December 10, 2025 • 27 mins
SoftBank is doubling down on artificial intelligence with a $30 billion commitment to OpenAI and ambitious plans for Stargate – even as risks mount. After a 200% surge earlier this year, shares tumbled 40% in November when Google’s Gemini 3.0 gained traction, threatening ChatGPT’s dominant position in the AI race. Credit stress is rising, with CDS widening to nearly 300 basis points as funding concerns build. Investors and creditors are increasingly uneasy as OpenAI now accounts for about 20% of SoftBank’s net assets, turning what was once seen as a bold growth play into a potential source of concentrated risk.
Bloomberg Intelligence equity analyst Kirk Boodry and credit analyst Sharon Chen join John Lee on the Asia Centric podcast. Together they unpack SoftBank’s AI ambitions, its reliance on margin loans and the implications of circular financing. They also weigh up Masayoshi Son’s track record – spectacular wins such as Alibaba and painful failures like WeWork – against his latest AI gamble.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Soft Bank is making one of the biggest bets in
its history on artificial intelligence. Founder and CEO Masayoshi Son,
famous for Boom and bus wages, has borrowed heavily to
fund massive investments in Open Ai and Stargate. But after
a two hundred percent surge earlier this year, shares plunged
(00:20):
forty percent in November when Google debuted Gemini three to
positive reviews, raising fears of tougher competition for Open Ai.
Credit risk is also flashing red. SoftBank's credit default swaps
has jumped to nearly three hundred with a twenty billion
funding gap looming. Will masses ai game will pay off?
(00:41):
Or is SoftBank taking too much risk for both equity
and credit investors. You're listening to Asia Centric from Bloomberg Intelligence.
I'm John Lee in Hong Kong. Today we have kerk Boudry,
equity analyst who has covered soft Bank for more than
fifteen years, and Sharon Chen, credit analyst based in Singapore,
both at Bloomberg Intelligence. Kirk and Sharon, welcome to the show.
Speaker 2 (01:04):
Thank you great to be here.
Speaker 1 (01:07):
SoftBank share price has been on a spectacular rollercoaster this year.
Can you give us some details as to why this
stock price ramped up so much earlier and why did
it sell off recently.
Speaker 2 (01:18):
Yeah, I mean you referred to that. I mean, just
talking about open ai is the major driver. What SoftBank
is doing is pivoting from being an investment company to
something that's more AI focused, and the investment open Ai
is a big part of it. That was big driver
of the share price because the valuation basically doubled from
(01:41):
the time that soft Bank first said it was investing
until now, and a lot of their investment is backloaded
at the original price, so people are just building in
this open Ai gain. One of other things that's important,
it's kind of been lost because of all the AI
frenzies that they're also pivoting out of their telecompositions, So
they made some fairly hefty sales of shares of Sprint
(02:04):
and Tea Mobile, which investors are quite happy to see.
Speaker 1 (02:07):
So if you look at SoftBank's position in open Ai,
what portion does it comprise of its say, total assets
or its net asset value.
Speaker 2 (02:15):
Yeah, I mean at this point pro former investment that
they still have to complete for in December, it's about
twenty percent of net asset value.
Speaker 1 (02:25):
So that's quite large.
Speaker 2 (02:26):
Yeah, it's between that and ARM because Arms forty percent.
You know, that's almost two thirds of their value just
in those two positions.
Speaker 1 (02:34):
How much is soft Bank riding on open Ai to
be successful, because on one hand, we've never seen a
company grow revenues as quickly as open Ai. It's forecast
to generate roughly thirteen billion dollars in sales this year
and it's only been around for a few years. But
on the other hand, it's losing about ten billion dollars
(02:55):
in losses each quarter.
Speaker 2 (02:57):
So, I mean, what's happened in the month of November.
Basically his investors are sort of taking a deeper look
into open ai and early on what you had, like
Soft Brank invested it basically a two hundred and sixty
billion dollar valuation, and then by August people were talking
about five hundred billion dollars and then got confirmed. And
(03:19):
underpinning these pretty aggressive numbers is forecasts from open Ai,
you know. So they go out and talk to investors
and they say, probably by twenty twenty nine, we're gonna
have one hundred and twenty five billion dollars in revenue. Right,
it's an unprecedented growth rate from where they are now.
To put it in context, if you look at back
(03:41):
last year and there's single digit billions and you go
to one hundred and twenty five billion in five years,
No one grows like that. I mean the closest is
Meta and Google. We were able to do it in
like eleven and twelve years. Companies that sell directly to consumers,
like open Ai, it took longer. So they're saying we're
going to grow faster than anyone ever has in history.
(04:03):
When they started talking about unprecedented growth, investors got excited
about that. When Gemini came out with a model that
was competitive, then people started thinking, ooh, market share may
be a bit of an issue. And if the market
is much more competitive, then the revenue targets may be aggressive.
(04:24):
Then the valuation may also be aggressive. When you look
at the opening eye valuation and how investors look at it,
I mean, we generally talk about revenue growth, but the
other side of that that is equally important is that
the operating losses are huge, and that the Microsoft accounting
implies that you know, you have something like ten billion
dollars in losses a quarter, and the guidance from the
(04:46):
company is that they're not going to break even and
for another four years probably, So you need all of
that revenue growth just to ensure that the company is
not going to continue to lose money. And if the
revenue growth is in question, then the idea that they
could reach break even is also an issue. I mean
break even ever. So, so I.
Speaker 1 (05:08):
Looked at some stats and this is from similar Web,
and it shows that Gemini's marketshare in web traffic in
the gen ai market has been rising. So I think
it was about five percent last year and recently it's
up as close to fourteen percent. So they are gaining
market share at the expense of Chat GPT. And by
the way, it's not just Gemini that open Ai has
(05:29):
to worry about. So if you look at other lms
like Claude, Grock, Deep Seek, they're also constantly improving and
could pose risks ahead.
Speaker 2 (05:38):
Yeah, I mean you'll see variations in market share all
the time. If you believe that open ai is worth
five hundred billion dollars, then you also have to believe
that they're going to be dominant in the market.
Speaker 1 (05:48):
Right, So I can see two risks, and you know,
please correct me. I can see two risks with soft Bank.
