Subscribe
Share
Share
Embed
In Mizuho | Greenhill’s Beyond The Obvious podcast channel, we uncover the value that others miss.
Our podcast is a source for the latest discussions on topics related to capital markets, dealmaking activity, business leadership and more.
Delve into insights from our investment & corporate banking thought leaders to hear their unique perspectives on current trends and market influences.
Learn how industry icons and influential figures began their journeys, overcame adversity and rose to success.
Discover insights that look deeper.
Podcast and Video Recording Disclaimer
This podcast or video recording is for informational purposes only and should not be copied, distributed, published, or reproduced in whole or in part without permission. The information contained in this podcast or video recording does not constitute research or recommendation from any Mizuho entity to the listener or viewer. Neither Mizuho nor any of its affiliates make any representation or warranty as to the accuracy or the completeness of the statements or any of the information contained in this podcast, and any liability therefor, including in respect of direct or indirect or consequential losses or damage, is expressly disclaimed.
The views expressed in this podcast or video recording are not necessarily those of Mizuho, and Mizuho is not providing any financial, economic, legal, accounting, or tax advice, or recommendations in this podcast. In addition, the receipt of the podcast by any listener is not to be taken as the giving of investment advice by Mizuho, nor to constitute that person a client of any Mizuho entity.
For additional disclaimers and regulatory disclosures, please visit Mizuhogroup.com/americas.
00:00:25:21 - 00:00:27:12
Hi, this is Jim Boland.
00:00:27:16 - 00:00:34:07
I work on the Leveraged Capital Markets desk at Mizuho, and this is Mizuho Markets Mindset.
00:00:34:09 - 00:00:38:21
With me today is Rahul Shah, who runs
our project finance business at Mizuho,
00:00:38:21 - 00:00:44:19
and Nick Brice, who runs our high yield sales
and trading business at Mizuho.
00:00:44:21 - 00:00:48:21
What we wanted to talk to you about today is the amount of activity that's going on
00:00:48:21 - 00:00:51:14
in the HPC data center market.
00:00:51:14 - 00:00:56:21
Rahul, as you've seen the expansion
in the project finance business
00:00:56:21 - 00:01:01:21
around these data centers,
you obviously had a very busy 2025,
00:01:01:21 - 00:01:04:10
what are you seeing in 2026?
00:01:04:10 - 00:01:07:15
You hear a lot of stories around –
there's been an AI bubble –
00:01:07:15 - 00:01:12:15
there's a lot more financing coming in 2026 alone.
00:01:12:17 - 00:01:17:08
What are the sentiments you're seeing in the project finance market?
00:01:17:10 - 00:01:22:08
And what, if anything, are you seeing changing as a result of that?
00:01:22:08 - 00:01:23:12
Yeah, sure. Thanks, Jim.
00:01:23:12 - 00:01:28:04
So, 2025 was a record year for the project finance market by every measure.
00:01:28:06 - 00:01:32:15
There was over $225 billion of volume
that came to market.
00:01:32:17 - 00:01:34:13
Three things really drove that.
00:01:34:15 - 00:01:36:06
Number one, renewables.
00:01:36:08 - 00:01:41:22
There was just a large push to renewables activity before the tax credit cliff really hits.
00:01:41:22 - 00:01:49:11
The second was LNG being back in play after being on the sideline and on pause for 2024.
00:01:49:13 - 00:01:54:12
But the third real driver and the biggest story to the market was data centers.
00:01:54:14 - 00:01:57:16
Deals grew to the point that the
largest deal that came to market
00:01:57:16 - 00:02:01:16
and the largest project finance deal ever
was a data center deal.
00:02:01:18 - 00:02:04:20
The number of lenders active in the market grew,
00:02:04:20 - 00:02:07:23
the number of deals that came to
market grew, deal size grew.
00:02:08:00 - 00:02:15:11
Everything's just grown to, just, numbers that no one expected seeing 12, 24, 36 months ago.
00:02:15:13 - 00:02:19:02
So with that backdrop and thinking
about what we see going forward,
00:02:19:02 - 00:02:22:15
one absolute issue or challenge for a lot of
00:02:22:15 - 00:02:26:06
the market right now is really just
hitting the point of saturation.
00:02:26:08 - 00:02:28:22
There's never been this much volume before.
00:02:29:00 - 00:02:32:11
And if you think about many of the
project finance players,
00:02:32:11 - 00:02:36:17
some of them, like us, are highly diversified, where we've been active in renewables,
00:02:36:17 - 00:02:39:07
active in energy, active in data centers.
00:02:39:09 - 00:02:42:22
A lot of the other players out there are
only playing in data centers.
