September 3, 2025 • 38 mins
In this episode, Philbert Shih, Founder and Managing Director of Structure Research, offers a deep dive into the evolving landscape of digital infrastructure through the lens of market intelligence and industry specialization. From launching his boutique firm in 2011 to supporting over $3 billion in M&A transactions, Phil shares insights into the macro trends reshaping the sector, including the rise of hyperscale cloud, the real-world impact of AI, and the mounting constraints of power, land, and regulatory hurdles. He also previews Structure Research’s annual Infrastructure Summit, emphasizing the 2025 theme of “Concurrence”, a reflection of the many concurrent forces redefining the industry.
Subscribe, leave a review, and stay connected.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:01):
Thanks for joining me on Let's Get Digital.
I'm Carrie Charles, your host, and today I have with me Phil Shee.
He is the founder and managing director of Structure Research.
Phil, thanks for coming on the show.
Thanks for having me.
So we met at ITW, right?
Yup, that wasn't that long ago.
(00:22):
Yeah, it sure wasn't.
Boy, time flies.
uh I know you have an interesting journey, and I'd love to hear a little bit about thefounding story of structure research, how you got to where you are today.
Yeah, mean, the whole story, it's just so much kind of just kind of live in life and thenjust kind of making decisions based on, you know, opportunities and doors that open ah and
(00:48):
kind of getting to the point in life where you thought, hey, maybe there's a better way todo this uh and kind of saying, hey, I got to take the leap.
If I'm ever going to do it, I got to do it now.
It was kind of like uh that kind of thinking, you know, and like you, I think many yearsago, you just said, hey,
What would it be like to not have a boss and to drive a lot of the decisions that Iincreasingly am not agreeing with in the firm I'm at?
(01:15):
I don't think that's an unusual, you know, that's not a unique story.
But honestly, that's what drove it.
It's like that question is like, hey, what if I did, you know, have the control and theability to make decisions and drive this in the way I want, but that's going to take a
leap as in, you know, saying goodbye to a regular paycheck.
predictability and the comfort and trying something completely different, which, you know,I think a lot of entrepreneurs or people who've gone on their own did not necessarily have
(01:42):
that in the plans, have that in store.
Again, it's about opportunity and, know, the field I was in, or the field that we are in,this industry, you know, was healthy then.
I saw, I think I saw the upside as many of us did.
And I thought, well, if you put those things together, wow, maybe this could work.
And yeah, I'm glad, I'm glad I made the leap.
was the
(02:03):
you know, something that I didn't expect, but I'm glad it's working out and uh I'mgrateful for the opportunity that the industry has provided.
How many years ago was that?
Yeah, 2011.
So late in 2011, I decided to make the jump and I've been at it ever since.
Okay, well you know what you're doing, so let's get into it.
(02:25):
Well first, tell us more about structure research.
Yeah, in a nutshell, for those of you who aren't familiar, we describe ourselves and thinkof ourselves as a boutique research firm, which is another way of saying we're smaller,
we're slightly understaffed, but we are specialists.
And that's kind of part and parcel of our value proposition.
(02:46):
We have seven analysts that spend all our time in this sector.
When I say this sector, we used to call it internet infrastructure.
But data centers, managed cloud, managed hosting, what used to be called managed hosting.
But the infrastructure services ecosystem, what fits now under the kind of broader rubricof digital infrastructure, we're kind of more focused on the compute-oriented
(03:12):
infrastructure side, rather than say wireless and connectivity.
We're very much, you know, I started my career basically, know, following servercompanies, server hosting companies.
And that's morphed over the years into managed services.
data centers and now hyper scale in AI.
So yeah, the boutique, that's who we are.
We spend all our time publishing industry focused research.
(03:33):
Think of us as living in the kind of middle ground between a Wall Street analyst or astreet analyst that's focused on earnings, the stock, the value of that stock and
recommending it to investor clients.
And then on the other kind of the other side of the spectrum, kind of more technical ortechnology oriented analysts that are focused on
(03:53):
the kind of details and the nuts and bolts of the technology pieces, we're very muchfocused on the industry, kind of macro trends, the landscape, the big picture, helping,
ultimately trying to provide information perspective and insights to help executives makedecisions.
And that's kind of, in a nutshell, who we are.
(04:16):
So you do compete in a space with massive analyst firms.
And I guess I would look at it a little bit like David and Goliath, but what makes yourboutique approach different and also advantageous for your clients?
Yeah, mean, it's again, the specialization kind of creates a certain, you know, certainamount of efficiency for the executive decision maker that subscribes to the service.
(04:42):
You know, they want information that's tailored and curated to their world, their focusand what they care about.
That's our pitch when we meet people.
It's like, hey, if you're spending a lot of time in this sector, if you care about what'sgoing on in this sector, if you have a stake, whether financial or kind of strategic,
ah you know, we could be the right service for you.
(05:03):
We don't kind of try to cover too much ground.
And that's how, you know, honestly, we too much ground in the sense that we don't cover alot of other sectors.
You know, we're not even covering broader digital infrastructure like towers, wires, Imentioned, wireless and connectivity is kind of slightly kind of referral to what we do.
And that's how we exist.
You know, you asked the question, how do we exist against some of the larger firms?
(05:26):
Well, we've selected a sector that we think is valuable.
that is frankly a bit undercover, don't let that get out, and say, hey, we're all aboutthis.
And if this is what you care about, we uh could be of service.
We're not for everybody.
ah And we're pretty comfortable with that.
And again, that's our pitch and that's how we survive.
The bigger firms tend to offer more and continue to offer more.
(05:50):
They're obviously have hundreds, thousands of people on their staff and they can cover alot of ground.
But we give you something very focused.
So we dive into, uh
we're able to provide a lot of depth and we give you a senior analyst perspective.
Everybody in our team, ah whether you're a junior analyst, but you quickly within our kindof environment ramp up pretty quickly because you're just doing this.
(06:16):
You're not focused on anything else.
Okay, that makes sense, and you've been doing it a long time.
Yeah, since 2011.
And kind of just gradually built up the team over the years.
yeah, that's the great world of remote working.
We were a remote team before COVID, the pandemic.
And it's working.
That's great.
(06:37):
That's great.
So you have an event.
I guess it's been going for six years now, right?
It's infrastructure.
um Infra is, uh I guess, small letters, then there's, uh you know, structure is all caps,but it's an infrastructure summit.
It's in Las Vegas, October 15th and 16th.
I know I'm gonna be there.
(06:58):
Can you tell us a little bit about the event?
What makes it different than all the other industry events as well?
So Infrastructure was founded, we had our first event in 2018.
It ah is the sixth edition this year, there were a couple years where we didn't have itdue to that pandemic.
ah And it started uh as and still is meant to be an executive summit ah that is unique inthe sense in that it's very uncommercialized.
(07:27):
So the content is obviously the agenda is backed, supported and led by an analyst teamthat is independent.
that is not tied to, you know, kind of any vendor or platform or investment vehicle.
uh Rather, we bring those parts of the ecosystem to our summit so that people can kind ofnetwork, you know, partner, engage in business development, and of course, discuss kind of
(07:50):
the present and future of the sector.
so research, you know, I believe, and I think our audience values is a really greatbackdrop and context for that discussion.
And so that's what we decided to do in 2018.
We slowly built up the event and you know our mantra for the summit is you know qualityover quantity.
We're not trying to build an expo.
(08:11):
We actually don't really have an exhibit hall.
We make this very clear to our sponsors that want to come.
That if you want to come it's a different type of engagement.
There's obviously room for business development but the way you go about it, the way youengage is a bit different because it's a summit because we're in the participation is in
the hundreds rather than thousands.
(08:31):
Yeah.
No, give me a little bit of a peek into the agenda, maybe some topics that you're going tocover and people who are going to attend maybe.
Yeah, so it's executive heavy.
So VPC level, a lot of owners, lot of entrepreneurs who run their business, a lot of CEOs.
(08:54):
And that's kind of the primary demographic.
uh But we surround them with uh two other major uh groups of people, which are investorsand financiers, or people on the finance side and real estate side of the game, and then
suppliers, people who sell products, services, or resources.
to the operating service providers so that they can build and deliver their services.
