Vertiv chairman Dave Cote: AI data-center demand 'very positive' for at least the next five years [4:40]

YouTube ^ | November 14, 2025 | CNBC Television

[SunkenCiv]

--> YouTube-Generated Transcript <--
0:01	·	order to be able to. I don't
0:04	·	even want to call it play
0:05	·	invest in this. And I'm because
0:06	·	it knows how to run the most
0:08	·	important part of the data
0:09	·	center. And Dave Cody is the
0:11	·	chairman of Vertiv, which I
0:13	·	think is when I think of data
0:15	·	center, I think of you because
0:16	·	you were the first one that
0:19	·	explained to me at $16 that
0:21	·	this was $16. Dave says, you
0:23	·	know, Jim, if you look at this
0:25	·	Vertiv, I said, oh man, it's
0:27	·	all the way at $16, because
0:28	·	have you thought about it the
0:30	·	other way? It's all the way at
0:31	·	$16. Dave Cody, thanks for
0:33	·	being on the show.
0:34	·	>> But actually the one that
0:35	·	I've always liked. You sent me
0:38	·	this clip once that said, I
0:38	·	feel like I'm sitting on a gold
0:40	·	mine. I little did I know, it
0:42	·	was a diamond mine that I was
0:43	·	sitting on.
0:44	·	>> Yeah, I love the Dave used
0:47	·	to live next door to me in
0:47	·	summit and he has a great,
0:49	·	really great, great conference
0:50	·	in in February and March that I
0:52	·	attend. Okay. So Dave, let me
0:53	·	ask you. Yeah, you're one of
0:55	·	the hardest nosed business
0:56	·	people I have ever met. Do you
0:58	·	worry that the data center AI
1:00	·	build out is now being financed
1:03	·	by so much debt, or are we okay?
1:06	·	>> Yeah, I don't know. I don't
1:07	·	worry about it all that that
1:10	·	much because, you know, the
1:11	·	amount of debt you can handle
1:13	·	always depends on what kind of
1:14	·	cash you can generate. And
1:15	·	these are big cash generating
1:17	·	assets. I mean, they they
1:19	·	generate a lot of cash. So no
1:20	·	I'm not that worried about that
1:21	·	okay. I mean usually the follow
1:24	·	on to that is are we in a
1:26	·	bubble of some kind?
1:27	·	>> I'm not going to give you
1:29	·	that because you know okay.
1:30	·	>> Good. Because I.
1:31	·	>> Think that I'm far more
1:32	·	concerned about the companies
1:34	·	whose stocks are going down
1:36	·	precipitously, whom I think are
1:38	·	doing well. For instance, I
1:39	·	have to tell you, I think
1:41	·	alphabet is doing incredibly
1:42	·	well, but it's getting taken
1:44	·	down with companies that don't
1:46	·	have cash. And I think that
1:47	·	Nvidia is real and that people
1:49	·	are finally they're giving up
1:50	·	on it, in part because it's
1:52	·	going down. And I think that I
1:54	·	remember when Birdie was at 100
1:56	·	and right after they reported,
1:58	·	no one said, Jim, you know it.
1:59	·	We're our orders are going up.
2:02	·	Yeah. It what is the state of
2:03	·	things is what I'm saying.
2:05	·	>> Oh, I'd say it's very
2:07	·	positive for at least the next
2:09	·	five years that everything is
2:11	·	five years. Yeah. It's kind of
2:12	·	tough to look much beyond that.
2:14	·	But if you just take a look at
2:16	·	the orders and you take a look
2:18	·	at how much their AI businesses
2:20	·	are growing, so there'll be a
2:21	·	tendency to look at, okay, data
2:23	·	center sales are up so much.
2:24	·	But if you take a look at AI
2:26	·	sales within some of these big
2:28	·	companies, a couple of them are
2:29	·	triple digit. Others just talk
2:31	·	about how robust it is and AI
2:34	·	usage is evolving. We still
2:36	·	think of it as ChatGPT you do
2:38	·	something on your machine which
2:39	·	you're going to start to get
2:41	·	more of, and this is happening
2:42	·	is like machine to machine. So
2:44	·	AI to AI, no human interface.
2:46	·	Right. So you'll see it in a
2:48	·	factory, for example, where
2:49	·	this machine is seeing a
2:51	·	problem that they if things can
2:53	·	be rectified down the line by
2:54	·	another machine, it can talk to
2:56	·	the other machine. At some
2:57	·	point you'll get to a point
2:59	·	where a company wants enough.
3:01	·	Trust exists. A company's AI
3:02	·	can talk to another company's
3:04	·	AI. I mean.
3:05	·	>> This is.
3:06	·	>> Very this is all just
3:07	·	beginning.
3:07	·	>> You think you think there'll
3:10	·	be whole divisions of companies
3:11	·	that aren't run by people?
3:12	·	>> Oh, wow. Not soon. Yeah,
3:17	·	okay. Not soon. But, I mean,
3:19	·	you keep extending this stuff.
3:21	·	At some point, you'd have to
3:22	·	say, yeah, probably. Probably
3:24	·	could.
3:25	·	>> Or at least strategic
3:25	·	missions will be written by not
3:28	·	people.
3:28	·	>> You'd have to you keep
3:29	·	carrying it on, you say. Yeah,
3:32	·	at some point that has to be
3:33	·	possible. But it's not
3:34	·	something that I would say is
3:35	·	anytime soon.
3:36	·	>> Now I want people to
3:38	·	understand where your company
3:38	·	fits in the food chain. If I'm
3:40	·	in a data center without a
3:42	·	naked data center, they run
3:43	·	really, really hot and
3:44	·	therefore they burn up. Can you
3:45	·	tell people in the food chain
3:48	·	why Vertiv is the one that you
3:51	·	would that would know more than
3:51	·	almost anybody?
3:52	·	>> Oh yeah. Well, we've got the
3:54	·	panoply of the innards of a
3:56	·	data center other than the
3:57	·	server. So you think about the
3:59	·	two of the big things that you
4:01	·	have to accomplish are
4:04	·	uninterruptible, sustainable
4:04	·	power coming in, which is a lot
4:06	·	more difficult than it sounds.
4:08	·	The other is how do you cool
4:09	·	everything in the building? It
4:11	·	used to be very simple. You
4:13	·	just had a big air conditioner,
4:15	·	you know, cool air in warm air
4:16	·	extricated. Now with this
4:17	·	liquid cooling where you have
4:18	·	to get the liquid directly to
4:20	·	the chip in order to remove
4:22	·	heat. Significant part of the
4:24	·	functioning of a data center.
4:25	·	Because if you're not cooling
4:27	·	those chips, the thing stops
4:29	·	working very quickly. Third
4:29	·	piece of it is the server rack,
4:31	·	and you think it sounds like a
4:33	·	simple thing that you just buy
4:35	·	in a store somewhere. But these
4:36	·	server racks that the servers
4:38	·	plug into very sophist...