Number one is that we're in AI bubble, and obviously,
you know, all tech suffers. The second risk is not
so much that we're in an AI bubble, but the
structure that market dynamics change away from say Chatchy, between
open Ai to Google.
Speaker 2 (06:08):
Yeah, everyone talked about AI bubble, you know once it
can a burst. Well, I don't know. I mean it's
a pretty transformative technology, you know, So we can shelve
that for now. But I think you're right is soft
Bank's recent weakness has a lot more to do with
concerns on open Ai in particular than it does with
AI bubble.
Speaker 1 (06:26):
Right, And I want to bring Sharon to this conversation.
Soft Bank is not just an equity story. It's also
a credit story. Now, Massa and soft Bank has financed
these huge bets in AI through heavy borrowings. Can you
just you know, let the listener know how much are
they borrowed to finance these investments.
Speaker 3 (06:45):
So this year alone, they've raised sixteen billion dollars worth
of bonds. Majority of that came from the domestic market,
but they've also been active issuers in US dollar and
eurobond markets. On top of that, they're increasingly reliant on
margin loan. So this year we've seen them upsize an
ARM backed margin loan to twenty billion dollars. Eleven and
(07:05):
a half billion dollars of that is still undrawn. And actually,
just very recently we've seen a company filing that they've
upsized a SoftBank Corp backed margin loan as well to
about eight billion dollars and about two point four billion
dollars is undrawn.
Speaker 1 (07:20):
So with their ARM holdings, like what percentage of their
holdings are actually on margin loan now?
Speaker 3 (07:26):
So about eighty percent of their ARM shares are pledged
against the margin loans. And the good thing is, even
though the company has been upsizing their margin loan, actually
they haven't had to pledge additional ARM shares because you know,
the valuation of ARM obviously has been rising over time.
Speaker 1 (07:43):
Okay, so you know they've promised to invest more into
open Ai as well as to fund Stargate. What's their
cash flow situation looking forward to the rest of this
year and next year?
Speaker 3 (07:55):
So in terms of cash outflows, we estimate that the
company needs about sixty five to seventy billion dollars and
that includes to be fair two years worth of bond
maturities because it's the company's financial policy to maintain that
much liquidity. We've also included investment in open Ai, so
the company has committed to a thirty billion dollar investment,
(08:17):
and twenty two and a half billion dollars of that
is due to be invested this month. Actually, on top
of that, we have five point four billion dollars worth
of ABB Robotics ARM deal. And there's just a lot
of refinancing needs as well, because the company funded the
first change of open Ai investment and also the mper
(08:38):
computing deal using one year bridge loans, so those who
need to be refinanced. They've also fully drawn down on
a six billion dollars short term commitment line, and that
I assume it's either going to be rolled over they
have to turn it out. So based on our calculation,
I think even though I've talked about how much funding
they've raised, including from assets sale, actually we still think
(09:00):
the company has about fifteen to twenty billion dollars of
fundraising needs in the next twelve months. This depends on
how much investment in Stargate as well, because you know,
the final amount hasn't been disclosed.
Speaker 1 (09:12):
Yet and how are they going to plug this short form.
Speaker 3 (09:16):
I think it's going to rely on the bond markets. Again,
just in terms of bonds that are due or callable.
In twenty twenty six, the company has closed to eight
billion dollars, so i'd expect a company to raise anything
from ten billion dollars in terms of bonds. The good
thing is SoftBank is well regarded in the domestic market.
They have strong access to the Japanese yen retail bond market,
(09:38):
and in fact, they are the largest issuer there, accounting
for about forty percent of the market based on the
company's presentation. On top of that, I expect the company
to sell more assets. So this year they've already raised
over ten billion dollars from asset sales, and this includes
T Mobile and Vidian Deutsche Telecom. They still have some
T Mobile steak and steaks in other listed technology stocks
(10:01):
like intern Symbotics, So I think they can raise over
ten billion dollars from asset sales as well.
Speaker 1 (10:07):
And this is like a huge outlin. Like you mentioned,
they're spending sixty five billion dollars in these investments. What's
the potential for a ratings downgrade?
Speaker 3 (10:17):
So just to clarify, I think the investments are closer
to forty five to fifty billion dollars to sixty five
include some refinancing needs as well for their bond maturities.
But in terms of their ratings, so Softbaang is rated
WB plus by S and P and there are two
downgrade triggers. One is around adjusted loan to value, So
(10:37):
the threshold for the rating is thirty five percent, and
based on our perform our calculation, we think it's going
to come to around thirty percent when we've included all
these investments and assuming you know, five to ten billion
dollars investment in Stargate, so it still has some rating
headroom and we think it can take about a fifteen
(10:58):
percent drop in portfolio of value before hitting the thirty
five percent threshold. The second down grade catalyst is coming
from the portfolio side, So if there's pressure on the
credit worthiness of the company's portfolio, a s NP may
consider to downgrade it as well. This is a bit
more subjective, obviously, but we do think the portfolio is
(11:18):
becoming a bit riskier because, as Kirk mentioned earlier, open
AI is going to account for a larger portion of
the portfolio and ARM is going to come down to
about forty percent. The company has been selling stakes in
mature listed stocks like T Mobile and Deutsche Telecom. Now,
from an equity perspective, it sounds positive, but from a
credit perspective, we view those as more stable, mature listed
(11:42):
companies that they can monetize if needed. Lasting to highlight
is the declining portion of listed stocks. So back in
December twenty twenty four, it was about seventy eight percent.
September we've reached below seventy percent, and we think it
could trend lower towards sixty percent, but it's hard to tell.
You know, at what point this could trigger a downgrade,
(12:03):
but the rating is still unstable outlook at this moment.
Speaker 1 (12:06):
So you mentioned that SoftBank has pledged, especially with its
ARM holdings, as collateral for these loans. So if the
ARMS share price falls, then they'll have to pretty much
you know, give you know, the lens. It's more money, right.
Speaker 3 (12:19):
So the terms haven't been fully disclosed, but the way
it works is that if there's a significant fall in
ARMS share price, then the company may have to repay
the loans or pledge some there's cash collateral pledging needs
having said that based on our calculation, the loan to
(12:40):
value for the ARM margin loan is still fairly low
at twenty percent or below, So I think the company
still has headroom. The only complication we see is that
if indeed, in a downside scenario, something goes probably wrong
with the Open AI investment, or indeed there's an AI
bubble and it bursts and the company needs to sell
more shares than you know what the T mobile and
(13:03):
other tech stocks that I've mentioned. It can sell some
of the ARM shares because you know, still about twenty
percent is unpledged, but then there's limited free float with
SoftBank still owning eighty seven percent of ARM, So you know,
we could see a scenario where if they're forced to
sell ARM shares, arms share price could tank, unless you know,
(13:24):
it brings in a strategic investor.