(09:17):
Real estate can also fall in that category.
Increasingly, we have energy and then of course software and hardware and constructionservices, commissioning services, various types of vendors that support this ecosystem.
ah So that's the audience and then the agenda uh tries to cater to the questions of theday um that everybody again that has a stake and interest in the sector
(09:43):
um you know, want answers to or want to hear discussion about or hear from, you know,leading minds and, you know, thought leaders in the space.
So, you know, you asked about the agenda this year.
ah You know, we try to find a balance.
So there's going to be some sessions, uh &A, you know, on the strategic landscape, uh bigpicture questions around, you know, the impact of AI, uh the impact of the recent
(10:11):
hyperscale pullbacks.
how that impacts the demand profile of the sector.
We have various topics on energy this year, how are energy constraints uh influencing orconstraining the ability of uh operating companies to continue to build in the face of
significant steam of demand.
And of course, there's always those who are a little skeptical and it's a great balance.
(10:37):
oh This year, we hope to have a lot of discussion.
m
between those who are a bit bearish and those who are a bit bullish.
That seems to come up every kind of cycle or phase in the sector where you see someskeptics rise up, know, and, you know, not without, you know, of legitimate concerns about
where things are going.
(10:58):
But that fuels a great debate and discussion.
And, you know, we love the diversity of opinions and takes.
So your 2025 theme is concurrence and I wanted you to talk more about what that means andwhy is it the right framing for today's infrastructure challenges.
(11:20):
Yeah, every year, so every year we try to pick one word as the theme, right?
So we've had, except for the first year, uh but we've had decentralization to kind ofcapture where the sector was moving a bit to the edge, uh acceleration coming out of the
pandemic in terms of kind of where the growth was.
(11:41):
And this year we picked concurrence, which is meant to capture kind of just the manymoving parts in this space.
Right.
So I did have to look up the dictionary definition of concurrence and it is on thewebsite.
kind of cut that out and put it in there.
Thanks, Merriam Webster.
But, you know, just try to capture kind of where the sector is in terms of how it'spreparing and engaging in building for this, you know, next wave that's coming.
(12:10):
And so a lot of moving parts, a lot of different things happening at the same time.
That's kind of what we're trying to capture, you know.
We're built, hyperscale is still, hyperscale cloud is still driving things.
AI is on the horizon.
Power constraints are developing.
How, you know, uh that opens up kind of a different regulatory picture.
uh Capital is flowing into the sector.
(12:31):
How do all these things moving kind of at the same time and kind of changing the waythings are done?
You know, how does that, you how do we make sense of that?
How does the sector organize around that to push forward?
So.
Yeah, um hopefully that makes sense.
We do try to be little edgy and different, but at the same time, capture what is the mostpressing dynamics in this space.
(13:00):
And that's what the concurrence is about.
So with over 20 years covering global infrastructure services, what have been the mostimpactful market shifts that you've observed?
Yeah, I mean, I think quite clearly the biggest shift was, you know, probably the, let'sjust agree the sector was born in the late 90s, early 2000s.
(13:24):
ah So the big shifts, I think, have been pretty clear that, you know, for those of you whoare a little older and remember the dot-com bubble, ah seems like, you know, it's almost
25 years ago now, quarter century.
ah That would be kind of the first thing, you know, one of the
(13:44):
the first big seismic events that kind of gave birth in many ways to the sector as a lotof operating companies kind of built out of the dregs of what happened.
And the sector was so young, a lot of assets were sold pennies on the dollar and the valueproposition of the space kind of took advantage of that situation to kind of build.
(14:05):
And then in 2000, the next kind of big development, 2006, AWS, Amazon decides to get intothe cloud business back then.
strictly known as an online book retailer, really not even, what did you buy from Amazonin 2006 besides books?
Nothing.
They got into the cloud business, which was a curiosity at the time, but set off a decadeplus move from kind of traditional infrastructure, which was still young and still
(14:33):
developing.
uh We didn't call it cloud then, ah but was still developing.
then at the same time, this
kind of hyperscale, what we now call hyperscale cloud, public cloud, was developing in thebackground, started to increasingly kind of become more prominent and now arguably, uh you
(14:55):
know, is probably the dominant kind of computing platform of choice for, or not to sayit's the only one, but it's definitely, you know, front and center and controls a
significant part of this ecosystem.
And so our sector has transitioned and built around that.
data centers, managed services, MSPs are all part of that.
And of course, there's still a lot of diversity in terms of other cloud platforms, butthat's the big one.
(15:20):
That's been a, depending on what part of the world you're sitting in, a decade, decade anda half.
It always amazes me how young some parts of the world, they're hyperscale.
The markets in those countries are so young, maybe just a few years with hyperscale cloud.
So that wave's still ongoing.
And I think the next big one most of us can agree is coming.
(15:42):
ah How it's going to play out, we don't all agree on, but AI and how that is going toaccelerate things is the next big inflection point wave that we'll pick up.
But interestingly, think overlap a bit with hyperscale cloud.
So you've led consulting mandates totaling over three billion in deals.
(16:04):
Is that, am I right?
My reading?
My research?
I think so.
Yeah, if we put it out there, we did a bit of quick math.
eh
Okay, good.
Can you share a standout transaction that really shaped Structure Research's strategicapproach?
Sure, ah so we're not a consulting firm, we're research firm uh at heart.
(16:26):
That's the of the majority of where our business model is built on that.
We do consulting on a strategic basis, on a selective basis, mostly around, as youmentioned, &A transactions.
We don't necessarily lead kind of that process.
We let the consultants do what they're good at.
We come in as the market expert, the subject matter expert, the market data source.
(16:49):
and support transaction activity in that way.
So, you know, we've done small transactions, large-scale ones.
We've done the first one I did under the banner of structural research was actually inEurope.
ah But we've done work in North America and we have quite a specialized practice in AsiaPacific.
ah Our head of research and myself, Hale, from the region still have strong ties there,spend a lot of time there.
(17:13):
And that's actually where the largest transaction we've done, which was the
kind of previous, not the recent one, but the previous investment in Australia's AirTrunk, which is, know, hyperscale data center platform.
uh And, you know, that value is a good chunk of the 3 billion you referenced.
uh But yeah, that's, that's, I mean, that's just one example of things we've done in thepast to support &A in the space.
(17:42):
Do you have proprietary em data sets that differentiate your insights?
Absolutely.
All our market share data, everything we put out there is built from the ground upinternally and organically.
It's a pretty painful, painstaking process.
It's a grind, we describe it, but we feel it pays off and it ensures the integrity of thedata that we put out there.
(18:10):
So yeah, every data set is based on what we feel is...
hard work, getting boots, we like to call it getting boots on the ground, talking tooperators, visiting sites, making sure we're double checking those facts with other
sources.
A lot of what we call triangulation, throwing out the high, throwing out the low, takingdata points, try to compare them and you tend to get closer to the truth if you undertake
(18:36):
that process.
yeah, absolutely everything we put out there, we've built, we own, we stand by it and tryto refine.
I think one of the other things that we, I think we do really well is that we refine overtime.
So we're always kind of, you know, uh we don't think we have the answer, the answers andat any given time.
(18:57):
And, you know, if we see other kind of facts or feedback or insights that tell us thatwe're wrong, we're going to adjust.
And so that, you know, over time we get closer to the truth and that's the approach we'vetaken.
So data center M &A hit record levels in 2024.
I was reading that also.
And what do you think is driving this?
(19:19):
mean, I probably we probably all know that.
But what should infrastructure leaders know about navigating this current M &Aenvironment?
Yeah, I mean, I think what's interesting, I mean, I would say not that we have, I have ahard data point, but my sense definitely is that the number of transactions is lower, but
the value is higher, right?
(19:40):
Is that, that's kind of, if you were measuring like, Hey, there was more value, you know,more transactions it's because the numbers are just so big now.
Um, you know, what would I say in terms of, you know, I
of kind of where we are and where we're going and how would operating executives kind oflook at that, I would say.
