HomeVideos

SpaceX’s $1.25 Trillion AI Bet | Prof G Markets

Now Playing

SpaceX’s $1.25 Trillion AI Bet | Prof G Markets

Transcript

813 segments

0:00

Today's number, 3.5 million. That's how

0:04

many Epstein files were released by the

0:07

Department of Justice last week.

0:09

Mentioned in the files was Elon Musk,

0:12

who apparently did not go to the island,

0:14

but he did try before being politely

0:17

rejected. [music]

0:19

We're not sure what's worse.

0:25

Welcome to Profy Markets. I'm Edson. It

0:28

is February 4th. Let's check in on

0:30

yesterday's market vitals.

0:32

The major indices declined amid a broad

0:35

tech sell-off. Software companies

0:37

weighed down the market after anthropic

0:39

released an automation tool for legal

0:41

work. The tech draw down hit Bitcoin as

0:43

well, dragging it to its lowest level

0:45

since November 2024. Meanwhile, Disney

0:48

shares slid again after the company

0:50

named Josh Dearo, chairman of the

0:52

experiences division as its next CEO. By

0:55

the way, Rich Greenfield predicted that

0:57

on our episode yesterday. And Netflix

0:59

shares fell as the Senate antitrust

1:02

hearing on the Warner Brothers deal

1:03

kicked off. A couple of stocks bucked

1:05

the downward trend. Walmart rose as much

1:07

as 3% and joined the $1 trillion market

1:10

cap club. And Palanteer gained more than

1:12

6% after reporting record revenues and

1:15

profits for the past quarter. Outside of

1:18

the stock market, gold and silver paired

1:20

some of their losses. And finally, oil

1:23

rose after the US shot down an Iranian

1:26

drone headed towards an aircraft carrier

1:29

in the Arabian Sea. Okay, what else is

1:32

happening?

1:35

Elon Musk's empire is consolidating.

1:38

SpaceX has acquired XAI, creating a

1:41

combined entity worth $1.25 trillion.

1:46

That makes it the most valuable private

1:48

company in history. Musk is calling it

1:51

quote the most ambitious vertically

1:53

integrated innovation engine on and off

1:56

Earth. And the logic behind the deal is

1:58

bold. It is a bet on space-based data

2:01

centers, which Musk believes are quote

2:04

obviously the only way to scale AI. The

2:07

Allstar deal comes just a week after

2:09

Musk announced that SpaceX will go

2:11

public later this year. Okay, here to

2:15

discuss the deal, we are speaking with

2:16

one of the reporters who actually broke

2:18

this story, Ed Lllo, co-host of

2:21

Bloomberg Technology on Bloomberg

2:23

television. Ed, thank you very much for

2:25

joining us again on Profy Markets.

2:27

>> Yeah, thank you for having me in Big

2:28

Story.

2:29

>> Big Story. Um, so I guess just tell us

2:33

the basics. You're the one who who broke

2:35

this SpaceX and XAI merging together.

2:39

Um, what do we know so far? Yeah, it's

2:42

an all stock transaction that values the

2:45

kind of entire entity at $1.25 trillion,

2:48

the SpaceX bit1 trillion, the XAI bit

2:52

$250 billion. But there's like some

2:54

structural things that are important,

2:56

which is the XAI basically operates as a

2:58

subsidiary of SpaceX, which is important

3:02

because SpaceX is subject to ITAR rules,

3:04

right? Rules that govern the use of

3:06

technology and defense applications. And

3:08

so like my understanding and our

3:10

understanding and our reporting is that

3:12

you know even though they're now

3:14

combined XAI kind of continues to

3:16

operate independently as its own

3:18

company.

3:18

>> How does this change things for the IPO?

3:21

Because obviously SpaceX is going to be

3:23

the biggest IPO of the year, one of the

3:25

biggest of all time. Does this change

3:28

things there?