Speaker 1 (13:26):
So Sharon and A're putting all this together, like what
are the risks for credit investors?
Speaker 3 (13:33):
So I think the risk is obviously SoftBank's debt is rising,
and the reason it's rising is because of all these
investments in unlisted companies. I mean, you mentioned Opening I
already has thirty billion dollars of revenue, but the valuations
seem pretty lofty, and there's competition from Gemini, so the
(13:53):
dead levels are rising and that just leaves less headroom
for the company to whether you know, potential market downturn
or something going wrong with their investments.
Speaker 1 (14:05):
Asia Centric is produced by Bloomberg Intelligence, where more than
five hundred experienced analysts and strategists work around the clock
to bring you timely, world class research. Our coverage spans
two hundred market industries, currencies, commodities, and industries, as well
as over two thousand equities and credits. To learn more
about Bloomberg Intelligence, visit bi go on the Bloomberg terminal.
(14:28):
If you like what you're here, don't forget to subscribe
and share. Kirk, is this a case of history repeating
itself because Massason is famous for these big bets that
have been wildly successful, but it's also made huge losses
as well.
Speaker 2 (14:43):
I mean, if you go back and look at you
soft Bank share Pricey's history, there is a definite pattern.
I mean, you saw way back in the late nineties
and two thousand when you first invested in Yahoo in
the US, and for a very short time he's the
richest man in the world, and then lost ninety nine
percent of when they bought Vodafone Japan in two thousand
(15:04):
and six, right before the Great Financial Crisis. They borrowed
heavily to do that. They're leveraged up six and a
half times net debt to EBIT down, which in today's
soft bank environment doesn't seem like a lot, but for
you know, the telco world at the time, it was
pretty crazy, you know. And then you had it again
around the COVID crisis and quantitative easing, you know where
(15:25):
they went massively in on Vision Fund. As the first
Vision Fund started to see companies, you know, coming to
market and surfacing value. They invested fifty billion dollars over
two years at the top of the market, which is
something that they're just now getting out from under. So
when the company decides there is a particular strategy they
(15:48):
want to follow, and I should say, when credit markets
are open, you know, they will go heavy. And this
is what we're seeing again.
Speaker 1 (15:57):
And Vision Fund is still loss making right well.
Speaker 2 (16:00):
This last quarter like it moved back on a cumulative basis,
it moved back to profitability, and the headline was the
games in Open Ai because some of that stake is
being held a vision fund two, but even if you'd
stripped that out, they would be back to break even,
which was a very good result.
Speaker 1 (16:18):
So you lived in Japan for many years, I think
it was like sixteen years. What's his perception as an
investor in Japan? Because, as you mentioned, it's had some
real big home runs, but he's also had some you know,
spectacular fails as well.
Speaker 2 (16:31):
Yeah, I mean it's SoftBank's very unique, like within the
Japanese sort of investing environment, because what it is is
it to play on like a lot of the foreign
tech themes always has been if you look at their
asset value, the only asset of real sort of substance
that they have a soft bank the telecom company. But
everything else that they've done is outside. So when investors
(16:54):
look at it, they go, oh, this is a way
we can play AI. You know that domestically, I mean,
there's it's not the only way. But when you're talking
about some of these sort of high profile you know,
global names like open AI and video whatnot software, could
you access to that?
Speaker 3 (17:12):
Okay?
Speaker 1 (17:12):
So if we look back in Son's history, what score
card would you give him as an investor, Because, like
he said, some spectacular home runs like Ali Baba. He
made I think over seventy billion dollars in Ali Baba.
He's also made a lot of money in Yahoo. Originally,
he's had some like, you know, spectacular losses, like obviously
we work as one that comes to mind. Ft X
(17:35):
is another. Yes, I forgot about that Green seal as well.
Speaker 2 (17:39):
Yeah, I mean I would give him you know, a
B A B minus maybe because obviously every time you
look at SoftBank, you know, when you go through these periods,
it's that those bubbles are high, are those asset valuations
are higher. So he is building something, but there have
been some pretty incredible losses and also lost opportunities.
Speaker 1 (18:03):
Referring to the in Video stake where he sold too
early in the first round.
Speaker 2 (18:07):
Right, there's actually three or four different ways to look
at what Softwaank's done with in Video. The first thing
is that they bought it for Vision Fund and over
a year they doubled their money and then they sold it.
And this is the one where you know, he's like,
I'm crying that I sold the shares. It was you know,
terrible because they would have owned like I don't even
remember what the number would have been.
Speaker 1 (18:28):
So he sold up before in Video had that most
recent share price r This.
Speaker 2 (18:33):
Was back in twenty seventeen and eighteen, and it was
before like Generitive AI was on anyone's radar screens. It
was just an investment. And at the time they doubled
their money. And this was at the end of twenty seventeen.
So we're going into the first quart of twenty eighteen,
in Vidia falls fifty percent in the quarter and everyone's like, oh, well,
(18:54):
you know, they made this beat on in Video for
the new Vision Fund. Doesn't look like it's working out,
And it turns out he'd locked in the games with
you know, derivative transactions or whatnot. And that was the thing.
Like at the time, Softbanks sold the in Video shares
because they wanted to make a point that Vision Fund
is going to be a money maker, and so they
(19:14):
accomplished that. Everyone was surprised. They're like, oh, that was
pretty smart trading. Even though now that the stock is
up like twenty times since then, you know, water under
the bridge. But there's other things like they were going
to sell Armed to Nvidiot one point and if they
had done that, they would have picked up twenty one
billion dollars of in video shares, and so this isn't
(19:36):
his fault because he wanted the deal to happen, but
it wouldn't. But that steak, you know, would be worth
twelve times as much, so it'd be worth like two
hundred and eighty billion or something. After that deal collapsed,
they bought a small token position. This is the only
time he's done fantastically well within video shares because they
bought like one hundred million dollars and that ended up
(19:56):
being worth one point four billion. This was part of
the stake that they just old last quarter. They made
huge investment to up that position earlier this year. So
his track record is he's made a little bit of
money on in video over time, but when he could
have really transformed SoftBank, he wasn't in video shows.