I mean, first of all, I think that's one thing that hasn't changed is that the sector hasalways seen consistent and steady strategic activity across the board from kind of early
(20:12):
gen cloud, smaller service providers to the largest operating platforms in the world.
ah I think it's a product of the fact that the sector has always been
you know, very kind of fragmented in its own way across different segments and now isincreasingly global, which has actually caused a cause, you know, led to kind of, you
could argue more fragmentation with regional markets, you know, and global operatingcompanies operating in different regions and countries around the world.
(20:41):
I think, I think that's just kind of part and parcel of how this stuff is built is thatbecause infrastructure is
Not like an iPhone or something.
It's not like a closed walled garden.
It's really aggregation and piecing together of different resources.
And they're done in different ways in different markets in different countries and regionsaround the world.
(21:02):
And so that creates a very heterogeneous environment that's conducive to M &A.
I don't think it's going away.
It goes through waves where there's more activity than not.
But in my experience, it's been steady just for so long.
And I know a lot of my &A and banking contacts are very happy about that, of course.
(21:25):
um But yeah, I would say to operating executives, that it's always something there tosupport strategic activity can always support growth and expansion.
There are great options out there.
um
the value of, you know, the valuation of assets in the space has always been high.
(21:48):
It continues to, you know, to climb depending on what part of the market you're in.
And that would be, I guess, the last kind of, I guess, if I could add a tidbit is tounderstand which segment you're in in the market and to kind of build, evolve and
transition your business accordingly.
Right.
So there are always going to be certain sectors that are, that have, that are healthier,certain sort, certain segments within the overall kind of data center infrastructures.
(22:14):
cloud infrastructure space that are trending in a better direction that have healthierupside.
There are some services now that we call legacy fairly because they are at a differentstage in their evolution.
And so when you're looking at operating assets with your, or an operating side and lookingto exit, or you're on the buying side and looking to build, expand, diversify, augment, uh
(22:38):
those are the, I think that is the most important thing to think about.
And hopefully a research firm can help provide some context to that.
So we can't have a conversation without talking about AI.
So here we go.
And everyone's talking about AI driving massive data center demand.
But you track the actual numbers.
So behind the hype, if you will.
(23:00):
So what does that real data tell us about AI's impact on the sector?
Yeah, it's a crazy time to be an analyst because the short answer to that question is, youknow, materialization of that demand is still very much at a very early stage of
development, right?
(23:21):
So that's not to say that things haven't landed and are running and impacting growth.
It certainly has.
But the the kind of surge or the wave that we expect if this technology continues to movein the direction we see it moving.
We're still probably a few years out, right?
To see that kind of really spike uh or shift in that trajectory on your PowerPoint chartsand everything.
(23:49):
Maybe it will look like a hockey stick, but not right now.
Right now, things are still being driven by what they've been driven by, which is publiccloud, hyperscale cloud.
But you're starting to see these models start to be built.
And then when that happens, there's kind of a spin-off effect across infrastructure.
(24:10):
So the models will be trained and they'll have to be used by the end user to deliver them.
They'll have to build inferencing infrastructure.
And then to maintain fine-tune and perfect those models, you're going to needinfrastructure that sits almost in between where the models are built, the large-scale
models, and then of course the inferencing.
And so that's more infrastructure.
(24:30):
And then you need connectivity to move the data, to access the data, which means moreinterconnection uptake, more network services.
And that spins off into service providers, et cetera, et cetera.
yeah, the way we're modeling it is we're trying, it's not an exact science as you canimagine, but we're trying to kind of predict when in the next few years some of these
(24:55):
things start to come online and that spin-off effect takes shape.
ah and then impacts kind of the entire ecosystem.
So our latest update, our latest numbers, you'll see a meaningful increase in what we'reprojecting for growth, but with a spike not quite coming just yet.
ah And it's very hard to predict and understand what that spike will look like.
(25:17):
So historical precedent is useful, but not always a perfect kind of comp comparison.
Can you just give me a little peek into what art, you're predicting?
Yeah, no, I mean, I think we had the data center co-location space last year growing atabout 13 and a half percent globally.
(25:39):
We now have it closer to 20 % and believe it'll be growing over 20 % in the next few yearsbased on kind of everything we're seeing.
So are we building fast enough and what are the bottlenecks that most people are just notseeing?
(26:00):
Yeah, it's difficult to say if we're or not building fast enough because the way thisdemand will materialize is hard to predict because it's so new.
Hyperscale cloud, if you're building availability zones for traditional cloudinfrastructure, we have enough of a sample size to be able to predict within reason what
(26:25):
it might look like.
are going to look very different.
So yeah, it's hard.
I say it's hard to say, and I'm kind of just, I think I kind of just missed the other partof your question as I was saying that, but right, right, right.
The bottlenecks.
(26:46):
So yeah, so it's difficult to say whether we're building too fast just because we don'tknow what's, we don't have a clear picture yet of what's coming.
But in terms of bottlenecks, absolutely.
I think you can see.
that the most clear characteristic of this next wave is the amount of energy that needs tobe consumed.
(27:08):
And there are increasingly signs that in the most mature markets, ah especially, ofcourse, in the United States, that uh if any of these predictions or projections are even
close, there's going to be significant uh shortage of energy.
And so that's kind of one of the biggest, if not the biggest constraint I would throw inthere as well, uh which related to energy is of course, land in it, real estate.
(27:37):
But there's also a shortage of expertise, right?
you need, uh hyperscale platforms have come to build on their own.
They face their self-build internally and not use data center co-location operators.
They've come across their own constraints and there are only so many
uh credible, uh reputable, established kind of data center operating platforms that canbuild at this level of this scale at the quality that's going to be required.
(28:07):
yeah, land, energy, expertise, and then the regulatory environment as well could be aconstraint given the constraints on energy are going to cause a shift to alternative
sources of energy, uh nuclear being that one that's been discussed quite a bit.
but also comes with a lot of regulatory implications.
(28:28):
yeah, it's a concurrence.
A lot of things happening at the same time that need to be resolved and solved before thesector can move forward.
So I'm a data center, let's say I'm a data center owner operator, what are some key uhdata points that I need to know that will help me navigate all of this, the concurrence?
(28:53):
Yeah, I would say come to the show and hear the discussion.
That's one.
Las Vegas in October is not a bad time of year.
um you know, it's a good question.
There's so many different kind of data points and factoids.
would say if you're an operating executive is to kind of look for honestly, I mean, themarket is healthy, but at the same time, there's a lot of competition, right?
(29:20):
There's a lot of capital flowing into space.
Everybody wants to get involved.
Um, you know, finding a market inefficiency is, you know, if, you're a new operating, youknow, whether you're a new operating platform or an existing one that's looking for other
opportunities, you know, finding where market inefficiencies are, uh, I think is going tobe increasingly important.
Uh, sometimes the industry kind of, I don't want to say, you know, over indexed, maybeover indexes a bit, or there's a lot of opportunity.
(29:47):
And so everyone flocks in one direction and then certain parts get neglected.
Right.
uh Obviously people push in a certain direction because for good reason, uh but in thatprocess, certain things come up, certain areas get slightly neglected.
so I think you're going to start to see, uh think investors, operating companies shouldexplore kind of what we've described as the edge hyperscale market, which is kind of
(30:17):
something that exists between enterprise and
and kind of large scale, hyper scale that everybody has been chasing, availability zones,regions that is a bit underserved, ah but there's no shortage of demand that's starting to
build towards that.
So edge hyper scale, pretty broad term meant to capture kind of secondary, second tiermarkets or even within major markets, but certain type of requirement that's broadly
(30:48):
speaking.
single digit megawatt to 10 or 20 versus the very large deployments.
And of course, the very small, the smaller kind of enterprise, is kind of a different,increasingly a different business in and of itself.
Let's keep looking at your crystal ball.
What are hyperscalers going to do next?
(31:09):
Yeah, mean, they, you know, I would say the one thing to keep in mind is that the CapExinvestments, the amount of investments they're making in data center and server
infrastructure has not slowed down.
Google yesterday, reported a, confirmed a increase in their projection by $10 billion fortheir 2025 CapEx.