3:29

>> Isn't that the the $ 1.5 trillion or

3:32

more question? you know, our our

3:34

reporting and our understanding is that

3:36

the work continues for SpaceX and now

3:39

the combined entity to do an IPO. In the

3:42

summer, we had reported it would be at

3:44

the midpoint of of this year. And while

3:48

like out there from the cell side, from

3:50

from lots of people that follow these

3:52

companies closely, there are still

3:54

questions about whether it happens. The

3:56

ultimate rationale behind this right is

3:58

uh data centers in space. And at the

4:00

time that we broke the story on why

4:03

SpaceX would go public um and and its

4:06

motivations for needing to raise tens of

4:08

billions of dollars, it was quite simple

4:11

that there is an ambition in place for

4:13

for space-based data center

4:14

infrastructure, but somebody's got to

4:16

buy the GPUs, right? That's that's the

4:18

kind of fixed cost. Um and and based on

4:21

our reporting and the days around this,

4:24

all of that is still holding true. So

4:26

this data centers in space thing which

4:29

yes that is what they've said he wrote

4:31

the idea is space-based AI that's the

4:34

only way to scale. So could you just

4:37

break down what that is? That is

4:39

literally let's take these data centers

4:41

that exist on the ground and instead of

4:43

having them on the ground we're going to

4:45

have them in orbit in space because

4:47

there's not enough space in the ground.

4:50

I mean why do we need to put data

4:52

centers in space? to to Elon Musk's mind

4:54

literally, you know, pardon the pun,

4:56

there's more space in space. You know,

4:59

the simple way that that that Musk

5:01

explains it is that right now the

5:03

limiting factor for scaling AI on Earth

5:06

is energy, right? You know, there's a

5:08

great um burden on grids across North

5:11

America and and other countries and

5:12

other jurisdictions. Water consumption

5:15

is an issue, impact on communities and

5:17

literal space to build the data centers.

5:20

If you put a data center in the form of

5:23

a satellite and put it into orbit, the

5:26

energy question is solved by solar. To

5:28

the mind of those advocating for this,

5:31

uh, you have to handle cooling and and

5:33

space is a vacuum and and to the the

5:35

mind of the engineers working on it,

5:37

that's easily solved. And you have

5:38

plenty of real estate, let's call it, in

5:41

space to to build on. And and that is

5:43

the vision that that Elon Musk is

5:44

pitching here. It's it's a horrible term

5:46

when you're when you're a financial news

5:48

journalist, but but what we're talking

5:49

about basically is vertical integration,

5:51

deeper vertical integration, right?

5:53

Starship provides the rocket that

5:55

carries these satellites into orbit and

5:57

deploys them. They are data centers to

5:59

all intents and purposes. XAI has

6:01

trained the models, but also those

6:04

models need to be run inference, which

6:05

you guys know that that is all outlined

6:08

in in the public coms that that SpaceX

6:10

and XI put out there, but also the

6:13

internal coms that we've reported on

6:14

around this too.