00:02:43:00 - 00:02:47:05
And so they're really grappling with concentration limits really to a subset of
00:02:47:05 - 00:02:53:01
hyperscale tenants who don't have all the same rating profile and some are stronger than others.
00:02:53:01 - 00:02:57:05
And so, looking into 2026, the pipeline is growing.
00:02:57:10 - 00:02:59:22
We expect to see a lot more volume,
00:02:59:22 - 00:03:06:05
but the supply of dollars available from banks is definitely not growing at the same clip.
00:03:06:07 - 00:03:11:02
And so what we have started to see and continue to expect to see next year
00:03:11:02 - 00:03:13:23
will be really alternative forms of capital.
00:03:14:01 - 00:03:18:18
So, going beyond the traditional bank market, reaching into the capital markets,
00:03:18:18 - 00:03:21:06
reaching into other pockets of
capital that are out there
00:03:21:06 - 00:03:24:04
to keep up with the amount of supply
that's coming through.
00:03:24:06 - 00:03:30:13
Now you've seen some migration into the private placement market, into the ABS market.
00:03:30:15 - 00:03:37:00
At this stage, at least at the construction phase, is that normal from what you're seeing historically?
00:03:37:00 - 00:03:43:13
It is and it's a good point because so much of what the banks are really focused on is capital recycling.
00:03:43:15 - 00:03:49:16
So, banks will come into a deal, they want to see that deal or that asset stabilized or completed,
00:03:49:16 - 00:03:56:19
and at that point, that deal gets refinanced in the ABS market, the private placement market.
00:03:56:21 - 00:03:58:02
If you take a step back and look at the
00:03:58:02 - 00:04:02:21
amount of volume that we've seen over the
past 12 months, 24 months, 36 months,
00:04:02:21 - 00:04:07:06
that recycling really hasn't hit the way that a lot of us want to see,
00:04:07:06 - 00:04:11:03
and it's not because projects are having issues
or facing challenges as
00:04:11:03 - 00:04:16:18
much as it is that there's been a really steep ramp, and while those markets are open,
00:04:16:18 - 00:04:21:03
there's just such a stronger pace of new deals coming to market than deals,
00:04:21:03 - 00:04:24:03
or new deals coming to market than those
deals being taken out.
00:04:24:06 - 00:04:24:20
Got it.
00:04:24:21 - 00:04:28:19
Now I'm going to hit Nick on this in a moment,
but I wonder if you could tell us,
00:04:28:19 - 00:04:35:22
we've seen some, I'd say, call them non-traditional type developers that have a history in the Bitcoin market
00:04:35:22 - 00:04:38:15
that have access to the high yield market.
00:04:38:19 - 00:04:43:01
Was that a selection of efficiency
and speed to market,
00:04:43:01 - 00:04:46:23
or did they look toward the project finance market
00:04:46:23 - 00:04:49:19
and just passed and opted to go into
the high yield market?
00:04:49:23 - 00:04:50:17
Yeah, it's a good question.
00:04:50:17 - 00:04:54:15
We've spent a lot of time on those situations,
and the answer is both.
00:04:54:18 - 00:04:58:10
If you think about the traditional
project finance market,
00:04:58:10 - 00:05:02:20
we look for investment grade offtakes, investment grade contracts
00:05:02:20 - 00:05:06:01
that are really underpinning the credit
profile of those deals,
00:05:06:01 - 00:05:10:10
and so when you think about the players
that you just mentioned with that profile,
00:05:10:10 - 00:05:16:17
they often don't meet or fit the mold of what the traditional bank market is really willing to do.
00:05:16:19 - 00:05:20:08
Now, there is voracious appetite in the
private credit market,
00:05:20:08 - 00:05:24:04
and so, many of those issuers have also
00:05:24:04 - 00:05:28:02
tried going down that path, but ultimately
steered to the capital markets
00:05:28:02 - 00:05:31:07
just because it felt like a more efficient
execution for them.
00:05:31:09 - 00:05:32:00
Got it.
00:05:32:03 - 00:05:36:23
And Nick, on that point, just in the
fourth quarter alone,
00:05:36:23 - 00:05:43:01
you saw just shy of $10 billion coming
from that universe.
00:05:43:01 - 00:05:46:17
What's been the reception so far
from high yield investors?
00:05:46:17 - 00:05:48:01
The reception has been great.
00:05:48:01 - 00:05:54:02
Our clients, they love the ability
to put net new money to work,
00:05:54:02 - 00:06:02:17
and data center financings have been and will continue to be the biggest source of net supply
00:06:02:17 - 00:06:09:22
for the credit markets, and so our investors are excited to have new credits to look at
00:06:09:22 - 00:06:16:22
and to gain exposure to artificial intelligence and high-performance compute data centers.