(31:34):
So it went to 75 to 85.
I'm sure some of you have already seen that.
uh So the hygroscaler is going to continue expanding.
ah going to, I think, behave as they always have, which is from the data center operatingperspective, a little fickle, a little kind of back and forth, a little...
(31:56):
They're stressed, as many of us, uh in terms of trying to build for the future and tryingto project.
And they kind of oscillate between going their own...
uh
and then working with their partners.
uh Sometimes they will kind of go through kind of phases where they emphasize one strategyover the other.
(32:21):
But it's an imperfect science.
They're going to miss.
They're going to have difficulties.
Engagement with partners, I think, will always be a part of the strategy.
The way these partnerships kind of move forward or the service that's provided them froman infrastructure perspective can change and is changing.
ah But I, you know, that would be my feedback to operating executives is in trying tounderstand what the hyperscalers are doing is that know that they have difficulty with
(32:48):
this and will this is not easy.
Know that there are, you know, they are, you could argue a bit understaffed when it comesto hyperscale procurement.
The teams that people are working with are not big.
ah So, you know, understand that they have a need for knowledge.
They're going to continue to need to collaborate and partner.
(33:09):
uh share information.
And so I do predict a ongoing collaborative relationship, even if it can be a littlebumpy, maybe even a little weird at times or, you know, frustrating and stressful.
But I think kind of if you take a step back and look at the big picture, uh it is stillgoing to and I think need to be uh collaborative and a relationship between partners.
(33:35):
uh
rather than a vendor selling something to a buyer or an end user.
Where else could hyperscale demand come from?
I still think we're in the phase of we're in this long kind of maybe it's going to be atwo decade build out around the world.
(33:56):
mean, a lot of hyperscale demand is coming from emerging markets in the world.
Our practice is a bit slanted to three major regions, North America, Western Europe, andAsia Pacific, not including China.
And we spend time.
in a lot of the second tier markets in those regions, especially Europe and Asia.
(34:20):
I can tell you that they are just getting started with some of this stuff.
the amount of hyperscale demand coming out of those regions and, you know, if you look atSouth America, Middle East, Africa, we don't spend as much time.
They're just getting started as well.
So I think the global, you know, the global.
kind of build out of cloud is still very much in its first phase and in the kind of secondand third tier.
(34:45):
ah And that may not be an opportunity for everybody, granted, ah but that is a significantdriver of growth ah in the future.
And of course, AI in the sense that the tools and services that are driving cloudconsumption, and ah I think sometimes this gets overlooked, that AI is fueling kind of
(35:07):
quote, quote, tradition or public cloud infrastructure growth in that people are runningcompute and storage on the cloud, but increasingly buying AI or AI type services that are
built and delivered through the cloud as a delivery vehicle.
And so that's why we're seeing public cloud growth in this quarter's earnings season isalready showing it.
know, you know, steady if not acceleration and growth, just driven by the uptake in thoseservices.
(35:34):
So Phil, you've seen multiple technology cycles.
What's your prediction for the next major inflection point in digital infrastructure?
I would say my prediction is that it's going to be a very long kind of wave, right?
I think the hyperscale build out is still going on and it's, like I said, it's been goingon for a decade, decade and a half.
(36:00):
AI, think, is going to take that to the next level.
So we'll probably think of it as a second wave, but really it is what it's just addingfuel to the fire.
I think it's an accelerant of what's going on now.
ah Again, strictly from an infrastructure perspective.
ah And so, yeah, I think it's going to start in a few years and we're talking, we could betalking about another decade or two of significant infrastructure growth ah and data
(36:26):
centers and digital infrastructure, internet infrastructure, whatever term you want tothrow on it, becoming just one of the just kind of major industries.
or that we think about, right?
Healthcare, ah financial services, something that's just kind of core to everybody's ahlife.
(36:52):
And if you just kind of look, I think I'll leave you with this.
If you kind of look at what's happening in the US in the Midwest, right?
Where a lot of data centers are taking over kind of decommissioned or abandoned.
ah
facilities from the manufacturing era, right?
(37:15):
It's kind of a great snapshot and a micro-osm of what's happening.
Traditional industries going by the, you know, retiring or moving to other parts of theworld, being replaced by digital infrastructure, data centers that, you know, that house
the servers that produce the digital content and applications workload that we all use andlive with every day that you access through your phone.
(37:37):
um
I've always thought of that, you know, and the other one would be there in Canada and theUS uh newspaper printing presses.
Now our data centers, right, been converted.
You know, we don't really consume newspapers anymore.
ah But where do we consume our news on websites?
Where are those websites on your phone?
(37:58):
Yes.
But where are they accessed through servers that sit in data centers?
And so you kind of see that old world transition to the new.
So if you think of things in those terms,
And these are inexorable, know, these things are in motion and there's nothing that'sgoing to stop it ah short of the world vaporizing.
So, ah but if you think of it in those terms, ah you can see how these waves or you cankind of, I think, come to see that these waves are significant.
(38:29):
They are, you know, going to happen over a long period of time.
I don't think it's hyperbole when people say this.
We had the Industrial Revolution in the late 1800s.
I don't necessarily think it's hyperbole to say that that's what we're undergoing, whichis a complete revolution.
It's just the digital infrastructure revolution.
(38:52):
It's a transition from economies based on industry, on commodities, on physical things todigital ah and internet-based.
um And if you believe that and if you see that that's
And if you believe that's going to continue, it becomes much more reasonable to think thatwhat's happening around data centers and cloud and servers, compute infrastructure is on a
(39:16):
similar path.
And then think about, yeah, it's only been happening for 20 odd years.
Imagine what, you know, if it's not going to slow down, what could happen in the next 20,25 years?
So that's why I'm optimistic and why I'm still working in this space.
I need employment.
Yes, speaking of that, my last question is if there's young people out there that arelooking at what am I going to do with my life, right?
(39:46):
The unemployment is so high for college graduates right now, or there's even people intheir midlife career cycle that are thinking, what am I going to do?
Is my job going to be replaced by AI?
What should I do next?
What should I be training for?
What are the opportunities in the space, Phil?
(40:08):
Do you see a few that you can talk about for people?
Oh, absolutely.
mean, I speak with operating companies all the time and a challenge they have ah isfinding people with, you know, industry or, you know, subject matter, knowledge,
experience, expertise.
ah It is not a common skill set to say, you know, to say, Hey, I worked in cloud or datacenter or hyperscale, or I understand, you know, how this industry works or I've done it
(40:33):
before.
That's not something easily found ah at the executive level, at the mid management level.
uh
You know, so I think that there are, there's, you know, room for people who understand thespace, who can analyze it, who can think about it critically and creatively to solve some
(40:54):
of these problems.
So we were ultimately, infrastructure is about solving problems.
So the, whether you're coming out of school, whether you're mid-career, if you're good atsolving problems or enjoy solving problems, looking at things from an analytical lens and
perhaps thinking a bit out of the box.
There's tons of opportunity and upside in a space that's built on being able to solveproblems that are not kind of easily visible.
(41:18):
So that's what I would say is those kinds of minds.
You can learn the space pretty quickly, right?
It's, you know, the building blocks and the ABCs of the space wouldn't be intimidated orscared if you're a young person listening coming out of school.
uh No, explore it, take a look at it, and you can apply skills that you have.
(41:38):
built throughout your life, your education, your career to this point, all can translate.
ah And like I said, if you're struggling, like you mentioned, Kerry, you're always lookingfor openings, shortages, of, where can I, my expertise or be, where is there scarcity?
(42:00):
I think you can find, I think this space there is definitely a level of scarcity thatpeople may be able to exploit further.
for their own gain.
You are right about that.
How can we reach structured research?
Your website.
Yep, uh sorry, we have two, so we have a website and an event site for our summit.
(42:25):
But if you go to our website, structureresearch.io, so just as it sounds,structureresearch.io, ah and there's a link to the event site there, but the event site
URL, sorry that it's long, is infrastructuresummit.io.
We like to use the .io, so yeah, that's structureresearch.io and infrastructuresummit.
(42:47):
dot i-o.
No fancy spellings, we're short, just the words.