Speaker 1 (20:18):
Right, But you have to say that, look, you know,
before we recorded this podcast, I looked at the Bloomberg
Rich Index and he's still worth over thirty billion dollars. Sure,
so he's still one of the most richest men in
the world. Having said that, he was worth over seventy
billion dollars in nineteen ninety nine. So it depends on
which way you look at it right.
Speaker 2 (20:39):
Yeah, I would say, like when I think about I
mean this question about whether Massa is a good investor
or not. The one thing that he does have in
his favor that he's absolutely fearless, you know, and he's
willing to stick by the conviction of his ideas and
borrow like just incredible amounts of money to do it.
So it does show that he believes in what he's doing,
(20:59):
and maybe that's it's more important even if some of
the investments haven't done as well as he would have lied.
Speaker 1 (21:05):
And Sharon, you speak to a lot of credit investors.
What are the major risks and perceptions for soft Bank
and Massa right now?
Speaker 3 (21:13):
I think there's a lot of concern from the fundraising perspective,
where are they going to get all this money from?
What are they going to sell? And for bond investors,
particularly in the US dollar europe on market, there's a
concern that the supply risk is going to lead the
bonds to underperform. On top of that, obviously there's less
concern about rating pressure at this moment, but I would
(21:35):
say there is concern that the debt is going higher
and to what level? Right? So the key question I
keep getting is is there an AI bubble? Which I'm
not sure I'm in a position to answer. But you know,
from a credit perspective, we just look at how much
headroom the company has, right, and we've done some scenario analysis,
(21:58):
and we think the company can actually use asset sales
to help manage the loan to value even if asset
values drop by So if it drops by thirty percent,
they might have to sell twenty billion dollars. That's like
a really downside scenario that you know, we're not saying
that's the base case at all.
Speaker 1 (22:13):
There's been a lot of you know, public criticisms of
the circular financing of a lot of the AI investments,
and soft Bank has been you know, obviously part of
this with its Stargate investment. Is this a risk?
Speaker 3 (22:27):
From my perspective, I think it's a risk more from
open ai being central to a lot of these circular transactions,
right because open ai has been working with Oracle and
VideA and obviously soft Bank, and these sort of transactions
where open a I agrees to lease these data centers
ends up pumping the valuation for all these AI related
(22:48):
stocks higher. But from a credit investor, perspective other than
you know, the potential valuation dropped for the portfolio. The
other concern that I didn't mention earlier is whether open
Ai will need to raise more funds in the future
because the company is expected to continue bleeding cash. And
what sort of role is soft Bank going to play
in that? Because SoftBank has already invested over thirty billion
(23:10):
dollars in open Ai, and if this is their big bet,
is there a risk that you know, Masayoshi Son likes
to go big, is there a risk that we'll see
more investments in the future.
Speaker 2 (23:22):
Yeah. I mean it's interesting because soft Bank is usually
not in this conversation globally about like the circular funding
and everything in Ai, but I think it's absolutely a
part of open Ai. I mean, they're investing thirty billion dollars.
Open Ai is going to use that money to lease
data centers that soft Bank is building through Stargate in
open Ai is also building using the money that SoftBank
(23:45):
invested in them. So all of this sort of like
data center strategy is supported by soft Bank doing the
open Ai investment. And then on top of that, soft
Bank has committed to spending three billion dollars a year
on open ai services in Japan, which you know, when
you think about like the size of that one those
(24:06):
first announced like last year when they said that they're
going to do this with open i, they were only
generating three point seven billion revenue in total. So it's
a lot of what got the open Ai valuation or
business kicked off earlier this year was because SoftBank was
willing to stand behind them.
Speaker 1 (24:26):
And Kirk before we let you go, I wanted to
ask about the valuations of SoftBank. Now, SoftBank is an
investment firm. It has these huge stakes in these both
listed and unlisted companies. Is the share price reflecting all
its investments, Like does a trade at a premium or
a discount too? You know what analysts called the sum
of the parts?
Speaker 2 (24:47):
Yeah, I mean it right now, it's at about a
thirty percent discount to net asset value, which is measured
on the latest public pricing for their liquid assets. And
you know what the carry value is for the private assets.
Back in the end of October, when SoftBank shows got
up to twenty seven thousand, it was trading at a premium,
(25:07):
which is basically the market saying we think open ai
is worth more. But a lot of that has come off.
And another way to look at it is if you
look at the history of the discount and we've started
measuring it since Ali Baba went public back in twenty fourteen,
and basically the history throughout this entire time, it's only
(25:27):
been a premium twice and the most recent was in October.
On average, SoftBank trades are the thirty to forty percent discount.
So at this level right now where we're at, if
investor perceptions of the assets don't change, then it's probably
where it should be.
Speaker 3 (25:45):
Sharon.
Speaker 1 (25:46):
As I mentioned earlier, the credit to fault swaps or
CDs for SoftBank Group has jumped up to almost three hundred.
What is that signaling to investors?
Speaker 3 (25:55):
So it's signaling all the caution I was talking earlier
about rising debt funded investments. It's up from low two
hundreds in early October. But back in twenty twenty two,
which is the last time you know, we saw SoftBank
had these large investments and at that time it was
in Vision Fund two, the CDs actually rose to over
(26:18):
five hundred basis points. So I think the market is
still giving the company some benefit because it has a
strong tread record of managing its balance sheet and its
credit risk. So back in twenty twenty two, what we
saw was the company essentially ended up selling down at
stake in Ali Baba via prepaid forward contracts to help
manage loan to value. So I think that, you know,
(26:40):
we probably see a repeat of this, and the company
is very committed to you know, having loan to value
below twenty five percent.
Speaker 1 (26:47):
And for the benefit of the listener, the credit to
vault swap is a gauge on whether the company could
go bankrupt or could default on its loans, and higher
than number one, more likely it will be right.
Speaker 3 (26:59):
Yes, it's a match default risk.
Speaker 1 (27:02):
This has been a really fascinating discussion and I think
we're in for a wild ride for shareholders and credit
investors on Softbak Kirk Sharon, thanks.