6:15

>> As I I am somewhat of a Elon Musk

6:19

skeptic and as the Musk skeptic, my view

6:22

when I see this is he's saying that the

6:25

idea is that we got to put these data

6:27

centers in space. How do we do it? Don't

6:29

ask too many questions. The idea is

6:31

we'll we'll get there. We'll figure it

6:33

out. At the same time, XAI is a company

6:37

that is competing with the likes of Open

6:40

AI and Anthropic, these companies that

6:42

are burning through billions of dollars

6:44

in cash and are having to raise billions

6:45

of dollars in cash. And XAI is itself

6:48

burning, I believe, a billion dollars a

6:51

month. And so when I see this, my

6:53

initial reaction is this is the saving

6:56

grace of XAI. If things aren't are

6:59

looking difficult over at the AI

7:01

company, well then why not just roll it

7:02

up into the bigger company which perhaps

7:05

has more uh cash to play with to invest

7:07

in in the AI models and training. Is

7:11

that too cynical? Uh what do you think

7:14

of that perspective? There are lots of

7:16

uh investors uh and people in the

7:18

markets that are are skeptical about the

7:20

the financial rationale for this

7:23

arrangement combining private XAI with

7:25

private SpaceX. And what we had also

7:27

reported last week is that Elon Musk had

7:30

looked at two distinct scenarios. One

7:33

scenario involved SpaceX combining with

7:36

Tesla which is of course a public

7:38

company. Bloomberg has reported that XAI

7:40

is burning a billion dollars of cash per

7:42

month and on a a quarter basis, you

7:44

know, you can put that in aggregate. It

7:46

also has a large debt burden which is

7:48

now on the balance sheet of SpaceX

7:50

technically. Um, those that would push

7:53

back against those concerns would say,

7:54

well, we also learned quite a bit about

7:57

SpaceX's financials. Revenue is growing

7:59

with Starlink now a majority contributor

8:01

as opposed to Launch being a majority

8:03

contributor. and on a an EBIT basis at

8:06

least it is profitable and so they're

8:09

not concerned about SpaceX needing to

8:11

service the XAI debt or account for the

8:14

losses at the XAI unit. Um but again,

8:17

you know, the the public commentary from

8:19

from SpaceX and from Musk is that this

8:21

is about deep vertical integration. It

8:24

makes the use of the talents and

8:25

resources of both companies to the

8:27

maximum effect which is AI development

8:30

whether that's here on Earth or up in

8:31

space. So Ed, I know you've been looking

8:34

at this for a long time, reporting on

8:36

it, observing what's happening, and you

8:38

know, your job is to be objective on

8:40

both sides. I would love to know what

8:42

you think the real motivation is because

8:46

it seems like there there are two things

8:47

at odds here. There's the financial

8:49

picture. You bring up the amount of cash

8:51

they're burning. You bring up the debt

8:52

picture at XAI, which is, you know, a

8:55

question. And then there's the deep

8:56

vertical integration pitch which is

8:58

coming from SpaceX and they're probably

9:00

saying no it's not about the financials.

9:02

It's because there are synergies here on

9:04

the business side. What do you think?

9:06

What what is the true uh strongest

9:09

primary motivator for Elon Musk to merge

9:11

these two things together?

9:12

>> Yeah. And none of this is that big a

9:15

surprise. You know, it's kind of

9:17

consistent with with what we've called

9:19

Elon Inc., right? You know, Elon Musk

9:22

has many ventures. They've always had

9:25

historically close engineering and

9:27

financial ties, right? There have been

9:29

engineers that have worked on materials

9:31

at Tesla and worked at SpaceX. You know,

9:34

some of the results of of SpaceX's more

9:37

recent success like um the heat

9:39

shielding on Starship. You know, a lot

9:41

of Tesla engineers touch those hands. A

9:43

lot of Tesla engineers were the reason

9:46

that XAI could build its data centers in

9:48

Tennessee so quickly. you know, this

9:50

kind of intercomp cooperation is is not

9:53

new. The financial part is also not that

9:57

new, right? XAI has been a buyer of

9:59

Tesla's energy products from Tesla's

10:01

earnings recently. The company disclosed

10:03

a $2 billion equity investment in XAI.

10:06

It's all kind of linked. Last August, we

10:09

we did this as the cover of Business

10:10

Week magazine. You know, 6 months into

10:13

the the Trump administration's term in

10:15

office. What was it that Elon Musk was

10:17

trying to get out of being in and around

10:19

the White House working with Doge? And

10:21

it was that a lot of the priorities that

10:23

he had for his different companies were

10:25

kind of aligned all around AI, you know,

10:28

and and having proximity to that

10:30

administration allowed him to try and

10:33

influence things that would help those

10:34

those priorities. But yeah, it is AI at

10:37

the end of the day. Tesla has hard

10:39

pivoted away from cars. it now needs,

10:41

you know, to develop the software part

10:43

of AI for for autonomous driving and and

10:45

the Optimus humanoid robot program. Um,

10:49

XAI per the the regulatory filing in the

10:52

week. It's not just a financial

10:54

investment that Tesla's made. They have

10:55

a framework to now share and work on the

10:58

technology. So, fast forward, the future

11:01

is that all of these companies kind of

11:03

work in parallel. What will the space

11:05

data centers actually be doing?

11:07

Inference in space, right? running

11:09

inference on Optimus or running

11:11

inference on on models that are that are

11:13

underpinning something else, be it the

11:15

the robo taxi network. It's kind of been

11:18

hiding in plain sight a little bit.

11:19

>> Do you think ultimately they all merge

11:22

with Tesla? I mean, the way you describe

11:24

it, Elon, Inc., it's a good point.

11:27

They're all kind of collaborating. It

11:29

seems as though maybe the ultimate goal

11:30

is, yeah, just roll it all up into one

11:33

company. How likely do you think that

11:34

is? Yeah, our reporting is still that

11:37

like this SpaceX XAI combined entity is

11:40

going to do an IPO later this year, but

11:42

actually loads and loads of people's

11:44

immediate reaction to the news was okay

11:46

Tesla must be next. You know, fold Tesla

11:48

in in in some way. People just think

11:51

that that is the logical step to make.