And I think it's important that we all attend these events and these conferences becausethings are changing so rapidly.
So I will for sure see you there in October.
And Phil, thank you so much for coming on.
This has been educational.
I just really love learning from you.
(43:11):
uh And I will learn more in October.
I can't wait.
Take care.
You too.
so that was awesome awesome awesome stop
Thanks for joining me on Let's Get Digital.
I'm Carrie Charles, your host, and today I have with me Phil Shee.
He is the founder and managing director of Structure Research.
Phil, thanks for coming on the show.
Thanks for having me.
So we met at ITW, right?
Yup, that wasn't that long ago.
(00:22):
Yeah, it sure wasn't.
Boy, time flies.
uh I know you have an interesting journey, and I'd love to hear a little bit about thefounding story of structure research, how you got to where you are today.
Yeah, mean, the whole story, it's just so much kind of just kind of live in life and thenjust kind of making decisions based on, you know, opportunities and doors that open ah and
(00:48):
kind of getting to the point in life where you thought, hey, maybe there's a better way todo this uh and kind of saying, hey, I got to take the leap.
If I'm ever going to do it, I got to do it now.
It was kind of like uh that kind of thinking, you know, and like you, I think many yearsago, you just said, hey,
What would it be like to not have a boss and to drive a lot of the decisions that Iincreasingly am not agreeing with in the firm I'm at?
(01:15):
I don't think that's an unusual, you know, that's not a unique story.
But honestly, that's what drove it.
It's like that question is like, hey, what if I did, you know, have the control and theability to make decisions and drive this in the way I want, but that's going to take a
leap as in, you know, saying goodbye to a regular paycheck.
predictability and the comfort and trying something completely different, which, you know,I think a lot of entrepreneurs or people who've gone on their own did not necessarily have
(01:42):
that in the plans, have that in store.
Again, it's about opportunity and, know, the field I was in, or the field that we are in,this industry, you know, was healthy then.
I saw, I think I saw the upside as many of us did.
And I thought, well, if you put those things together, wow, maybe this could work.
And yeah, I'm glad, I'm glad I made the leap.
was the
(02:03):
you know, something that I didn't expect, but I'm glad it's working out and uh I'mgrateful for the opportunity that the industry has provided.
How many years ago was that?
Yeah, 2011.
So late in 2011, I decided to make the jump and I've been at it ever since.
Okay, well you know what you're doing, so let's get into it.
(02:25):
Well first, tell us more about structure research.
Yeah, in a nutshell, for those of you who aren't familiar, we describe ourselves and thinkof ourselves as a boutique research firm, which is another way of saying we're smaller,
we're slightly understaffed, but we are specialists.
And that's kind of part and parcel of our value proposition.
(02:46):
We have seven analysts that spend all our time in this sector.
When I say this sector, we used to call it internet infrastructure.
But data centers, managed cloud, managed hosting, what used to be called managed hosting.
But the infrastructure services ecosystem, what fits now under the kind of broader rubricof digital infrastructure, we're kind of more focused on the compute-oriented
(03:12):
infrastructure side, rather than say wireless and connectivity.
We're very much, you know, I started my career basically, know, following servercompanies, server hosting companies.
And that's morphed over the years into managed services.
data centers and now hyper scale in AI.
So yeah, the boutique, that's who we are.
We spend all our time publishing industry focused research.
(03:33):
Think of us as living in the kind of middle ground between a Wall Street analyst or astreet analyst that's focused on earnings, the stock, the value of that stock and
recommending it to investor clients.
And then on the other kind of the other side of the spectrum, kind of more technical ortechnology oriented analysts that are focused on
(03:53):
the kind of details and the nuts and bolts of the technology pieces, we're very muchfocused on the industry, kind of macro trends, the landscape, the big picture, helping,
ultimately trying to provide information perspective and insights to help executives makedecisions.
And that's kind of, in a nutshell, who we are.
(04:16):
So you do compete in a space with massive analyst firms.
And I guess I would look at it a little bit like David and Goliath, but what makes yourboutique approach different and also advantageous for your clients?
Yeah, mean, it's again, the specialization kind of creates a certain, you know, certainamount of efficiency for the executive decision maker that subscribes to the service.
(04:42):
You know, they want information that's tailored and curated to their world, their focusand what they care about.
That's our pitch when we meet people.
It's like, hey, if you're spending a lot of time in this sector, if you care about what'sgoing on in this sector, if you have a stake, whether financial or kind of strategic,
ah you know, we could be the right service for you.
(05:03):
We don't kind of try to cover too much ground.
And that's how, you know, honestly, we too much ground in the sense that we don't cover alot of other sectors.
You know, we're not even covering broader digital infrastructure like towers, wires, Imentioned, wireless and connectivity is kind of slightly kind of referral to what we do.
And that's how we exist.
You know, you asked the question, how do we exist against some of the larger firms?
(05:26):
Well, we've selected a sector that we think is valuable.
that is frankly a bit undercover, don't let that get out, and say, hey, we're all aboutthis.
And if this is what you care about, we uh could be of service.
We're not for everybody.
ah And we're pretty comfortable with that.
And again, that's our pitch and that's how we survive.
The bigger firms tend to offer more and continue to offer more.
(05:50):
They're obviously have hundreds, thousands of people on their staff and they can cover alot of ground.
But we give you something very focused.
So we dive into, uh
we're able to provide a lot of depth and we give you a senior analyst perspective.
Everybody in our team, ah whether you're a junior analyst, but you quickly within our kindof environment ramp up pretty quickly because you're just doing this.
(06:16):
You're not focused on anything else.
Okay, that makes sense, and you've been doing it a long time.
Yeah, since 2011.
And kind of just gradually built up the team over the years.
yeah, that's the great world of remote working.
We were a remote team before COVID, the pandemic.
And it's working.
That's great.
(06:37):
That's great.
So you have an event.
I guess it's been going for six years now, right?
It's infrastructure.
um Infra is, uh I guess, small letters, then there's, uh you know, structure is all caps,but it's an infrastructure summit.
It's in Las Vegas, October 15th and 16th.
I know I'm gonna be there.
(06:58):
Can you tell us a little bit about the event?
What makes it different than all the other industry events as well?
So Infrastructure was founded, we had our first event in 2018.
It ah is the sixth edition this year, there were a couple years where we didn't have itdue to that pandemic.
ah And it started uh as and still is meant to be an executive summit ah that is unique inthe sense in that it's very uncommercialized.
(07:27):
So the content is obviously the agenda is backed, supported and led by an analyst teamthat is independent.
that is not tied to, you know, kind of any vendor or platform or investment vehicle.
uh Rather, we bring those parts of the ecosystem to our summit so that people can kind ofnetwork, you know, partner, engage in business development, and of course, discuss kind of
(07:50):
the present and future of the sector.
so research, you know, I believe, and I think our audience values is a really greatbackdrop and context for that discussion.
And so that's what we decided to do in 2018.
We slowly built up the event and you know our mantra for the summit is you know qualityover quantity.
We're not trying to build an expo.
(08:11):
We actually don't really have an exhibit hall.
We make this very clear to our sponsors that want to come.
That if you want to come it's a different type of engagement.
There's obviously room for business development but the way you go about it, the way youengage is a bit different because it's a summit because we're in the participation is in
the hundreds rather than thousands.
(08:31):
Yeah.
No, give me a little bit of a peek into the agenda, maybe some topics that you're going tocover and people who are going to attend maybe.
Yeah, so it's executive heavy.
So VPC level, a lot of owners, lot of entrepreneurs who run their business, a lot of CEOs.
(08:54):
And that's kind of the primary demographic.
uh But we surround them with uh two other major uh groups of people, which are investorsand financiers, or people on the finance side and real estate side of the game, and then
suppliers, people who sell products, services, or resources.
to the operating service providers so that they can build and deliver their services.
(09:17):
Real estate can also fall in that category.
Increasingly, we have energy and then of course software and hardware and constructionservices, commissioning services, various types of vendors that support this ecosystem.
ah So that's the audience and then the agenda uh tries to cater to the questions of theday um that everybody again that has a stake and interest in the sector
(09:43):
um you know, want answers to or want to hear discussion about or hear from, you know,leading minds and, you know, thought leaders in the space.