[SunkenCiv]

I'm pretty sure I'd never heard of this one before. When there's a big construction project going on, the drive-throughs surrounding the site have steady business...

[bakeneko]

Anybody considered that the “AI Race” is kind of like how Reagan bankrupted the Soviet Union with “Star Wars”, except this time China (or shadier globalist principalities) is doing it to us?

[SunkenCiv]

The US is actually quite far ahead. The main threat from China is their usual weaponization of technology to run scams, since the whole country is about a billion criminals.

[dpetty121263]

AI is an example of how a Nation destroys its self from within and ceases to exist and it is right on schedule.

[RoosterRedux]

Bfl

[MMusson]

There is not enough electricity. You can have electric cars or AI. But, you cannot have both.

[ridesthemiles]

A HUGE DATA CENTER HAS BEEN APPROVED BY THE COUNTRY COMMISSIONERS IN OUR SMALL RURAL HIGH DESERT N NEVADA VALLEY.

1070 ACRES...THAT IS HUGE FOR THIS AREA.

WE ALL HAVE WATER LIMITS ON OUR PARCELS....2.2 ACRE FEET A YEAR FOR DOMESTIC USAGE.

DATA CENTERS CAN USE OVER 450,000 GALLONS A DAY—— NOT RECYCLED. UP INTO THE DRY DESERT AIR...GONE.

THAT EQUALS THE DAILY WATER USAGE OF 252 PARCELS IN THIS VALLEY.

THE IMPACT WILL BE VERY NEGATIVE FOR US, I FEAR.