11:54

Um when we did the August edition of

11:56

Business Week, you know, it was it was

11:58

well reported and and in conversation

12:00

with all of these sources in close

12:02

proximity to the companies and Elon

12:03

Musk, there is a a perceive a

12:05

perception, you know, a realization that

12:08

if you took these massive companies and

12:10

tried to to fold them into one

12:11

conglomerate in any environment, that

12:14

would be tricky. You know, regulators

12:15

would look at it for one reason or

12:17

another. The mechanics of it um would be

12:19

difficult. And again, I go back just to

12:21

the reporting about um the two options

12:23

that we believe were looked at, SpaceX

12:25

plus Tesla or SpaceX plus XAI. And in in

12:28

the first instance, they went with

12:30

SpaceX plus X AI. But there were a body

12:32

of SpaceX investors, you know, with with

12:35

a meaningful footprint on the cap table

12:37

that that looked at it and and to them,

12:39

the Tesla combination made a lot more

12:41

sense um than the XAI one did. Not just

12:44

because of the sort of financial health

12:46

of such a transaction but because that

12:48

what they were working on you know at

12:50

scale manufacturing vertical integration

12:53

of of those more analogous industries to

12:55

that to those investors it it was

12:57

completely obvious. It's

12:58

>> really fascinating. Ed Lllo co-host of

13:00

Bloomberg Technology on Bloomberg

13:02

television. Ed, thank you very much for

13:04

joining us.

13:05

>> Thank you for having me. We'll be right

13:06

back. And if you're enjoying the show so

13:08

far, be sure to like and subscribe to

13:10

the Profod YouTube channel at the link

13:13

below.

13:17

Support for the show comes from Public,

13:19

the investing platform for those who

13:21

take it seriously. On Public, you can

13:22

build a multi-asset portfolio of stocks,

13:24

bonds, and options. And now, generated

13:26

assets, which allow you to turn any idea

13:28

into an investable index with AI. It all

13:31

starts with your prompt. From renewable

13:33

energy companies with high free cash

13:35

flow to semiconductor suppliers growing

13:37

revenue over 20% a year overear. You can

13:40

literally type any prompt and put the AI

13:42

to work. It screens thousands of stocks,

13:44

builds a one-of-a-kind index, and let

13:45

you back test it against the S&P 500.

13:47

Then you can invest in a few clicks.

13:50

Generated assets are like ETFs with

13:52

infinite possibilities, completely

13:53

customizable, and based on your thesis,

13:55

not someone else's. Go to

13:57

public.com/provg

13:59

and earn an uncapped 1% bonus when you

14:01

transfer your portfolio. That's

14:03

public.com/provg.

14:05

Paid for by public investing, broker

14:07

services by Open to the Public

14:08

Investing, Inc., member FINRA and SIPC.

14:10

The advisory services by public advisors

14:12

LLC SEC registered advisor generated

14:15

assets is an interactive analysis tool.

14:17

Output is forformational purposes only

14:18

and is not an investment recommendation

14:19

or advice. [music] Complete disclosures

14:22

available at public.com/disclosures.

14:28

We're back with Profy Markets.

14:31

Oracle has set investors on edge. On

14:34

Sunday night, the company announced

14:36

plans to raise up to $50 billion to fund

14:38

its cloud infrastructure. It launched

14:40

the first part of that financing on

14:42

Monday with a $25 billion bond sale that

14:44

actually drew record demand. But

14:47

enthusiasm quickly turned to concern as

14:50

questions resurfaced about a related

14:52

matter, specifically Nvidia's investment

14:55

in open AI. As a reminder, back in

14:58

September, Nvidia announced it plan to

15:00

invest up to hundred billion in open AI.

15:03

However, CEO Jensen Huang recently

15:05

clarified that that figure was quote

15:07

never a commitment. and it appears that

15:10

the investment has stalled. That poses a

15:12

serious risk for Oracle as roughly half

15:15

of its $625 billion in future revenues

15:18

are made up of spending commitments from

15:20

OpenAI. So, the question is now louder

15:23

than ever. Will Open AI be able to pay

15:27

up. In total, Oracle stock has fallen

15:30

roughly 10% since Sunday [music] night.

15:33

All right, here to walk us through what

15:35

is going on at Oracle, we're speaking

15:38

with Gil Lura, head of technology

15:40

research at DAD Davidson. Gil, thank you

15:42

for joining us again on Profy Markets.

15:44

>> Thanks for having me.