So, you know, you asked about the agenda this year.
ah You know, we try to find a balance.
So there's going to be some sessions, uh &A, you know, on the strategic landscape, uh bigpicture questions around, you know, the impact of AI, uh the impact of the recent
(10:11):
hyperscale pullbacks.
how that impacts the demand profile of the sector.
We have various topics on energy this year, how are energy constraints uh influencing orconstraining the ability of uh operating companies to continue to build in the face of
significant steam of demand.
And of course, there's always those who are a little skeptical and it's a great balance.
(10:37):
oh This year, we hope to have a lot of discussion.
m
between those who are a bit bearish and those who are a bit bullish.
That seems to come up every kind of cycle or phase in the sector where you see someskeptics rise up, know, and, you know, not without, you know, of legitimate concerns about
where things are going.
(10:58):
But that fuels a great debate and discussion.
And, you know, we love the diversity of opinions and takes.
So your 2025 theme is concurrence and I wanted you to talk more about what that means andwhy is it the right framing for today's infrastructure challenges.
(11:20):
Yeah, every year, so every year we try to pick one word as the theme, right?
So we've had, except for the first year, uh but we've had decentralization to kind ofcapture where the sector was moving a bit to the edge, uh acceleration coming out of the
pandemic in terms of kind of where the growth was.
(11:41):
And this year we picked concurrence, which is meant to capture kind of just the manymoving parts in this space.
Right.
So I did have to look up the dictionary definition of concurrence and it is on thewebsite.
kind of cut that out and put it in there.
Thanks, Merriam Webster.
But, you know, just try to capture kind of where the sector is in terms of how it'spreparing and engaging in building for this, you know, next wave that's coming.
(12:10):
And so a lot of moving parts, a lot of different things happening at the same time.
That's kind of what we're trying to capture, you know.
We're built, hyperscale is still, hyperscale cloud is still driving things.
AI is on the horizon.
Power constraints are developing.
How, you know, uh that opens up kind of a different regulatory picture.
uh Capital is flowing into the sector.
(12:31):
How do all these things moving kind of at the same time and kind of changing the waythings are done?
You know, how does that, you how do we make sense of that?
How does the sector organize around that to push forward?
So.
Yeah, um hopefully that makes sense.
We do try to be little edgy and different, but at the same time, capture what is the mostpressing dynamics in this space.
(13:00):
And that's what the concurrence is about.
So with over 20 years covering global infrastructure services, what have been the mostimpactful market shifts that you've observed?
Yeah, I mean, I think quite clearly the biggest shift was, you know, probably the, let'sjust agree the sector was born in the late 90s, early 2000s.
(13:24):
ah So the big shifts, I think, have been pretty clear that, you know, for those of you whoare a little older and remember the dot-com bubble, ah seems like, you know, it's almost
25 years ago now, quarter century.
ah That would be kind of the first thing, you know, one of the
(13:44):
the first big seismic events that kind of gave birth in many ways to the sector as a lotof operating companies kind of built out of the dregs of what happened.
And the sector was so young, a lot of assets were sold pennies on the dollar and the valueproposition of the space kind of took advantage of that situation to kind of build.
(14:05):
And then in 2000, the next kind of big development, 2006, AWS, Amazon decides to get intothe cloud business back then.
strictly known as an online book retailer, really not even, what did you buy from Amazonin 2006 besides books?
Nothing.
They got into the cloud business, which was a curiosity at the time, but set off a decadeplus move from kind of traditional infrastructure, which was still young and still
(14:33):
developing.
uh We didn't call it cloud then, ah but was still developing.
then at the same time, this
kind of hyperscale, what we now call hyperscale cloud, public cloud, was developing in thebackground, started to increasingly kind of become more prominent and now arguably, uh you
(14:55):
know, is probably the dominant kind of computing platform of choice for, or not to sayit's the only one, but it's definitely, you know, front and center and controls a
significant part of this ecosystem.
And so our sector has transitioned and built around that.
data centers, managed services, MSPs are all part of that.
And of course, there's still a lot of diversity in terms of other cloud platforms, butthat's the big one.
(15:20):
That's been a, depending on what part of the world you're sitting in, a decade, decade anda half.
It always amazes me how young some parts of the world, they're hyperscale.
The markets in those countries are so young, maybe just a few years with hyperscale cloud.
So that wave's still ongoing.
And I think the next big one most of us can agree is coming.
(15:42):
ah How it's going to play out, we don't all agree on, but AI and how that is going toaccelerate things is the next big inflection point wave that we'll pick up.
But interestingly, think overlap a bit with hyperscale cloud.
So you've led consulting mandates totaling over three billion in deals.
(16:04):
Is that, am I right?
My reading?
My research?
I think so.
Yeah, if we put it out there, we did a bit of quick math.
eh
Okay, good.
Can you share a standout transaction that really shaped Structure Research's strategicapproach?
Sure, ah so we're not a consulting firm, we're research firm uh at heart.
(16:26):
That's the of the majority of where our business model is built on that.
We do consulting on a strategic basis, on a selective basis, mostly around, as youmentioned, &A transactions.
We don't necessarily lead kind of that process.
We let the consultants do what they're good at.
We come in as the market expert, the subject matter expert, the market data source.
(16:49):
and support transaction activity in that way.
So, you know, we've done small transactions, large-scale ones.
We've done the first one I did under the banner of structural research was actually inEurope.
ah But we've done work in North America and we have quite a specialized practice in AsiaPacific.
ah Our head of research and myself, Hale, from the region still have strong ties there,spend a lot of time there.
(17:13):
And that's actually where the largest transaction we've done, which was the
kind of previous, not the recent one, but the previous investment in Australia's AirTrunk, which is, know, hyperscale data center platform.
uh And, you know, that value is a good chunk of the 3 billion you referenced.
uh But yeah, that's, that's, I mean, that's just one example of things we've done in thepast to support &A in the space.
(17:42):
Do you have proprietary em data sets that differentiate your insights?
Absolutely.
All our market share data, everything we put out there is built from the ground upinternally and organically.
It's a pretty painful, painstaking process.
It's a grind, we describe it, but we feel it pays off and it ensures the integrity of thedata that we put out there.
(18:10):
So yeah, every data set is based on what we feel is...
hard work, getting boots, we like to call it getting boots on the ground, talking tooperators, visiting sites, making sure we're double checking those facts with other
sources.
A lot of what we call triangulation, throwing out the high, throwing out the low, takingdata points, try to compare them and you tend to get closer to the truth if you undertake
(18:36):
that process.
yeah, absolutely everything we put out there, we've built, we own, we stand by it and tryto refine.
I think one of the other things that we, I think we do really well is that we refine overtime.
So we're always kind of, you know, uh we don't think we have the answer, the answers andat any given time.
(18:57):
And, you know, if we see other kind of facts or feedback or insights that tell us thatwe're wrong, we're going to adjust.
And so that, you know, over time we get closer to the truth and that's the approach we'vetaken.
So data center M &A hit record levels in 2024.
I was reading that also.
And what do you think is driving this?
(19:19):
mean, I probably we probably all know that.
But what should infrastructure leaders know about navigating this current M &Aenvironment?
Yeah, I mean, I think what's interesting, I mean, I would say not that we have, I have ahard data point, but my sense definitely is that the number of transactions is lower, but
the value is higher, right?
(19:40):
Is that, that's kind of, if you were measuring like, Hey, there was more value, you know,more transactions it's because the numbers are just so big now.
Um, you know, what would I say in terms of, you know, I
of kind of where we are and where we're going and how would operating executives kind oflook at that, I would say.
I mean, first of all, I think that's one thing that hasn't changed is that the sector hasalways seen consistent and steady strategic activity across the board from kind of early
(20:12):
gen cloud, smaller service providers to the largest operating platforms in the world.
ah I think it's a product of the fact that the sector has always been
you know, very kind of fragmented in its own way across different segments and now isincreasingly global, which has actually caused a cause, you know, led to kind of, you
could argue more fragmentation with regional markets, you know, and global operatingcompanies operating in different regions and countries around the world.