00:06:17:00 - 00:06:21:08
Now, construction, historically, it's been done before in high yield,
00:06:21:08 - 00:06:25:00
a ton of construction financing
coming into the high yield market.
00:06:25:00 - 00:06:28:06
How do you think that's going to be received?
00:06:28:06 - 00:06:35:19
So far, it's been received well, mostly because
of the tenant quality.
00:06:35:21 - 00:06:40:16
For the most part, these are hyperscaler IG counterparties.
00:06:40:16 - 00:06:44:06
We’ve taken a lot of the characteristics of
Rahul's market with
00:06:44:06 - 00:06:50:12
amortization, excess cash sweeps, and
tenant backstops, and security,
00:06:50:12 - 00:06:57:06
five-year tenors and BB ratings to make sure
our investors can get
00:06:57:06 - 00:07:01:11
comfortable with the construction risk
and some of the execution risk
00:07:01:11 - 00:07:04:19
around a relatively new asset class.
00:07:04:19 - 00:07:08:20
And any points of caution you're
seeing with investors?
00:07:08:20 - 00:07:11:06
Yes. Similar to what's been mentioned,
00:07:11:06 - 00:07:20:09
but construction risk, access to power, and
operator and tenant quality.
00:07:20:11 - 00:07:29:05
Standing up and operating an
HPC data center is different than
00:07:29:05 - 00:07:35:04
your legacy data centers that were built
in the last couple of decades.
00:07:35:06 - 00:07:39:15
So, certainly teams with more
experience are preferred.
00:07:39:20 - 00:07:40:04
Got it.
00:07:40:04 - 00:07:44:02
And when you think about the high yield market last year being about
00:07:44:02 - 00:07:49:21
$325 billion in total volume, and you hear numbers thrown around anywhere from $40
00:07:49:21 - 00:07:55:23
to $60 billion of potential data center financings, and that's just data centers
00:07:55:23 - 00:08:00:02
themselves away from the other parts of the value chain, whether it's the GPU
00:08:00:02 - 00:08:04:16
financings or just the power itself, how do you think the high yield market's
00:08:04:16 - 00:08:05:17
going to absorb that?
00:08:05:17 - 00:08:08:05
I think there'll be the demand there for it.
00:08:08:07 - 00:08:13:19
I think the credit market in general
right now is very healthy.
00:08:13:21 - 00:08:17:23
The recent deals multiple times oversubscribed,
00:08:17:23 - 00:08:24:05
and we have all of our blue-chip long-only asset managers, insurance companies, pension funds
00:08:24:05 - 00:08:29:03
and hedge funds are all important participants
00:08:29:03 - 00:08:35:00
and have bought sizable portions of the deals that have come to market so far.
00:08:35:00 - 00:08:40:20
And as long as we have the characteristics, the structural enhancements that we talked about
00:08:40:20 - 00:08:48:16
in place, and the tenant quality is there, then
there should be plenty of demand.
00:08:48:16 - 00:08:54:03
So would you say that the majority of these books are broad-based investors,
00:08:54:03 - 00:08:58:14
not just segmented towards hedge funds or
just traditional high yield?
00:08:58:17 - 00:09:03:00
It depends a little bit on the quality of the deal.
00:09:03:03 - 00:09:12:07
So far, we've seen mostly BB, seven-handle, high-quality deals come to market.
00:09:12:09 - 00:09:18:06
And in that ballpark, you can attract
00:09:18:06 - 00:09:23:15
investors from across the spectrum and you should have plenty of demand.
00:09:23:20 - 00:09:24:05
Got it.
00:09:24:05 - 00:09:30:02
And as you get away from the hyperscaler community into sort of lower-quality guys,
00:09:30:02 - 00:09:33:18
how do you think the market reacts there?
00:09:33:20 - 00:09:35:16
That'll be the real test.
00:09:35:18 - 00:09:38:20
That's probably the open-ended
question at this point.
00:09:38:20 - 00:09:42:03
You'll probably have to get a little more creative with the structural enhancements,
00:09:42:03 - 00:09:51:07
because when we start talking about higher-risk investments in the AI ecosystem,
00:09:51:07 - 00:09:57:10
credit is not exactly how you want to express
those types of trades.
00:09:57:10 - 00:10:02:06
You want more equity-like returns
for that type of risk.
00:10:02:08 - 00:10:08:03
And high yield bonds, unless they come very wide with bells and whistles,
00:10:08:03 - 00:10:10:17
don't usually offer equity-like returns.
00:10:10:17 - 00:10:11:17
Got it.
00:10:11:19 - 00:10:13:18
Going back to Rahul, what can you tell us about 2026, your pipeline?
00:10:13:18 - 00:10:18:07
Going back to Rahul, what can you tell us about 2026, your pipeline?