15:46

>> So, there are a lot of moving parts

15:48

here. Um, Oracle says that they're going

15:52

to raise uh $50 billion worth of debt

15:55

and equity. they go out and they raise

15:58

the debt successfully,

16:01

but the stock is sliding

16:04

um down around 10% in the past couple of

16:07

days, and it seems to all have to do

16:09

with Open AI. Can you just give us the

16:12

the playbyplay? What on earth is going

16:14

on with Oracle right now?

16:16

>> No, you're exactly right. There's just a

16:18

lot of moving pieces, especially the

16:20

last couple of days. So let's just start

16:22

with the fact that all of software is

16:24

down a lot and um all of more than all

16:28

of Oracle's cash flow comes from its

16:31

software business. Let's not forget that

16:32

Oracle is mostly almost entirely a

16:36

company that sells database software uh

16:39

enterprise resource planning software.

16:41

That's where they get more than all

16:42

their profits. And and right now the

16:45

market feels like software is just a bad

16:46

idea in general. So that's been

16:48

happening at once. And then you have the

16:51

the fact that they're raising capital

16:53

and then they said that this is all the

16:55

capital they're going to need this year.

16:56

That's good. The fact they were able to

16:58

do at least part of the debt. We're not

17:00

sure they're done, but at least part of

17:02

the debt, that's good. Then you start

17:04

getting into some of the other

17:05

challenges. one is that's that's it for

17:08

this year. But for them to be able to

17:11

build the capacity that they promised

17:13

Open AI, they're going to have to do

17:15

this again next year and probably again

17:16

the year after that. Meaning raise

17:18

another $50 billion next year and the

17:22

year after that. What's happened is if

17:25

you look closely at this debt issuance,

17:27

it was not cheap. the rating agencies

17:30

gave their secured debt, the the senior

17:33

secured debt a triple B minus rating,

17:36

which is to say just above junk bond

17:39

status. This is a 50-year-old company

17:41

that is now issuing debt, almost a junk

17:44

bond rate. So that's not good. They

17:47

won't have any cash to pay that debt

17:49

down. They're going to have negative

17:51

cash flow this year. So a year from now

17:53

when they need to go back to the debt

17:55

markets, that's going to be junk bonds.

17:57

So even higher yield, higher interest

18:00

expense. To add insult to injury,

18:03

there's also the equity offering.

18:05

They're doing what's called an at the

18:07

market offering. That means that instead

18:08

of doing a full issuance uh overnight,

18:12

they're going to sell a little bit of

18:14

stock every day. And our back of the

18:16

envelope method, I could walk you

18:17

through it, is that that's probably 10

18:20

weeks of them selling 10% of the daily

18:24

active trading volume. uh every trading

18:28

day. So that's going to keep putting

18:30

pressure on the stock possibly for the

18:32

next 10 weeks. So those are those are

18:35

some of the puts and takes here in in

18:37

what is a very dynamic complicated

18:39

situation. And where does OpenAI fit

18:42

into all of this? Because there was

18:44

reporting that OpenAI had this deal with

18:47

Nvidia that maybe isn't as secure as

18:51

people once thought. Uh and then Oracle

18:55

puts out this really fascinating

18:56

statement. They say, quote, "The NVIDIA

18:59

OpenAI deal has zero impact on our

19:01

financial relationship with OpenAI. We

19:03

remain highly confident in OpenAI's

19:04

ability to raise funds and meet its

19:06

commitments." And then that statement

19:08

goes viral and everyone's kind of

19:10

laughing and pointing fingers at the

19:12

statement and at Oracle right now. So

19:14

talk a little bit about how Open AI fits

19:17

into this slightly ugly picture for

19:20

Oracle.