[RoosterRedux]

Vertiv is on my watch list. It is a steadily growing "picks and shovels" part of the AI gold rush, providing cooling, power, racks, and connectivity infrastructure (i.e., not chips and servers).

Here's it's 3rd quarter 2025 10-Q.

[RoosterRedux]

AI is an example of how a Nation destroys its self from within and ceases to exist and it is right on schedule.

How do you figure?

If the US doesn't develop AI faster than China, we will all be speaking Mandarin shortly.

[RoosterRedux]

I’m not saying that AI won’t eventually lead to things that can be used to destroy humanity, but every major technological leap has triggered the same kind of fear and exaggeration we’re hearing about AI.

When steam engines and railroads arrived, people sincerely believed traveling 20–30 mph would rupture your organs, cause brain damage, and/or cause women’s uteruses to fall out. Newspapers warned that trains would stop cows from giving milk, make chickens stop laying eggs, panic horses, and set farms on fire. Moralists insisted railroads would corrupt society by letting people move too easily, upsetting the social order and turning stations into dens of vice.

Established industries—canals, stagecoaches, toll roads—claimed railroads would wreck the economy, eliminate jobs, and destabilize entire regions. Critics said steam engines were inherently unsafe, prone to explode, and that railroads violated nature and even God’s intended order.

[SunkenCiv]

Nonsense.

[SunkenCiv]

Thx!

[SunkenCiv]

Building tech facilities that use a lot of water, especially clean freshwater, are more appropriately sited in place like, say, the Great Lakes region.

[dpetty121263]

Well lets think about this a minute AI is being sold as a way to save tons of money for business with compute power as whole departments and organizations are being cut as people are the biggest expense for business. Now AI was sold at the beginning AI would free up staff to do other things but that is not happening as they found they do not need that staff anymore. Ok what are all these people supposed to do if no jobs are available any more? In China they do not care what happens to their people as if they lost 100 million it would be a blessing as it would be that fewer mouths to feed. Case in point when Reagan sent over a delegation to discuss Nuclear disarmament the Chinese told them they could afford to lose 100 million and would not even notice it. So with all these job losses coming what are you going to do with the massive unemployment created with AI, let them starve and go homeless or what? Overload your Social Services budgets? This AI race is a race we need to think about and freeze any AI development till this is thought out... The Government may be forced to proved a living paycheck to all unemployed compared to what they were earning so what you were trying to avoid in becoming a Communist country may well happen or a real Revolution...

[montag813]

There is not enough electricity. You can have electric cars or AI. But, you cannot have both. ?

We need dozens of nuclear plants, ASAP. Musk is a fool to say Solar-Uber-Alles. It's just not enough and never can be.

[RoosterRedux]

You’re looking only at layoffs, but not at the demand side. Many companies—including Palantir and a long list of AI-driven firms—are actually desperate for new workers right now. The economy isn’t shrinking; it’s shifting. Some roles are going away, but new ones are being created faster than companies can hire for them. The real challenge is helping people transition, not assuming jobs are disappearing forever.

And ultimately, AI is coming whether we like it or not. And if we don't lead the way, China will.

[montag813]

WE ALL HAVE WATER LIMITS ON OUR PARCELS....2.2 ACRE FEET A YEAR FOR DOMESTIC USAGE.

We should have diverted the Mississippi River decades ago to serve the Southwest states, along with desalinization plants along the California coast.

Desalinating Water Is Becoming “Absurdly Cheap”

How cheap is cheap? Energy Monitor notes that “globally, around 1% of the world’s drinking water is desalinated, but in Israel, that figure is around 25%.” Israel’s desalinated water comes from five desalination plants. The Sorek B plant has a capacity to desalinate 52.8 billion gallons a year and is contracted to produce water for $0.41 per cubic meter. There are around 264 gallons per cubic meter, so this puts the cost at about a penny per 6.4 gallons.

One hundred percent of the municipal water supply in the United Arab Emirates is desalinated. Dubai bloomed out of the desert with desalination technology. There are some 186 desalination facilities under construction or at the pre-construction phase around the world.

[RoosterRedux]

As an aside, a lot of jobs lost due to AI isn't because AI is replacing people. It's because AI is being used to identify inefficiencies and unnecessary jobs. This makes companies and industries stronger.

As my doctor told me recently, it isn't that doctors are going to be replaced by AI. It's that doctors who use AI are going to replace doctors who don't.