Speaker 2 (27:10):
For joining, Thanks for having me.
Speaker 3 (27:12):
Thanks John, you've.
Speaker 1 (27:14):
Been listening to Asia Centric from Bloomberg Intelligence. I'm John
Lee in Hong Kong. You can listen to all our
episodes on Apple Podcasts, Spotify or review Listen and this
podcast was produced and edited by Clara Chen. Thanks for listening.
Soft Bank is making one of the biggest bets in
its history on artificial intelligence. Founder and CEO Masayoshi Son,
famous for Boom and bus wages, has borrowed heavily to
fund massive investments in Open Ai and Stargate. But after
a two hundred percent surge earlier this year, shares plunged
(00:20):
forty percent in November when Google debuted Gemini three to
positive reviews, raising fears of tougher competition for Open Ai.
Credit risk is also flashing red. SoftBank's credit default swaps
has jumped to nearly three hundred with a twenty billion
funding gap looming. Will masses ai game will pay off?
(00:41):
Or is SoftBank taking too much risk for both equity
and credit investors. You're listening to Asia Centric from Bloomberg Intelligence.
I'm John Lee in Hong Kong. Today we have kerk Boudry,
equity analyst who has covered soft Bank for more than
fifteen years, and Sharon Chen, credit analyst based in Singapore,
both at Bloomberg Intelligence. Kirk and Sharon, welcome to the show.
Speaker 2 (01:04):
Thank you great to be here.
Speaker 1 (01:07):
SoftBank share price has been on a spectacular rollercoaster this year.
Can you give us some details as to why this
stock price ramped up so much earlier and why did
it sell off recently.
Speaker 2 (01:18):
Yeah, I mean you referred to that. I mean, just
talking about open ai is the major driver. What SoftBank
is doing is pivoting from being an investment company to
something that's more AI focused, and the investment open Ai
is a big part of it. That was big driver
of the share price because the valuation basically doubled from
(01:41):
the time that soft Bank first said it was investing
until now, and a lot of their investment is backloaded
at the original price, so people are just building in
this open Ai gain. One of other things that's important,
it's kind of been lost because of all the AI
frenzies that they're also pivoting out of their telecompositions, So
they made some fairly hefty sales of shares of Sprint
(02:04):
and Tea Mobile, which investors are quite happy to see.
Speaker 1 (02:07):
So if you look at SoftBank's position in open Ai,
what portion does it comprise of its say, total assets
or its net asset value.
Speaker 2 (02:15):
Yeah, I mean at this point pro former investment that
they still have to complete for in December, it's about
twenty percent of net asset value.
Speaker 1 (02:25):
So that's quite large.
Speaker 2 (02:26):
Yeah, it's between that and ARM because Arms forty percent.
You know, that's almost two thirds of their value just
in those two positions.
Speaker 1 (02:34):
How much is soft Bank riding on open Ai to
be successful, because on one hand, we've never seen a
company grow revenues as quickly as open Ai. It's forecast
to generate roughly thirteen billion dollars in sales this year
and it's only been around for a few years. But
on the other hand, it's losing about ten billion dollars
(02:55):
in losses each quarter.
Speaker 2 (02:57):
So, I mean, what's happened in the month of November.
Basically his investors are sort of taking a deeper look
into open ai and early on what you had, like
Soft Brank invested it basically a two hundred and sixty
billion dollar valuation, and then by August people were talking
about five hundred billion dollars and then got confirmed. And
(03:19):
underpinning these pretty aggressive numbers is forecasts from open Ai,
you know. So they go out and talk to investors
and they say, probably by twenty twenty nine, we're gonna
have one hundred and twenty five billion dollars in revenue. Right,
it's an unprecedented growth rate from where they are now.
To put it in context, if you look at back
(03:41):
last year and there's single digit billions and you go
to one hundred and twenty five billion in five years,
No one grows like that. I mean the closest is
Meta and Google. We were able to do it in
like eleven and twelve years. Companies that sell directly to consumers,
like open Ai, it took longer. So they're saying we're
going to grow faster than anyone ever has in history.
(04:03):
When they started talking about unprecedented growth, investors got excited
about that. When Gemini came out with a model that
was competitive, then people started thinking, ooh, market share may
be a bit of an issue. And if the market
is much more competitive, then the revenue targets may be aggressive.
(04:24):
Then the valuation may also be aggressive. When you look
at the opening eye valuation and how investors look at it,
I mean, we generally talk about revenue growth, but the
other side of that that is equally important is that
the operating losses are huge, and that the Microsoft accounting
implies that you know, you have something like ten billion
dollars in losses a quarter, and the guidance from the
(04:46):
company is that they're not going to break even and
for another four years probably, So you need all of
that revenue growth just to ensure that the company is
not going to continue to lose money. And if the
revenue growth is in question, then the idea that they
could reach break even is also an issue. I mean
break even ever. So, so I.
Speaker 1 (05:08):
Looked at some stats and this is from similar Web,
and it shows that Gemini's marketshare in web traffic in
the gen ai market has been rising. So I think
it was about five percent last year and recently it's
up as close to fourteen percent. So they are gaining
market share at the expense of Chat GPT. And by
the way, it's not just Gemini that open Ai has
(05:29):
to worry about. So if you look at other lms
like Claude, Grock, Deep Seek, they're also constantly improving and
could pose risks ahead.
Speaker 2 (05:38):
Yeah, I mean you'll see variations in market share all
the time. If you believe that open ai is worth
five hundred billion dollars, then you also have to believe
that they're going to be dominant in the market.
Speaker 1 (05:48):
Right, So I can see two risks, and you know,
please correct me. I can see two risks with soft Bank.
Number one is that we're in AI bubble, and obviously,
you know, all tech suffers. The second risk is not
so much that we're in an AI bubble, but the
structure that market dynamics change away from say Chatchy, between
open Ai to Google.
Speaker 2 (06:08):
Yeah, everyone talked about AI bubble, you know once it
can a burst. Well, I don't know. I mean it's
a pretty transformative technology, you know, So we can shelve
that for now. But I think you're right is soft
Bank's recent weakness has a lot more to do with
concerns on open Ai in particular than it does with
AI bubble.
Speaker 1 (06:26):
Right, And I want to bring Sharon to this conversation.