19:21

>> Yeah. Well, then there's all that and

19:23

and that [laughter] that tweet was also

19:26

unfortunate that that would do protest

19:27

too much, right? And this is what a

19:29

corporate account should be tweeting. I

19:31

I would take a look at that and and I'm

19:33

not sure who what the instructions were,

19:35

but I I we've talked about the fact that

19:38

Oracle is raising all this capital to

19:41

build data centers for open AI. So open

19:45

Oracle has to raise a lot of capital to

19:47

build data service for open air

19:48

>> for open to be able to pay for this

19:51

compute capacity. They need to raise a

19:53

lot of capital as well. And in fact if

19:56

you think about where Oracle stands in

19:57

line it's important to see how much

20:00

OpenAI can raise. If they can't raise

20:02

enough Oracle's not getting paid because

20:05

Oracle is probably third or fourth in

20:07

line in terms of who OpenAI is going to

20:09

get paid who opened AAI is going to pay

20:12

when they raise the capital. If open

20:14

doesn't raise enough capital only

20:15

Microsoft gets paid maybe Amazon

20:19

if we Oracle has to hope that open

20:21

raises hundred billion dollars or more

20:23

to be able to afford its grand ambitions

20:27

which include this open AI capacity. So

20:30

one company raising debt at almost junk

20:32

ratings in order to build capacity for a

20:35

startup a money losing startup that's

20:37

still struggling to raise capital so it

20:40

can pay for it. We've talked about this.

20:42

this is bubbleicious behavior. This is

20:44

not okay. This is unhealthy behavior. It

20:47

may still work out because this AI stuff

20:48

is great, but it's very very risky uh

20:52

type of situation for really for both

20:55

companies.

20:56

>> And it's really fascinating how it all

20:57

kind of ties back to Nvidia because it

21:00

was all about Nvidia investing the money

21:03

into Open AI and then that's the money

21:04

they're going to use to pay Oracle. But

21:07

there was this fascinating interview

21:09

with Jensen Huang which I would love to

21:12

to get your reactions to. He was asked

21:15

about this 100 billion dollar commitment

21:18

deal that was talked about with OpenAI.

21:21

Uh he was asked about what's going to

21:23

happen with that and this is what he

21:24

said.

21:25

>> There was never a commitment. They

21:27

invited us to uh invest up to a hundred

21:30

million dollars

21:32

>> and of course we were we were uh very

21:34

happy and honored that they invited us

21:36

but we will invest uh one step at a

21:39

time.

21:40

>> It seems that he's kind of frustrated in

21:43

that video um that people are saying

21:46

that oh you're going to you're going to

21:47

invest hundred billion dollars and he's

21:49

saying no that's not what's happening. I

21:51

would just love to get your reactions to

21:53

how he sounded and also what he said.

21:55

Yeah, there's a lot of things going on

21:56

there, too. So, uh, first of all, that

21:58

hund00 million that was talked about a

22:00

while ago, that was really just a

22:02

framework for what I would call a

22:04

rebate. Nvidia Nvidia talked to Nvidia

22:07

agreed with OpenAI for every gigawatt of

22:10

capacity that you get out there, we'll

22:12

invest $10 billion at your current

22:14

valuation, which at the time was $500

22:16

billion. Um, OpenAI hasn't even deployed

22:19

a gigawatt yet. So, they haven't gotten

22:20

that first 10 billion. So that was a

22:22

framework and and and I understand that

22:24

Jensen's saying that. By the way, the

22:26

conversation now is for Nvidia to

22:28

actually invest tens of billions of

22:29

dollars actually at this level, not as a

22:32

rebate, as is the discussion for

22:35

Microsoft and Amazon and and to also

22:38

invest in OpenAI. So it again it is

22:40

quite possible that OpenAI raised this

22:42

capital but there's a lot of noise and

22:44

the last couple of days there were news

22:46

items about Nvidia being frustrated with

22:48

OpenAI spreading itself too thin in

22:51

terms of the businesses it's it's in

22:53

deploying in terms of the chips that

22:55

it's buying. So there was frustration

22:56

Nvidia and coincidentally or not there

23:00

was a leak from open air that they're

23:01

frustrated with Nvidia's chip

23:03

performance. So, a little bit of uh

23:06

childish um conflict here, but at some

23:10

point the adults are in the room. Nvidia

23:12

has a very significant interest that

23:14

OpenAI does well, right? Remember, if if

23:17

there's no OpenAI, there's no anthropic.

23:19

Everybody's using Google TPUs. So,

23:22

Nvidia needs OpenAI to do well and it

23:24

has the capital to invest as does

23:26

Microsoft, as does Amazon. So it is

23:28

likely that those three will fund the

23:31

continued expansion of OpenAI at least

23:33

this year. But if you can imagine that

23:36

if Nvidia invested OpenAI, there's

23:37

strings attached. There's no more

23:39

discussion of other chips and there

23:41

certainly will not be communication out

23:43

of OpenAI about Nvidia chips not

23:45

performing. So that's where we're at. A

23:47

little bit of middle school drama, but

23:48

at the end of the day, uh there's a lot

23:50

of incentives going around to make sure

23:52

that the investment continues.

