August 25, 2025 • 38 mins
Brian Comiskey, Senior Director of Innovation and Trends at the Consumer Technology Association (CTA), explores how emerging technologies—including cybersecurity, AI, robotics, and cloud computing—are evolving into digital utilities, with important implications for investors.
----------------------------------------------------------------------------------------------
Subscribe Here to the ROI Podcast & other First Trust Market News
Website: First Trust Portfolios
Connect with us on LinkedIn: First Trust LinkedIn
Follow us on X: First Trust on X
Subscribe to the First Trust YouTube Channel
Subscribe to the ROI Podcast YouTube Channel
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Ryan (00:09):
Hi, welcome to this episode of the First Trust ROI
podcast.
I'm Ryan Isakainen, etfstrategist at First Trust.
Today I'm joined by BrianKomsky, senior Director of
Innovation and Trends at theConsumer Technology Association.
Brian is an expert ontechnology and that's what we're
going to discuss.
We'll talk about cybersecurity,robotics, artificial
intelligence, cloud computingand a variety of other topics.
(00:31):
Thanks for joining us.
So, looking around the world,there's quite a bit of stress,
there's quite a bit of conflict,and one of the things that
strikes me is that the newfrontier of war isn't
necessarily kinetic war, butmore information and cyber war,
a lot of potential cyber threatsfrom some of the players on the
(00:54):
geopolitical stage.
So my question for you, brian,is, as you kind of think about
cybersecurity quite a bit andsome of the related technologies
, how do you expect that willactually impact spending on
cybersecurity from companies,from nations?
Do you think that'll have animpact?
Brian (01:14):
Certainly, and part of the experience that I like to
draw upon and think about is Iused to do some consulting with
the US Coast Guard, and one ofthe areas that was a major focus
of my time there was buildingout new policies around
cybersecurity and cyberoperations, because they
recognized that we were movingfrom three traditional theaters
or kinetic theaters of war, land, sea and air into a fourth
(01:35):
theater that was equallyimportant, which was the
cyberspace, and so you alreadysaw that investment going in
from the US government side, andthat's something that's
definitely replicated throughoutthe world, where you're seeing
budgets increase in the EU,you're seeing those budgets
increase in APAC region, and socertainly governments are
recognizing that they need toplay a role in amping up
(01:55):
cybersecurity from the idea thatthere are nation states that
might attack them in that regard.
But one of my favorite stats isonly 25% of cyber attacks are
state sponsored.
Most of them about 75% arefinancially motivated, which
means that they're notnecessarily targeting a nation
state.
They might be targeting abusiness in this case, and if
(02:17):
it's such a lucrative illiciteconomy that emerges, then
companies need to make sure thatthey're spending in order to
protect from themselves againstit.
So we expect about like $212billion in information security
spending this year alone perGartner.
Ryan (02:31):
So is that ransomware sort of attacks?
What's the financial motivewhen you're talking about
financial motives, is itransomware?
Is it corporate espionage?
What sort of financial motivesare we talking about?
Brian (02:42):
It can be a bit of both Corporate espionage, for sure
but that usually means thatyou've got to pull back a layer
and there might be a nationstate behind that.
But usually it's ransomwarewhich has a lot of payments.
And if you look at the statsgoing in from 2023 to 2024, you
saw a 5x increase in the mediumransomware payout, which, when
we think about what thosepayouts look like, it's not just
(03:03):
oh, here's the money that I'mspending in millions of dollars
to get my data back, but you'vehad an impact to your operations
as a business, which means,uh-oh, I have more costs that
are associated with it.
So when you think about payingout ransomware, it's not
actually just a one-time onlypayment.
It's usually a three-year costwindow that you're paying out,
(03:23):
where you're paying 66% of yourtotal cost from the impact the
first year and then another 22%second year, and then you finish
out with that 11% in your thirdyear.
Ryan (03:31):
So explain that a little bit more.
The cost is spread out becausethe cyber criminals come back
and say, okay, you paid me once,you need to pay me again.
Or am I missing something?