Soft Bank is not just an equity story. It's also
a credit story. Now, Massa and soft Bank has financed
these huge bets in AI through heavy borrowings. Can you
just you know, let the listener know how much are
they borrowed to finance these investments.
Speaker 3 (06:45):
So this year alone, they've raised sixteen billion dollars worth
of bonds. Majority of that came from the domestic market,
but they've also been active issuers in US dollar and
eurobond markets. On top of that, they're increasingly reliant on
margin loan. So this year we've seen them upsize an
ARM backed margin loan to twenty billion dollars. Eleven and
(07:05):
a half billion dollars of that is still undrawn. And actually,
just very recently we've seen a company filing that they've
upsized a SoftBank Corp backed margin loan as well to
about eight billion dollars and about two point four billion
dollars is undrawn.
Speaker 1 (07:20):
So with their ARM holdings, like what percentage of their
holdings are actually on margin loan now?
Speaker 3 (07:26):
So about eighty percent of their ARM shares are pledged
against the margin loans. And the good thing is, even
though the company has been upsizing their margin loan, actually
they haven't had to pledge additional ARM shares because you know,
the valuation of ARM obviously has been rising over time.
Speaker 1 (07:43):
Okay, so you know they've promised to invest more into
open Ai as well as to fund Stargate. What's their
cash flow situation looking forward to the rest of this
year and next year?
Speaker 3 (07:55):
So in terms of cash outflows, we estimate that the
company needs about sixty five to seventy billion dollars and
that includes to be fair two years worth of bond
maturities because it's the company's financial policy to maintain that
much liquidity. We've also included investment in open Ai, so
the company has committed to a thirty billion dollar investment,
(08:17):
and twenty two and a half billion dollars of that
is due to be invested this month. Actually, on top
of that, we have five point four billion dollars worth
of ABB Robotics ARM deal. And there's just a lot
of refinancing needs as well, because the company funded the
first change of open Ai investment and also the mper
(08:38):
computing deal using one year bridge loans, so those who
need to be refinanced. They've also fully drawn down on
a six billion dollars short term commitment line, and that
I assume it's either going to be rolled over they
have to turn it out. So based on our calculation,
I think even though I've talked about how much funding
they've raised, including from assets sale, actually we still think
(09:00):
the company has about fifteen to twenty billion dollars of
fundraising needs in the next twelve months. This depends on
how much investment in Stargate as well, because you know,
the final amount hasn't been disclosed.
Speaker 1 (09:12):
Yet and how are they going to plug this short form.
Speaker 3 (09:16):
I think it's going to rely on the bond markets. Again,
just in terms of bonds that are due or callable.
In twenty twenty six, the company has closed to eight
billion dollars, so i'd expect a company to raise anything
from ten billion dollars in terms of bonds. The good
thing is SoftBank is well regarded in the domestic market.
They have strong access to the Japanese yen retail bond market,
(09:38):
and in fact, they are the largest issuer there, accounting
for about forty percent of the market based on the
company's presentation. On top of that, I expect the company
to sell more assets. So this year they've already raised
over ten billion dollars from asset sales, and this includes
T Mobile and Vidian Deutsche Telecom. They still have some
T Mobile steak and steaks in other listed technology stocks
(10:01):
like intern Symbotics, So I think they can raise over
ten billion dollars from asset sales as well.
Speaker 1 (10:07):
And this is like a huge outlin. Like you mentioned,
they're spending sixty five billion dollars in these investments. What's
the potential for a ratings downgrade?
Speaker 3 (10:17):
So just to clarify, I think the investments are closer
to forty five to fifty billion dollars to sixty five
include some refinancing needs as well for their bond maturities.
But in terms of their ratings, so Softbaang is rated
WB plus by S and P and there are two
downgrade triggers. One is around adjusted loan to value, So
(10:37):
the threshold for the rating is thirty five percent, and
based on our perform our calculation, we think it's going
to come to around thirty percent when we've included all
these investments and assuming you know, five to ten billion
dollars investment in Stargate, so it still has some rating
headroom and we think it can take about a fifteen
(10:58):
percent drop in portfolio of value before hitting the thirty
five percent threshold. The second down grade catalyst is coming
from the portfolio side, So if there's pressure on the
credit worthiness of the company's portfolio, a s NP may
consider to downgrade it as well. This is a bit
more subjective, obviously, but we do think the portfolio is
(11:18):
becoming a bit riskier because, as Kirk mentioned earlier, open
AI is going to account for a larger portion of
the portfolio and ARM is going to come down to
about forty percent. The company has been selling stakes in
mature listed stocks like T Mobile and Deutsche Telecom. Now,
from an equity perspective, it sounds positive, but from a
credit perspective, we view those as more stable, mature listed
(11:42):
companies that they can monetize if needed. Lasting to highlight
is the declining portion of listed stocks. So back in
December twenty twenty four, it was about seventy eight percent.
September we've reached below seventy percent, and we think it
could trend lower towards sixty percent, but it's hard to tell.
You know, at what point this could trigger a downgrade,
(12:03):
but the rating is still unstable outlook at this moment.
Speaker 1 (12:06):
So you mentioned that SoftBank has pledged, especially with its
ARM holdings, as collateral for these loans. So if the
ARMS share price falls, then they'll have to pretty much
you know, give you know, the lens. It's more money, right.
Speaker 3 (12:19):
So the terms haven't been fully disclosed, but the way
it works is that if there's a significant fall in
ARMS share price, then the company may have to repay
the loans or pledge some there's cash collateral pledging needs
having said that based on our calculation, the loan to
(12:40):
value for the ARM margin loan is still fairly low
at twenty percent or below, So I think the company
still has headroom. The only complication we see is that
if indeed, in a downside scenario, something goes probably wrong
with the Open AI investment, or indeed there's an AI
bubble and it bursts and the company needs to sell
more shares than you know what the T mobile and
(13:03):
other tech stocks that I've mentioned. It can sell some
of the ARM shares because you know, still about twenty
percent is unpledged, but then there's limited free float with
SoftBank still owning eighty seven percent of ARM, So you know,
we could see a scenario where if they're forced to
sell ARM shares, arms share price could tank, unless you know,
(13:24):
it brings in a strategic investor.
Speaker 1 (13:26):
So Sharon and A're putting all this together, like what
are the risks for credit investors?