23:55

the debt picture that you described at

23:57

Oracle in concert with the capital

24:00

problem over at OpenAI and the fact that

24:02

they need more money to spend on these

24:05

commitments than they actually have. I

24:07

it's all very concerning. It just like

24:10

very flashing bright red signs when you

24:13

hear that. We were, you know, when when

24:16

or a couple of years ago when Oracle

24:17

started to position itself as a real AI

24:19

player, we were quite bullish on the

24:21

company and that actually uh, you know,

24:24

paid dividends going forward. But it's

24:26

now getting to the point where there are

24:28

so many red flags here. Do you think

24:30

that this is just a company that

24:31

investors probably shouldn't touch at

24:34

this point? Are we getting into a a

24:36

territory that is genuinely uh dangerous

24:40

if you're if you're buying Oracle or is

24:42

or is is it safer than than I'm

24:44

portraying? No, it is risky. And let me

24:47

just put it in context. Oracle is still

24:49

trading at 20 times forward earnings on

24:52

what what looks like low teens growth

24:54

next year. Salesforce is going to grow

24:57

low teens growth next year. It's trading

24:58

at 16 times. Adobe is going to grow low

25:01

teens next year. It's trading at 12

25:03

times. So Oracle by Again, given where

25:05

we are on software, Oracle is not cheap

25:08

by any stretch of the imagination. They

25:10

have good business the the the baseline

25:12

database business does generate a lot of

25:14

cash. The baseline Oracle cloud business

25:17

was okay. It was growing fast, not very

25:19

profitably, growing fast. [snorts] The

25:22

AI part of the business, the AI compute

25:24

part of the business appears to be such

25:27

low margin that it's not even clear that

25:29

it that they should be investing in it.

25:32

And yet again, to your point, they just

25:34

borrowed a lot and they're issuing at

25:37

least 5% of their shares in order to

25:40

fund this, which again may grow their

25:43

revenue a lot, but may not add any

25:44

profit anytime soon. Certainly not any

25:47

cash flow anytime soon.

25:49

>> So very risky, especially compared to

25:52

much better software companies that are

25:54

trading at a lower multiple right now.

25:56

>> Okay, Gil Larry, head of technology

25:58

research at DA Davidson. Thank you so

26:01

much, Gil. Lots of crazy stuff happening

26:04

in tech world. I really appreciate you

26:05

taking us through it.

26:06

>> Thank you.

26:07

>> We'll be right back. And if you're

26:08

enjoying the show so far, be sure to

26:10

like and subscribe to the Profod YouTube

26:13

channel at the link below.

26:19

Support for the show comes [music] from

26:20

LinkedIn. It's a shame when the best B2B

26:23

marketing gets wasted on the wrong

26:24

audience. Like imagine running an ad for

26:26

cataract surgery on Saturday morning

26:28

cartoons [music] or running a promo for

26:30

this show on a video about Roblox or

26:33

something. No offense to [music] our Gen

26:34

Alpha listeners, but that would be a

26:36

waste of anyone's ad budget. So [music]

26:39

when you want to reach the right

26:40

professionals, you can use LinkedIn ads.

26:42

LinkedIn has grown to a network of

26:44

[music] over 1 billion professionals and

26:46

130 million decision makers according to

26:48

their data. That's where [music] it

26:49

stands apart from other ad buys. You can

26:52

target your buyers by job title,

26:53

industry, company role, seniority,

26:55

skills, [music] company revenue. All so

26:57

you can stop wasting budget on the wrong

26:59

audience. That's why LinkedIn ads boasts

27:01

one of the highest B2B [music] return on

27:02

ad spend of all online ad networks.

27:05

Seriously, all of them. Spend $250 on

27:07

your first campaign [music] on LinkedIn

27:08

ads and get a free $250 credit for the

27:11

next one. Just go to linkedin.com/scott.

27:14

That's linkedin.com/scott.

27:16

Terms and conditions [music] apply.

27:20

We're back with Profy Markets.