Brian (03:42):
It's more about the cost and impact of operations when
you're losing all thatopportunity costs.
Opportunity costs and that's oneof the most important things
where a lot of companies areconsidering.
They're like, well, what if Ijust have backups right, like I
can beat the hackers?
I can beat people.
We're going to try and hold mydata ransom by having more and
more data in different storageareas, in different off premise
(04:04):
orpremise or in maybe differentcloud environments.
But those hackers are gettingsmarter, which means that
they're targeting backups moreand more.
That's one of the weedingtrends that we're seeing.
So what that does is it holdscompanies to account, saying
that you actually need to makesure that you're budgeting for
cybersecurity in this instance,because it's really a
requirement for basically modernenterprise operation.
Ryan (04:26):
Yeah, you can't afford to not spend, yeah, so what's?
We've talked about this before,but if you could discuss a
little bit more on the linkagebetween cybersecurity and the
need for cybersecurity and sortof this new frontier of
artificial intelligence, how arethey linked?
Brian (04:43):
Oh, they're certainly linked.
I mean, I think a lot of it is.
Cybersecurity's threatlandscape is already under
constant evolution.
It's always changing.
That's where solutions have hadto meet and match up to it.
But what artificialintelligence does is rapidly
increase that evolution timelineand lifecycle.
So when we think about whatartificial intelligence does, it
increases the speed ofransomware attacks.
(05:05):
We've seen an 1800 or sopercent increase in distributed
denial of service attacks.
That's usually bot generatedand we even see the idea that
phishing attacks which usuallyyou know it when you see it
right, when you get a bad email,you can tell that this isn't
addressed to me those emails aregetting a lot more
sophisticated.
If you haven't noticed thatthis isn't addressed to me,
those emails are getting a lotmore sophisticated, if you
haven't noticed, and a lot ofthat is it's turning large-scale
(05:27):
phishing attacks into spearphishing, which means they're
much more targeted to theindividual.
So you have a threat landscapefor artificial that is driven by
artificial intelligence that isgetting much more complex and
rapid.
And if you look at that fromthe negative side, what does
that mean?
In terms of the positive side?
Well, you can leverageartificial intelligence towards
maybe being better at automatedthreat response in particular.
Ryan (05:56):
So there's both sort of two sides of that AI coin when
it comes to its relationship tocybersecurity.
You've got the cybercrime side,where they're getting more
efficient and better at theirjobs, and then the response is
also getting better by usingsimilar types of resources.
Brian (06:06):
Exactly Like in anything in cybersecurity space, there's
always what you would call redteam or blue team.
Right, the red team that'sengaging in the hacking activity
, blue team is your defensiveunit.
Well, in this case, you'reprobably seeing a future where
artificial intelligence leads tored teams being made of bots
and blue teams being made ofbots in return, which then
effectively helps if you'redefending against an attack.
(06:27):
Triaging you get, so that yourhuman resources, which are
limited, because if you were totake every single certified
professional in the world andtry to fill out every position,
you'd only fill two-thirds ofthem.
That's how much of a gap thereis in terms of qualified
professionals on the cyber front.
So you really need thoseautomated detection units, and
there's companies likeCrowdStrike, sentinel One,
(06:49):
darktrace, that are really goodat these automated threat
responses.
They're very much AI firstcybersecurity companies at this
stage, and so what that does isit allows you to triage
effectively down so that yourhuman responders can respond to
the major breakthroughs andbreaches that require a little
bit more of that humancreativity side.
Ryan (07:10):
One of the things that I've begun to hear a little bit
more about and I think mostagree that it's still on.
Think of the relationshipbetween quantum computing and
cryptography and the ability toprotect, you know, digital
(07:31):
assets.
As a result of that, If you cancrack someone's password almost
instantaneously, all of asudden the digital world that we
live in becomes a verydangerous place.
So what's kind of coming on thehorizon related to quantum
computing and cybersecurity?
Brian (07:49):
Yeah, and we're talking about quantum computing.
It feels like it's very faraway and, let's be honest,
commercially viable.
Quantum computers in terms ofbeing widespread and in use is
probably a 2030s innovation.