Speaker 3 (13:33):
So I think the risk is obviously SoftBank's debt is rising,
and the reason it's rising is because of all these
investments in unlisted companies. I mean, you mentioned Opening I
already has thirty billion dollars of revenue, but the valuations
seem pretty lofty, and there's competition from Gemini, so the
(13:53):
dead levels are rising and that just leaves less headroom
for the company to whether you know, potential market downturn
or something going wrong with their investments.
Speaker 1 (14:05):
Asia Centric is produced by Bloomberg Intelligence, where more than
five hundred experienced analysts and strategists work around the clock
to bring you timely, world class research. Our coverage spans
two hundred market industries, currencies, commodities, and industries, as well
as over two thousand equities and credits. To learn more
about Bloomberg Intelligence, visit bi go on the Bloomberg terminal.
(14:28):
If you like what you're here, don't forget to subscribe
and share. Kirk, is this a case of history repeating
itself because Massason is famous for these big bets that
have been wildly successful, but it's also made huge losses
as well.
Speaker 2 (14:43):
I mean, if you go back and look at you
soft Bank share Pricey's history, there is a definite pattern.
I mean, you saw way back in the late nineties
and two thousand when you first invested in Yahoo in
the US, and for a very short time he's the
richest man in the world, and then lost ninety nine
percent of when they bought Vodafone Japan in two thousand
(15:04):
and six, right before the Great Financial Crisis. They borrowed
heavily to do that. They're leveraged up six and a
half times net debt to EBIT down, which in today's
soft bank environment doesn't seem like a lot, but for
you know, the telco world at the time, it was
pretty crazy, you know. And then you had it again
around the COVID crisis and quantitative easing, you know where
(15:25):
they went massively in on Vision Fund. As the first
Vision Fund started to see companies, you know, coming to
market and surfacing value. They invested fifty billion dollars over
two years at the top of the market, which is
something that they're just now getting out from under. So
when the company decides there is a particular strategy they
(15:48):
want to follow, and I should say, when credit markets
are open, you know, they will go heavy. And this
is what we're seeing again.
Speaker 1 (15:57):
And Vision Fund is still loss making right well.
Speaker 2 (16:00):
This last quarter like it moved back on a cumulative basis,
it moved back to profitability, and the headline was the
games in Open Ai because some of that stake is
being held a vision fund two, but even if you'd
stripped that out, they would be back to break even,
which was a very good result.
Speaker 1 (16:18):
So you lived in Japan for many years, I think
it was like sixteen years. What's his perception as an
investor in Japan? Because, as you mentioned, it's had some
real big home runs, but he's also had some you know,
spectacular fails as well.
Speaker 2 (16:31):
Yeah, I mean it's SoftBank's very unique, like within the
Japanese sort of investing environment, because what it is is
it to play on like a lot of the foreign
tech themes always has been if you look at their
asset value, the only asset of real sort of substance
that they have a soft bank the telecom company. But
everything else that they've done is outside. So when investors
(16:54):
look at it, they go, oh, this is a way
we can play AI. You know that domestically, I mean,
there's it's not the only way. But when you're talking
about some of these sort of high profile you know,
global names like open AI and video whatnot software, could
you access to that?
Speaker 3 (17:12):
Okay?
Speaker 1 (17:12):
So if we look back in Son's history, what score
card would you give him as an investor, Because, like
he said, some spectacular home runs like Ali Baba. He
made I think over seventy billion dollars in Ali Baba.
He's also made a lot of money in Yahoo. Originally,
he's had some like, you know, spectacular losses, like obviously
we work as one that comes to mind. Ft X
(17:35):
is another. Yes, I forgot about that Green seal as well.
Speaker 2 (17:39):
Yeah, I mean I would give him you know, a
B A B minus maybe because obviously every time you
look at SoftBank, you know, when you go through these periods,
it's that those bubbles are high, are those asset valuations
are higher. So he is building something, but there have
been some pretty incredible losses and also lost opportunities.
Speaker 1 (18:03):
Referring to the in Video stake where he sold too
early in the first round.
Speaker 2 (18:07):
Right, there's actually three or four different ways to look
at what Softwaank's done with in Video. The first thing
is that they bought it for Vision Fund and over
a year they doubled their money and then they sold it.
And this is the one where you know, he's like,
I'm crying that I sold the shares. It was you know,
terrible because they would have owned like I don't even
remember what the number would have been.
Speaker 1 (18:28):
So he sold up before in Video had that most
recent share price r This.
Speaker 2 (18:33):
Was back in twenty seventeen and eighteen, and it was
before like Generitive AI was on anyone's radar screens. It
was just an investment. And at the time they doubled
their money. And this was at the end of twenty seventeen.
So we're going into the first quart of twenty eighteen,
in Vidia falls fifty percent in the quarter and everyone's like, oh, well,
(18:54):
you know, they made this beat on in Video for
the new Vision Fund. Doesn't look like it's working out,
And it turns out he'd locked in the games with
you know, derivative transactions or whatnot. And that was the thing.
Like at the time, Softbanks sold the in Video shares
because they wanted to make a point that Vision Fund
is going to be a money maker, and so they
(19:14):
accomplished that. Everyone was surprised. They're like, oh, that was
pretty smart trading. Even though now that the stock is
up like twenty times since then, you know, water under
the bridge. But there's other things like they were going
to sell Armed to Nvidiot one point and if they
had done that, they would have picked up twenty one
billion dollars of in video shares, and so this isn't
(19:36):
his fault because he wanted the deal to happen, but
it wouldn't. But that steak, you know, would be worth
twelve times as much, so it'd be worth like two
hundred and eighty billion or something. After that deal collapsed,
they bought a small token position. This is the only
time he's done fantastically well within video shares because they
bought like one hundred million dollars and that ended up
(19:56):
being worth one point four billion. This was part of
the stake that they just old last quarter. They made
huge investment to up that position earlier this year. So
his track record is he's made a little bit of
money on in video over time, but when he could
have really transformed SoftBank, he wasn't in video shows.
Speaker 1 (20:18):
Right, But you have to say that, look, you know,
before we recorded this podcast, I looked at the Bloomberg
Rich Index and he's still worth over thirty billion dollars. Sure,
so he's still one of the most richest men in
the world. Having said that, he was worth over seventy
billion dollars in nineteen ninety nine. So it depends on
which way you look at it right.