(20:41):
I think, I think that's just kind of part and parcel of how this stuff is built is thatbecause infrastructure is
Not like an iPhone or something.
It's not like a closed walled garden.
It's really aggregation and piecing together of different resources.
And they're done in different ways in different markets in different countries and regionsaround the world.
(21:02):
And so that creates a very heterogeneous environment that's conducive to M &A.
I don't think it's going away.
It goes through waves where there's more activity than not.
But in my experience, it's been steady just for so long.
And I know a lot of my &A and banking contacts are very happy about that, of course.
(21:25):
um But yeah, I would say to operating executives, that it's always something there tosupport strategic activity can always support growth and expansion.
There are great options out there.
um
the value of, you know, the valuation of assets in the space has always been high.
(21:48):
It continues to, you know, to climb depending on what part of the market you're in.
And that would be, I guess, the last kind of, I guess, if I could add a tidbit is tounderstand which segment you're in in the market and to kind of build, evolve and
transition your business accordingly.
Right.
So there are always going to be certain sectors that are, that have, that are healthier,certain sort, certain segments within the overall kind of data center infrastructures.
(22:14):
cloud infrastructure space that are trending in a better direction that have healthierupside.
There are some services now that we call legacy fairly because they are at a differentstage in their evolution.
And so when you're looking at operating assets with your, or an operating side and lookingto exit, or you're on the buying side and looking to build, expand, diversify, augment, uh
(22:38):
those are the, I think that is the most important thing to think about.
And hopefully a research firm can help provide some context to that.
So we can't have a conversation without talking about AI.
So here we go.
And everyone's talking about AI driving massive data center demand.
But you track the actual numbers.
So behind the hype, if you will.
(23:00):
So what does that real data tell us about AI's impact on the sector?
Yeah, it's a crazy time to be an analyst because the short answer to that question is, youknow, materialization of that demand is still very much at a very early stage of
development, right?
(23:21):
So that's not to say that things haven't landed and are running and impacting growth.
It certainly has.
But the the kind of surge or the wave that we expect if this technology continues to movein the direction we see it moving.
We're still probably a few years out, right?
To see that kind of really spike uh or shift in that trajectory on your PowerPoint chartsand everything.
(23:49):
Maybe it will look like a hockey stick, but not right now.
Right now, things are still being driven by what they've been driven by, which is publiccloud, hyperscale cloud.
But you're starting to see these models start to be built.
And then when that happens, there's kind of a spin-off effect across infrastructure.
(24:10):
So the models will be trained and they'll have to be used by the end user to deliver them.
They'll have to build inferencing infrastructure.
And then to maintain fine-tune and perfect those models, you're going to needinfrastructure that sits almost in between where the models are built, the large-scale
models, and then of course the inferencing.
And so that's more infrastructure.
(24:30):
And then you need connectivity to move the data, to access the data, which means moreinterconnection uptake, more network services.
And that spins off into service providers, et cetera, et cetera.
yeah, the way we're modeling it is we're trying, it's not an exact science as you canimagine, but we're trying to kind of predict when in the next few years some of these
(24:55):
things start to come online and that spin-off effect takes shape.
ah and then impacts kind of the entire ecosystem.
So our latest update, our latest numbers, you'll see a meaningful increase in what we'reprojecting for growth, but with a spike not quite coming just yet.
ah And it's very hard to predict and understand what that spike will look like.
(25:17):
So historical precedent is useful, but not always a perfect kind of comp comparison.
Can you just give me a little peek into what art, you're predicting?
Yeah, no, I mean, I think we had the data center co-location space last year growing atabout 13 and a half percent globally.
(25:39):
We now have it closer to 20 % and believe it'll be growing over 20 % in the next few yearsbased on kind of everything we're seeing.
So are we building fast enough and what are the bottlenecks that most people are just notseeing?
(26:00):
Yeah, it's difficult to say if we're or not building fast enough because the way thisdemand will materialize is hard to predict because it's so new.
Hyperscale cloud, if you're building availability zones for traditional cloudinfrastructure, we have enough of a sample size to be able to predict within reason what
(26:25):
it might look like.
are going to look very different.
So yeah, it's hard.
I say it's hard to say, and I'm kind of just, I think I kind of just missed the other partof your question as I was saying that, but right, right, right.
The bottlenecks.
(26:46):
So yeah, so it's difficult to say whether we're building too fast just because we don'tknow what's, we don't have a clear picture yet of what's coming.
But in terms of bottlenecks, absolutely.
I think you can see.
that the most clear characteristic of this next wave is the amount of energy that needs tobe consumed.
(27:08):
And there are increasingly signs that in the most mature markets, ah especially, ofcourse, in the United States, that uh if any of these predictions or projections are even
close, there's going to be significant uh shortage of energy.
And so that's kind of one of the biggest, if not the biggest constraint I would throw inthere as well, uh which related to energy is of course, land in it, real estate.
(27:37):
But there's also a shortage of expertise, right?
you need, uh hyperscale platforms have come to build on their own.
They face their self-build internally and not use data center co-location operators.
They've come across their own constraints and there are only so many
uh credible, uh reputable, established kind of data center operating platforms that canbuild at this level of this scale at the quality that's going to be required.
(28:07):
yeah, land, energy, expertise, and then the regulatory environment as well could be aconstraint given the constraints on energy are going to cause a shift to alternative
sources of energy, uh nuclear being that one that's been discussed quite a bit.
but also comes with a lot of regulatory implications.
(28:28):
yeah, it's a concurrence.
A lot of things happening at the same time that need to be resolved and solved before thesector can move forward.
So I'm a data center, let's say I'm a data center owner operator, what are some key uhdata points that I need to know that will help me navigate all of this, the concurrence?
(28:53):
Yeah, I would say come to the show and hear the discussion.
That's one.
Las Vegas in October is not a bad time of year.
um you know, it's a good question.
There's so many different kind of data points and factoids.
would say if you're an operating executive is to kind of look for honestly, I mean, themarket is healthy, but at the same time, there's a lot of competition, right?
(29:20):
There's a lot of capital flowing into space.
Everybody wants to get involved.
Um, you know, finding a market inefficiency is, you know, if, you're a new operating, youknow, whether you're a new operating platform or an existing one that's looking for other
opportunities, you know, finding where market inefficiencies are, uh, I think is going tobe increasingly important.
Uh, sometimes the industry kind of, I don't want to say, you know, over indexed, maybeover indexes a bit, or there's a lot of opportunity.
(29:47):
And so everyone flocks in one direction and then certain parts get neglected.
Right.
uh Obviously people push in a certain direction because for good reason, uh but in thatprocess, certain things come up, certain areas get slightly neglected.
so I think you're going to start to see, uh think investors, operating companies shouldexplore kind of what we've described as the edge hyperscale market, which is kind of
(30:17):
something that exists between enterprise and
and kind of large scale, hyper scale that everybody has been chasing, availability zones,regions that is a bit underserved, ah but there's no shortage of demand that's starting to
build towards that.
So edge hyper scale, pretty broad term meant to capture kind of secondary, second tiermarkets or even within major markets, but certain type of requirement that's broadly
(30:48):
speaking.
single digit megawatt to 10 or 20 versus the very large deployments.
And of course, the very small, the smaller kind of enterprise, is kind of a different,increasingly a different business in and of itself.
Let's keep looking at your crystal ball.
What are hyperscalers going to do next?
(31:09):
Yeah, mean, they, you know, I would say the one thing to keep in mind is that the CapExinvestments, the amount of investments they're making in data center and server
infrastructure has not slowed down.
Google yesterday, reported a, confirmed a increase in their projection by $10 billion fortheir 2025 CapEx.
(31:34):
So it went to 75 to 85.
I'm sure some of you have already seen that.
uh So the hygroscaler is going to continue expanding.
ah going to, I think, behave as they always have, which is from the data center operatingperspective, a little fickle, a little kind of back and forth, a little...
(31:56):
They're stressed, as many of us, uh in terms of trying to build for the future and tryingto project.