But if you're going to be asavvy business or savvy
government in in this case tooright, because we're talking
about there's interest inincreased spend by governments
as well as enterprises you haveto start thinking about now what
(08:12):
the quantum landscape lookslike.
Quantum computing can reallybreak through traditional
encryption quite fast and quiteeasily.
So what you have to start doingis build what's called post
quantum cryptography.
So how do you start buildingand planning for that future?
Well, that's already happeningright now.
There's contests from groupslike the National Institutes of
Standards and Technology thatare really asking for companies
(08:33):
to put forward tools that canbuild and reflect this quantum
future.
How do they do that?
Well, oftentimes they're usingsimulated quantum environments
that use a cloud infrastructure.
So think like AWS or Azure, andyou can use AI to actually
simulate a quantum environment.
And while this sounds like it'sfar off, they've already found
(08:55):
two companies that have actuallybeen able to meet some
post-quantum cryptographystandards.
So you're already seeing someof the solutions be built now,
which should be a sigh of relieffor a lot of people Because,
again, the 2030s are comingsooner than we think.
We're already halfway into the2020s, so we'll be at a time of
quantum computers faster than weknow it.
Ryan (09:14):
Yeah, I've heard stories and you can maybe tell me if
this is actually happening or ifmaybe I'm worried about nothing
happening, or if maybe I'mworried about nothing, of cyber
criminals actually getting ahold of data that is protected.
It's encrypted data With thethought of later on, once I've
(09:38):
got quantum, I can break thatcode and actually have access to
the data.
So, even though you haveencrypted data that maybe even a
nation is stealing to store forlater, once they have access to
that quantum computing, is thatsomething that's actually
happening or am I worried aboutnothing?
Brian (09:53):
I would imagine that's probably a strategy that people
are deploying, so I don't thinkit's a worry out for nothing
Seems like a good.
Ryan (09:59):
hopefully I haven't given all the cyber criminals good
ideas now on this podcast.
Brian (10:03):
That is a pretty good idea.
Maybe you have a side careerthat I don't know about.
Ryan (10:07):
Okay.
So the bottom line is we're ata point now where every company,
every government agency,individuals everyone needs this
as a product or service, and itseems to me that the spending
has become non-discretionary inthis and your CTA.
(10:30):
You guys have come up with areally good term.
You call it digital utilities.
Can you talk a bit about whyyou call it digital utilities?
Brian (10:35):
Yeah, of course.
So we view cybersecurity, cloudcomputing and AI robotics as
the new digital utilities.
So the way that we view it islike water or electricity that
are required.
Say, you want to open abusiness, you sure want running
water and electricity to poweryour building and allow for good
facilities, right?
Well, in this case, we viewcybersecurity, cloud and AI as
(10:56):
all being part of this new waveof digital utilities, which are
requirements for any modernenterprise to operate in an
increasingly digital world.
So what they offer are what wewould say are three S's Security
, straightforward, right?
You want to make sure that yourdata is secure, that your
operations are operating in away that's not going to be
hacked anytime soon.
Cloud offers scalability theability of, say, a small
(11:19):
business to have greater reachthrough e-commerce tools, having
cloud infrastructure to hostall of their data from all of
their not just their customers,but also their employees.
And the last one is AI androbotics offer simulation.
We produce so much data on aday-to-day basis about 1.7
quintillion bytes, which is 1.7,followed by 18 zeros, which is
(11:41):
a number that's very hard tofathom.
Seems like a lot of data.
Seems like a lot of data.
Seems like a lot of data andit's impossible for humans to be
able to process on a givenbasis.
So you want AI robotics tosimulate human productivity as
much as you can.
So security, scalability andsimulation Okay.
Ryan (11:58):
So the CTA, that's, the Consumer Technology Association,
is, I believe, the largesttrade group for technology in
North America.
Is that correct?
That's correct.
You put on the ConsumerElectronics Show, or the CES.
Yep Is what it's called now,yep the.
Brian (12:14):
CES because about 40% of our exhibitors are doing
enterprise products.