Speaker 2 (20:39):
Yeah, I would say, like when I think about I
mean this question about whether Massa is a good investor
or not. The one thing that he does have in
his favor that he's absolutely fearless, you know, and he's
willing to stick by the conviction of his ideas and
borrow like just incredible amounts of money to do it.
So it does show that he believes in what he's doing,
(20:59):
and maybe that's it's more important even if some of
the investments haven't done as well as he would have lied.
Speaker 1 (21:05):
And Sharon, you speak to a lot of credit investors.
What are the major risks and perceptions for soft Bank
and Massa right now?
Speaker 3 (21:13):
I think there's a lot of concern from the fundraising perspective,
where are they going to get all this money from?
What are they going to sell? And for bond investors,
particularly in the US dollar europe on market, there's a
concern that the supply risk is going to lead the
bonds to underperform. On top of that, obviously there's less
concern about rating pressure at this moment, but I would
(21:35):
say there is concern that the debt is going higher
and to what level? Right? So the key question I
keep getting is is there an AI bubble? Which I'm
not sure I'm in a position to answer. But you know,
from a credit perspective, we just look at how much
headroom the company has, right, and we've done some scenario analysis,
(21:58):
and we think the company can actually use asset sales
to help manage the loan to value even if asset
values drop by So if it drops by thirty percent,
they might have to sell twenty billion dollars. That's like
a really downside scenario that you know, we're not saying
that's the base case at all.
Speaker 1 (22:13):
There's been a lot of you know, public criticisms of
the circular financing of a lot of the AI investments,
and soft Bank has been you know, obviously part of
this with its Stargate investment. Is this a risk?
Speaker 3 (22:27):
From my perspective, I think it's a risk more from
open ai being central to a lot of these circular transactions,
right because open ai has been working with Oracle and
VideA and obviously soft Bank, and these sort of transactions
where open a I agrees to lease these data centers
ends up pumping the valuation for all these AI related
(22:48):
stocks higher. But from a credit investor, perspective other than
you know, the potential valuation dropped for the portfolio. The
other concern that I didn't mention earlier is whether open
Ai will need to raise more funds in the future
because the company is expected to continue bleeding cash. And
what sort of role is soft Bank going to play
in that? Because SoftBank has already invested over thirty billion
(23:10):
dollars in open Ai, and if this is their big bet,
is there a risk that you know, Masayoshi Son likes
to go big, is there a risk that we'll see
more investments in the future.
Speaker 2 (23:22):
Yeah. I mean it's interesting because soft Bank is usually
not in this conversation globally about like the circular funding
and everything in Ai, but I think it's absolutely a
part of open Ai. I mean, they're investing thirty billion dollars.
Open Ai is going to use that money to lease
data centers that soft Bank is building through Stargate in
open Ai is also building using the money that SoftBank
(23:45):
invested in them. So all of this sort of like
data center strategy is supported by soft Bank doing the
open Ai investment. And then on top of that, soft
Bank has committed to spending three billion dollars a year
on open ai services in Japan, which you know, when
you think about like the size of that one those
(24:06):
first announced like last year when they said that they're
going to do this with open i, they were only
generating three point seven billion revenue in total. So it's
a lot of what got the open Ai valuation or
business kicked off earlier this year was because SoftBank was
willing to stand behind them.
Speaker 1 (24:26):
And Kirk before we let you go, I wanted to
ask about the valuations of SoftBank. Now, SoftBank is an
investment firm. It has these huge stakes in these both
listed and unlisted companies. Is the share price reflecting all
its investments, Like does a trade at a premium or
a discount too? You know what analysts called the sum
of the parts?
Speaker 2 (24:47):
Yeah, I mean it right now, it's at about a
thirty percent discount to net asset value, which is measured
on the latest public pricing for their liquid assets. And
you know what the carry value is for the private assets.
Back in the end of October, when SoftBank shows got
up to twenty seven thousand, it was trading at a premium,
(25:07):
which is basically the market saying we think open ai
is worth more. But a lot of that has come off.
And another way to look at it is if you
look at the history of the discount and we've started
measuring it since Ali Baba went public back in twenty fourteen,
and basically the history throughout this entire time, it's only
(25:27):
been a premium twice and the most recent was in October.
On average, SoftBank trades are the thirty to forty percent discount.
So at this level right now where we're at, if
investor perceptions of the assets don't change, then it's probably
where it should be.
Speaker 3 (25:45):
Sharon.
Speaker 1 (25:46):
As I mentioned earlier, the credit to fault swaps or
CDs for SoftBank Group has jumped up to almost three hundred.
What is that signaling to investors?
Speaker 3 (25:55):
So it's signaling all the caution I was talking earlier
about rising debt funded investments. It's up from low two
hundreds in early October. But back in twenty twenty two,
which is the last time you know, we saw SoftBank
had these large investments and at that time it was
in Vision Fund two, the CDs actually rose to over
(26:18):
five hundred basis points. So I think the market is
still giving the company some benefit because it has a
strong tread record of managing its balance sheet and its
credit risk. So back in twenty twenty two, what we
saw was the company essentially ended up selling down at
stake in Ali Baba via prepaid forward contracts to help
manage loan to value. So I think that, you know,
(26:40):
we probably see a repeat of this, and the company
is very committed to you know, having loan to value
below twenty five percent.
Speaker 1 (26:47):
And for the benefit of the listener, the credit to
vault swap is a gauge on whether the company could
go bankrupt or could default on its loans, and higher
than number one, more likely it will be right.
Speaker 3 (26:59):
Yes, it's a match default risk.
Speaker 1 (27:02):
This has been a really fascinating discussion and I think
we're in for a wild ride for shareholders and credit
investors on Softbak Kirk Sharon, thanks.
Speaker 2 (27:10):
For joining, Thanks for having me.
Speaker 3 (27:12):
Thanks John, you've.
Speaker 1 (27:14):
Been listening to Asia Centric from Bloomberg Intelligence. I'm John
Lee in Hong Kong. You can listen to all our
episodes on Apple Podcasts, Spotify or review Listen and this
podcast was produced and edited by Clara Chen. Thanks for listening.
Hosts And Creators
John Lee
Tom Corbett
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!