reason you want it down that deep is just that the water's pretty clean, and the water pressure's very high.

And the way that modern desalination works is you're pushing water across a membrane with tiny holes that only admit water molecules, and the holes are so tiny that they don't let salt through, and that's how you desalinate water.

So the whole idea is, why don't we use, use the enormous pressure at the bottom of the ocean to push seawater through this membrane.

And yeah, you still gotta pump some fresh water up to the surface, but it just takes a lot less energy.

And that is the real bugaboo of desalination.

Tell us some of the ways that these companies are testing this technology in this way.

China has plenty of other sources, so how do they stack up against the American chip giant?

So China has been cut off not just from US high-end chips for AI, like from NVIDIA and AMD, but also from the tools and in some cases, even the chemicals and other materials required to build the latest generation of their own AI chips.

And so really since about 2018, China has been on a tear building up its domestic capacity to fabricate high-end chips and build everything else in that whole really complicated supply chain, which of course starts with silicon, literally sand, but includes so many other things in order to create the most advanced chips.

They haven't been able to replace everything in that supply chain from domestic sources, but they're kind of on the path and there's been a lot of interesting innovation.

China's largest chip manufacturer SMIC, which is their answer to TSMC, the Taiwanese chip manufacturer which makes chips for Apple, has basically enhanced older chip-making technology so it can make newer kinds of chips.

And this has allowed companies like Huawei to continue to make some of the most advanced AI supercomputers in the world.

And even the CEO of NVIDIA has said Huawei's latest AI supercomputer is comparable to or better than the ones made by NVIDIA.

Is that true? And what happens next?

So Christopher, you're essentially saying that we're nowhere near AGI, is that right?

That's correct.

That's the right takeaway.

We are definitely nowhere near AGI, and anyone who tells you differently, I think, honestly, just hasn't looked that deeply into what intelligence actually is.

It turns out that what today's transformer-based AIs, and that's the kind of AIs that underlies ChatGPT and a lot of other generative AIs, the way that they work is just they have this kind of almost infinitely long list of little rules of thumb that they apply.

And so to give you a concrete example, one reason, historically, these models have been really bad at math, even if you show them a million math problems and their correct answers, is that they learn weird stuff.

You know, if you ask them to multiply two numbers and one of them is between 200 and 211, it has a different set of little rules of thumb it uses for multiplying those numbers than it does for any other numbers anywhere else on the, the number timeline.

Can you talk about that and how it explains the bag of heuristics theory?

So, what are you stocking up on right now? (laughs)

You know, my day job is as a tech reporter.

So it's the place where I am the most updated in my knowledge about the impacts of these tariffs.

And, uh, you know, you don't need to go out and buy a new phone because they, uh, gave, uh, an exception for those, you know, probably 'cause, uh, you know, Donald Trump's friend Tim Apple, as he likes to call him, you know, has made multiple trips to the White House and has, has greased those skids for his whole industry.

But every other type of good that isn't like a phone or a tablet or like a laptop that's on a, a piece of electronics, like that's definitely gonna go up in cost.

It's funny enough, there are deals I think because companies are selling out the inventory that they have now or that they moved into the country in advance of the tariffs.

So if there's some big complicated object that has a bunch of electronics in it that you need now, like probably now's the time to do it.

That's grimmer than empty store shelves.

Chris, how does this electric motor minus rare earths work?

There's a company called Conifer that are based in Silicon Valley and they have found a cheap way to make a rare earths-free motor.

Their first application is going to be, like, Vespa-style scooters.

Turns out that's a huge market, there's a billion of those in the world.

But, you know, this could go into those vacuum cleaners, it could go into your HVAC system.

It could be scaled up, it could go into electric vehicles.

So that's a big deal.

They just drop it in where you would use a normal motor or how does that work?

So Christopher, just put into perspective for us how highly anticipated the Switch 2's launch is both in and outside your own household.

(laughs) Within my own household, my two sons, a teenager and a nearly teenager, uh, have been saving all of their pennies in order to buy one of these, and turns out that's not atypical.

There are 190 million gamers in the United States of America.

Nintendo has sold more than 50 million of the original Switch consoles in the eight years that it has been out, and the level of anticipation for the Switch 2 is through the roof.

This is singly the best-selling, uh, video game console in the US certainly, and so the fact that kids might not be able to get their hands on it here in the US because of tariffs has made tariffs a subject of conversation among my children and apparently in many other households.

Could that dollar amount go even higher because of the tariffs?

Initially it seemed like Nintendo was either going to have to not sell the Switch in the US, which really is an impossibility for them, or possibly charge a great deal more under the original tariff scheme where they were looking at, we have to toggle back in our minds 'cause it changed so much, but originally it was, like, 50% tariffs from China or 50% tariffs approximately from Vietnam, and those are the two places where they manufacture almost all of the Switch 2s.

It now appears with the tariff pause that Nintendo could potentially sell the Switch 2 at its original $450 price because the tariffs are only, and I'm putting that in air quotes, "10%" from Vietnam.

Is Zipline really gonna make this a widespread reality?

Yes.

The startup called Zipline thinks that they're gonna make it a widespread reality.