27:24

So, [music] lots of AI drama this week

27:27

and at the center of it all again, it's

27:30

open AI. Through multiple reports,

27:32

[music] we now know that those $100

27:34

billion that Nvidia was going to commit

27:38

to Open AI, whatever that means, that is

27:40

now quote on ice. But what might be even

27:44

more interesting is the extent to which

27:47

Oracle has been implicated in all of

27:50

this. And even more interesting than

27:52

that is perhaps what Oracle said about

27:55

it. This was the public statement from

27:58

the official Oracle account on X which I

28:00

said to Gil. They said quote the Nvidia

28:03

OpenAI deal has zero impact on our

28:05

financial relationship with OpenAI. We

28:07

remain highly confident in OpenAI's

28:10

ability to raise funds and meet its

28:12

commitments. There are several things

28:14

that are striking about this. Number

28:16

one, why is a large corporation

28:19

commenting on a third party deal with

28:22

which they are not involved? That is

28:24

kind of strange. Number two, why do they

28:26

feel the need to say anything about it?

28:28

And number three, why bring more

28:30

attention to this deal than it already

28:32

has? These are all the questions that

28:34

everyone's asking themselves. is the

28:36

question that I'm asking myself. But I'm

28:38

only asking them because Oracle posted

28:41

about this. If Oracle hadn't said

28:44

anything, then perhaps we wouldn't be

28:47

asking these questions. But they did.

28:49

And so now we are. And the answer to

28:51

those questions is quite obvious.

28:53

They're saying these things because they

28:55

are worried about them. When they say

28:57

this has no impact on our relationship

28:59

with OpenAI, that's how you know it does

29:01

have an impact on their relationship

29:03

with OpenAI. When they say they're

29:05

highly confident that OpenAI will meet

29:07

its commitments, that's how you know

29:09

they're not highly confident that OpenAI

29:12

will meet its commitments. This is

29:13

public relations 101. This is what is

29:16

known as the Stryand effect. And that is

29:18

whenever you try to suppress or downplay

29:21

an issue that you don't want lots of

29:23

people talking about. All that happens

29:25

when you do that is it makes people talk

29:27

about it even more. And this is the

29:28

perfect example. That statement, that

29:30

press release received 5 million views

29:34

on X. To put that into perspective, that

29:37

is more than double the prime time

29:39

viewership of Fox News. And no, people

29:41

were not saying it was a good statement.

29:42

It wasn't going viral for that reason.

29:44

People said it was a bad statement. They

29:46

were trolling Oracle. They were saying

29:48

how embarrassing it was. And crucially,

29:50

in the hours after that statement was

29:52

published, Oracle stock fell 6%. So this

29:56

should be a learning moment in public

29:59

relations. I think companies need to

30:01

remember that business is now more

30:04

popular than ever as a form of

30:06

entertainment. It is a sport now. It is

30:08

a series. You Oracle are now part of the

30:12

show. And so what that means is when you

30:14

screw up like they did, the stakes are a

30:18

lot higher because everyone is watching.

30:20

Everyone is interested in how this plays

30:22

out. Now, I don't know who made the call

30:25

to post that statement. It seems to have

30:27

come from someone potentially higher up

30:29

in the company, but if you're running a

30:32

half a trillion dollar company, which

30:33

Oracle is, you need to be better. The

30:35

language needs to be way less defensive.

30:38

It needs to deal with these concerns a

30:39

lot more delicately than it did. And

30:41

most importantly, someone should have

30:43

asked, "Do we really need to post this?"

30:46

And the answer should have been no. You

30:49

know, maybe 10 years ago, people

30:51

wouldn't have noticed this, but the year

30:53

is 2026. We are in a digital era. We

30:56

live in an age of virality. The costs of

30:59

these mistakes are higher. And in this

31:01

case, it cost them roughly $25 billion

31:05

in market cap. It is quite simple.

31:08

You've got to be smarter. Thanks for

31:10

listening to Prophy Markets from Profy

31:12

Media. If you liked what you heard,

31:14

subscribe to our YouTube channel and

31:16

tune in tomorrow for more. [music]

Interactive Summary

This episode of Profy Markets explores the major consolidation within Elon Musk's empire, specifically the merger of SpaceX and xAI into a $1.25 trillion entity focused on space-based AI scaling. The discussion also covers market vitals, including Walmart's entry into the $1 trillion market cap club and a deep dive into Oracle's precarious financial situation. Oracle is facing significant risks due to its heavy reliance on OpenAI's future spending, especially as NVIDIA's rumored $100 billion investment appears stalled, leading to a PR blunder that cost Oracle $25 billion in market value.

Suggested questions

4 ready-made prompts