00:10:18:09 - 00:10:19:17
Is it manageable?
00:10:19:21 - 00:10:21:15
Are you turning business away?
00:10:21:15 - 00:10:22:17
Yeah, the pipeline's growing.
00:10:22:17 - 00:10:30:00
I continue to be surprised by just the intensity of inbounds for new projects and opportunities.
00:10:30:02 - 00:10:34:11
So I take a step back, I think about AI
and just data centers overall.
00:10:34:13 - 00:10:37:20
No doubt, volume will be up year over year.
00:10:37:22 - 00:10:43:07
Whether or not the banks can keep up with that volume, I think the answer is probably not.
00:10:43:07 - 00:10:47:21
I think we will have to explore alternatives
to the bank market,
00:10:47:21 - 00:10:52:05
just because there's so much supply out there, based on what we see coming.
00:10:52:07 - 00:10:58:14
From a structure perspective, these deals have really stayed at pricing terms.
00:10:58:17 - 00:11:00:11
Nothing has really changed.
00:11:00:12 - 00:11:05:13
Because there's too much supply for anyone to really try to get too aggressive on things.
00:11:05:14 - 00:11:07:09
And I think that'll stay the same as well.
00:11:07:09 - 00:11:13:19
I think about this space two years ago, many people thought of it as "there's winners and losers".
00:11:13:19 - 00:11:15:17
The winners are the folks active in data centers.
00:11:15:20 - 00:11:19:00
The losers are the folks who are
watching it from the sideline.
00:11:19:02 - 00:11:21:23
And when I think about the space today,
00:11:21:23 - 00:11:25:21
I think there will be winners and losers
who are active in data centers.
00:11:25:23 - 00:11:29:11
What I mean by that is some projects
will hit delays.
00:11:29:13 - 00:11:34:01
There will be construction issues, there will be supply chain issues, things like that.
00:11:34:03 - 00:11:38:00
The underlying credit profiles are strong,
and so the structures are tight.
00:11:38:02 - 00:11:44:18
But I think we will see delays on certain projects, and that's just ordinary course for project finance.
00:11:44:18 - 00:11:49:08
And how do you think about the
operator, developer, lessor,
00:11:49:08 - 00:11:52:16
and your decision making around
which deals you chase?
00:11:52:18 - 00:11:56:08
Proven track record goes a long way.
00:11:56:10 - 00:12:03:10
Having a demonstrated ability to work well
with a number of hyperscalers,
00:12:03:10 - 00:12:08:02
longstanding relationships, things like that matter.
00:12:08:02 - 00:12:12:00
Because when things go sideways or
when you hit challenges,
00:12:12:00 - 00:12:15:20
the strength of those relationships I think really make the difference in projects.
00:12:15:22 - 00:12:20:12
And so that has always been a part of
our calculus in terms of deal selection.
00:12:20:14 - 00:12:27:02
And I think the market in many ways is starting to think about that the same way.
00:12:27:04 - 00:12:31:00
I think about the last 18 months and a lot
of new entrants in the space
00:12:31:00 - 00:12:36:03
were trying to credentialize themselves to be able to build some creds,
00:12:36:03 - 00:12:41:08
show that they're able to deliver on these deals, that they can structure them.
00:12:41:10 - 00:12:44:01
There's a lot of new players who
have been able to do that.
00:12:44:01 - 00:12:47:13
And I think we'll see more discipline around
deal selection going forward
00:12:47:13 - 00:12:50:03
to capture exactly those factors.
00:12:50:03 - 00:12:55:09
When you look at the private credit community, a lot of them manage insurance company money
00:12:55:09 - 00:12:58:22
who can be very competitive from a pricing standpoint.
00:12:59:00 - 00:13:04:22
Are they winning the battle, losing the battle when you guys compete against them?
00:13:05:00 - 00:13:07:14
Because there's so much supply right now,
00:13:07:14 - 00:13:12:13
it hasn't been a matter of us losing to those players or those players losing to us.
00:13:12:15 - 00:13:15:17
It feels like there's enough for everyone.
00:13:15:19 - 00:13:17:18
Might that change over time?
00:13:17:20 - 00:13:18:18
Perhaps.
00:13:18:23 - 00:13:22:17
Does that change in 2026 based on
what we see in the pipeline so far?
00:13:22:17 - 00:13:23:19
I don't think so.
00:13:23:19 - 00:13:24:10
Okay, great.
00:13:24:15 - 00:13:28:00
With that, I'd like to wrap it up and thank you for joining us today.
00:13:28:02 - 00:13:31:01
And I'd like to thank Rahul Shah and Nick Brice
00:13:31:01 - 00:13:36:15
for joining us for this very important discussion on the data center market.
00:13:36:17 - 00:13:37:01
Thank you.