I tend to believe them just because they have some deep-pocketed competitors who also say they're gonna make it a widespread reality, and they include Amazon and Wing, which is a sister company to Google.

Zipline is different in that they have been doing drone delivery for so long.

If you haven't heard of them, it's because y- you don't live in Rwanda where they deliver the overwhelming majority of medicines and blood products from centralized blood banks.

One study said that they had slashed maternal mortality across the entire country just by delivering blood all over the country with their original drone, which looked like a...... small remote-controlled airplane.

How does Zipline and even its other competitors think about that when it comes to the future of drone delivery in America?

Just walk us through what their plan is.

So, one thing that's been happening in Hollywood and video games lately is folks are using AI to eliminate the jobs of concept artists, for example.

They'll replace them with AI image generation tools, and voice actors 'cause they'll use AI clones of actors' voices.

The fear is that eventually, we'll just start using AI to generate movies themselves.

And so, the Russo brothers' whole idea is, "We don't want to leave the future of AI in filmmaking in the hands of tech companies because they will be dastardly in how they use it." And they want to try to use AI to, for example, accelerate the creative process.

Now, they're pretty vague on what that looks like, but based on previous interviews with their chief scientist, Dominic Hughes, it sounds like they believe that AI can be a companion when you are coming up with new ideas, for example.

There is research to back that up.

So, walk us through that concept.

One thing you mention is that World is trying to be this sort of everything app. We've seen everything apps be popular in places like China.What exactly is an everything app, and why haven't we really seen something like that in the US?

So in China, in particular, WeChat, but also in South Korea in the form of KakaoTalk, there are these apps where inside of the app you can do everything. You can shop, you can order the equivalent of an Uber. You can do your banking. And these have been really popular in Asia for a very long time, and they never really took hold in the West because we kinda didn't need them. We had the Apple App Store, we had the Google Play Store, and so instead of a quote unquote everything app, you just download whatever apps you need. People like Elon Musk, Mark Zuckerberg, and now Sam Altman have said, "Wouldn't it be great if we could create a everything app in the West?" Because it's extremely lucrative to be the place where people conduct their entire online lives. World is offering a new mini app store within their World app, and they are saying, you know, that this might be a way to possibly succeed at creating an everything app for the West.

Are there risks to scanning things like our fingers and our eyeballs for authentication? How do we think about privacy and security?

So biometrics have come a long way in the past, I don't know, 15 or 20 years. A couple decades ago, it was valid to worry about using some pattern on your body, whether it's your fingerprint or your eye or your face or the veins in your hand as an identifier, but engineers have figured out pretty clever ways to scan parts of our body and then create this mathematical encrypted abstraction of that, that image or that data and then use that to securely identify us. We're already doing this on-device, right? Every time you unlock your phone with your face, we're already taking a big step in that direction. So, you know, there are definitely systems that in the future will prove to be insecure. Like all systems, there's always a backdoor somewhere or another, but World, Amazon with their hand-scanning technology, and others, they seem really serious about not losing people's trust by creating insecure systems.

(instrumental music plays) Newly powered by AI brains, humanoid robots are becoming increasingly capable. Currently they're mostly used in what's known as materials handling, basically moving things around in warehouses and manufacturing facilities. And while these are generally low stakes tasks, humanoid robots could save money for companies that use them. Here to tell us about the tech powering these bots and what's next for the industry is our tech columnist Christopher Mims. Can you talk to me about some of the startups and companies and some specific humanoid robots that you've seen lately?

Okay. This is Christopher Mims' top humanoid robot companies in order of likelihood that you'll actually hear about them again five years from now and they won't just be gone. All right, number one, Apptronic. Apptronic is a Austin, Texas based company. Their focus is on making humanoid robot bodies as affordable as possible, so that's good in terms of making them viable. Agility Robotics, they are the first in the world to really be used in production in a Spanx warehouse outside of Atlanta, Georgia. They're moving shapewear. Reflex Robotics, they're doing cool stuff with a logistics company called GXO. Picture a robot picking stuff out of bins and, and putting them in boxes to be shipped to you. Can't count out Tesla's Optimus robots. There's Figure. They just got a boatload of money and they claim to be building kind of start to finish an AI-powered robot. There's Boston Dynamics. They've been at it forever. They make those terrifying big dog robots that you've seen videos of on the internet, and now they make humanoid robots. And they don't have any tests with customers that they're willing to tell me about yet, but they're good at what they do and they've had more experience at it than anybody.

What's the tech that's helping current robots do more than before?

So robots, they're different than other computers in that they have bodies. So...One big thing that's helping them is all the parts that make up their bodies are getting cheaper, they're getting more capable, their camera eyes, that all comes out of the cell phone supply chain. We've gotten really good at making that kind of thing. Their joints, their limbs, their actuators, their motors, their equivalent of our muscles, people have done a lot of work in terms of making those stronger, more precise, lighter, use less power. That's really key, because these robots, unless you want them plugged in all the time, they gotta run on batteries, and they apparently really chew through batteries pretty quick. Then there are their brains, and there have been a lot of breakthroughs lately in terms of using the same kind of underlying AI that drives things like ChatGPT to learn how to animate a robot body. And pretty much all these companies are betting that that innovation's gonna continue at a pretty swift pace in order to make these robots more versatile.