And they kind of oscillate between going their own...
uh
and then working with their partners.
uh Sometimes they will kind of go through kind of phases where they emphasize one strategyover the other.
(32:21):
But it's an imperfect science.
They're going to miss.
They're going to have difficulties.
Engagement with partners, I think, will always be a part of the strategy.
The way these partnerships kind of move forward or the service that's provided them froman infrastructure perspective can change and is changing.
ah But I, you know, that would be my feedback to operating executives is in trying tounderstand what the hyperscalers are doing is that know that they have difficulty with
(32:48):
this and will this is not easy.
Know that there are, you know, they are, you could argue a bit understaffed when it comesto hyperscale procurement.
The teams that people are working with are not big.
ah So, you know, understand that they have a need for knowledge.
They're going to continue to need to collaborate and partner.
(33:09):
uh share information.
And so I do predict a ongoing collaborative relationship, even if it can be a littlebumpy, maybe even a little weird at times or, you know, frustrating and stressful.
But I think kind of if you take a step back and look at the big picture, uh it is stillgoing to and I think need to be uh collaborative and a relationship between partners.
(33:35):
uh
rather than a vendor selling something to a buyer or an end user.
Where else could hyperscale demand come from?
I still think we're in the phase of we're in this long kind of maybe it's going to be atwo decade build out around the world.
(33:56):
mean, a lot of hyperscale demand is coming from emerging markets in the world.
Our practice is a bit slanted to three major regions, North America, Western Europe, andAsia Pacific, not including China.
And we spend time.
in a lot of the second tier markets in those regions, especially Europe and Asia.
(34:20):
I can tell you that they are just getting started with some of this stuff.
the amount of hyperscale demand coming out of those regions and, you know, if you look atSouth America, Middle East, Africa, we don't spend as much time.
They're just getting started as well.
So I think the global, you know, the global.
kind of build out of cloud is still very much in its first phase and in the kind of secondand third tier.
(34:45):
ah And that may not be an opportunity for everybody, granted, ah but that is a significantdriver of growth ah in the future.
And of course, AI in the sense that the tools and services that are driving cloudconsumption, and ah I think sometimes this gets overlooked, that AI is fueling kind of
(35:07):
quote, quote, tradition or public cloud infrastructure growth in that people are runningcompute and storage on the cloud, but increasingly buying AI or AI type services that are
built and delivered through the cloud as a delivery vehicle.
And so that's why we're seeing public cloud growth in this quarter's earnings season isalready showing it.
know, you know, steady if not acceleration and growth, just driven by the uptake in thoseservices.
(35:34):
So Phil, you've seen multiple technology cycles.
What's your prediction for the next major inflection point in digital infrastructure?
I would say my prediction is that it's going to be a very long kind of wave, right?
I think the hyperscale build out is still going on and it's, like I said, it's been goingon for a decade, decade and a half.
(36:00):
AI, think, is going to take that to the next level.
So we'll probably think of it as a second wave, but really it is what it's just addingfuel to the fire.
I think it's an accelerant of what's going on now.
ah Again, strictly from an infrastructure perspective.
ah And so, yeah, I think it's going to start in a few years and we're talking, we could betalking about another decade or two of significant infrastructure growth ah and data
(36:26):
centers and digital infrastructure, internet infrastructure, whatever term you want tothrow on it, becoming just one of the just kind of major industries.
or that we think about, right?
Healthcare, ah financial services, something that's just kind of core to everybody's ahlife.
(36:52):
And if you just kind of look, I think I'll leave you with this.
If you kind of look at what's happening in the US in the Midwest, right?
Where a lot of data centers are taking over kind of decommissioned or abandoned.
ah
facilities from the manufacturing era, right?
(37:15):
It's kind of a great snapshot and a micro-osm of what's happening.
Traditional industries going by the, you know, retiring or moving to other parts of theworld, being replaced by digital infrastructure, data centers that, you know, that house
the servers that produce the digital content and applications workload that we all use andlive with every day that you access through your phone.
(37:37):
um
I've always thought of that, you know, and the other one would be there in Canada and theUS uh newspaper printing presses.
Now our data centers, right, been converted.
You know, we don't really consume newspapers anymore.
ah But where do we consume our news on websites?
Where are those websites on your phone?
(37:58):
Yes.
But where are they accessed through servers that sit in data centers?
And so you kind of see that old world transition to the new.
So if you think of things in those terms,
And these are inexorable, know, these things are in motion and there's nothing that'sgoing to stop it ah short of the world vaporizing.
So, ah but if you think of it in those terms, ah you can see how these waves or you cankind of, I think, come to see that these waves are significant.
(38:29):
They are, you know, going to happen over a long period of time.
I don't think it's hyperbole when people say this.
We had the Industrial Revolution in the late 1800s.
I don't necessarily think it's hyperbole to say that that's what we're undergoing, whichis a complete revolution.
It's just the digital infrastructure revolution.
(38:52):
It's a transition from economies based on industry, on commodities, on physical things todigital ah and internet-based.
um And if you believe that and if you see that that's
And if you believe that's going to continue, it becomes much more reasonable to think thatwhat's happening around data centers and cloud and servers, compute infrastructure is on a
(39:16):
similar path.
And then think about, yeah, it's only been happening for 20 odd years.
Imagine what, you know, if it's not going to slow down, what could happen in the next 20,25 years?
So that's why I'm optimistic and why I'm still working in this space.
I need employment.
Yes, speaking of that, my last question is if there's young people out there that arelooking at what am I going to do with my life, right?
(39:46):
The unemployment is so high for college graduates right now, or there's even people intheir midlife career cycle that are thinking, what am I going to do?
Is my job going to be replaced by AI?
What should I do next?
What should I be training for?
What are the opportunities in the space, Phil?
(40:08):
Do you see a few that you can talk about for people?
Oh, absolutely.
mean, I speak with operating companies all the time and a challenge they have ah isfinding people with, you know, industry or, you know, subject matter, knowledge,
experience, expertise.
ah It is not a common skill set to say, you know, to say, Hey, I worked in cloud or datacenter or hyperscale, or I understand, you know, how this industry works or I've done it
(40:33):
before.
That's not something easily found ah at the executive level, at the mid management level.
uh
You know, so I think that there are, there's, you know, room for people who understand thespace, who can analyze it, who can think about it critically and creatively to solve some
(40:54):
of these problems.
So we were ultimately, infrastructure is about solving problems.
So the, whether you're coming out of school, whether you're mid-career, if you're good atsolving problems or enjoy solving problems, looking at things from an analytical lens and
perhaps thinking a bit out of the box.
There's tons of opportunity and upside in a space that's built on being able to solveproblems that are not kind of easily visible.
(41:18):
So that's what I would say is those kinds of minds.
You can learn the space pretty quickly, right?
It's, you know, the building blocks and the ABCs of the space wouldn't be intimidated orscared if you're a young person listening coming out of school.
uh No, explore it, take a look at it, and you can apply skills that you have.
(41:38):
built throughout your life, your education, your career to this point, all can translate.
ah And like I said, if you're struggling, like you mentioned, Kerry, you're always lookingfor openings, shortages, of, where can I, my expertise or be, where is there scarcity?
(42:00):
I think you can find, I think this space there is definitely a level of scarcity thatpeople may be able to exploit further.
for their own gain.
You are right about that.
How can we reach structured research?
Your website.
Yep, uh sorry, we have two, so we have a website and an event site for our summit.
(42:25):
But if you go to our website, structureresearch.io, so just as it sounds,structureresearch.io, ah and there's a link to the event site there, but the event site
URL, sorry that it's long, is infrastructuresummit.io.
We like to use the .io, so yeah, that's structureresearch.io and infrastructuresummit.
(42:47):
dot i-o.
No fancy spellings, we're short, just the words.
And I think it's important that we all attend these events and these conferences becausethings are changing so rapidly.
So I will for sure see you there in October.
And Phil, thank you so much for coming on.
This has been educational.
I just really love learning from you.
(43:11):
uh And I will learn more in October.
I can't wait.
Take care.
You too.
so that was awesome awesome awesome stop
Popular Podcasts
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.
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
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!