Ryan (12:18):
So it's no longer just consumer.
Brian (12:19):
It's not just consumer and if you think about that,
that line's been blurring for along time.
This idea of well, aninnovation in the enterprise
will come to consumers.
I think 5G in particular, whereit was adopted by businesses
for rapid scale deployment ofdigital operations.
Then it really came to yourphones and got better
connectivity over time.
That's one example of it, butsometimes it might go the other
(12:42):
way.
Where smart glasses, forexample, you have meta-ray bands
that are quite popular withconsumers.
Well, there's enterpriseapplications in the long run,
like augmented reality onglasses, helping surgical
applications.
So it might be where you seethe popularization in the
consumer landscape actuallyinfluence the enterprise.
So that line only got blurred,I think, more during the
(13:02):
pandemic, where you had workfrom home tools, so common
consumer device, like an earbudsor AirPods, anything that you
can think of.
Well, those are also workdevices now, because it's how
many people are taking calls onthem and using them as an
enterprise product.
So it's really nice to be atthe nexus of it all at CES.
Ryan (13:32):
And CTA has an index, a thematic index program, which is
why we have First Trust, has arelationship with CTA and NASDAQ
, who create some of the indexesthat our funds track, and for
First Trust, one of the thingsthat's really valuable in that
relationship is sort of the edgethat you have in understanding
technology and understandingsome of the companies.
So could you talk a little bitabout that, the edge that you
have at somebody who's veryclose to the technology industry
?
How does that make you betterable to identify companies that
(13:55):
might make it into one of theseindexes?
Brian (13:57):
Yeah, well, I think it comes with some of it is the
Roots for our trade association.
So we represent over 1200members, so we have this ability
to see companies from yourlarge hyperscalers, your Amazon,
google's, but the majority ofour membership overwhelming,
like around 80% are startups andsmall businesses, so we see
that early stage ideations ofnew technologies occur, so
(14:18):
that's allows us to see theentire technology lifecycle.
You have CES, which is a globalshow.
About 40% of our attendees areactually from abroad, so they're
from APAC Europe.
Across the board.
We're able to not hold to justbeing a North American trade
association.
We can take a very global viewto innovation, which I think is
quite important when you thinkabout how technology diffuses
(14:39):
across the world.
But then I think one of themost important things about this
and this is why we love havingour partnership with NASDAQ when
we develop indexes NASDAQ is anincredible institution when it
comes to finance andquantitative measures.
They've been doing the indexbusiness.
They've been doing thefinancial business for decades.
We let them focus on that.
(15:01):
That's their expertise whichallows us and frees us as a team
myself included to focus on thetechnology itself and really
spend our time looking at patentportfolios, research and
development budgets, merger andacquisition activity in the tech
field.
Or our big one is calculatingthematic revenue, which is, how
much revenue do you derive froma given theme?
(15:21):
That's usually a prettyintensive manual calculation
because you're having a humancomb through financial
statements and take theirexpertise of a technology and
say, okay, this is how much theyget from cybersecurity or this
is how much they get fromrobotics.
So I think that's really whatsets it up is we can lean on the
partnership at the end of theday to let us be CTA and be
(15:43):
technology experts.
Ryan (15:44):
Yeah, so you're not.
I'll paraphrase what I justheard you say you're not
necessarily just looking at thefinancial statements, but
there's a little bit more.
There is the quantitativeaspect to it, but there's also a
qualitative understanding ofthe companies.
Brian (15:56):
Is that fair?
Exactly, I think that's a veryfair assessment.
I think qualitative is reallywhere our bread and butter is.
It's what we're leaning into.
Ryan (16:02):
Okay, so we've talked a bit about cybersecurity.
I want to talk a little bitabout what everyone's been
talking about over the lastcouple of years, and that's
artificial intelligence, andagain we touched on it a bit.
But as you kind of look atwhat's coming down the road over
the next year or two, what isit that you find to be kind of
most exciting about AI?
Brian (16:23):
Yeah, well, I think there's really three frontiers
of AI innovation that I'mexcited about.
The first is agentic AI.
So when we think about agents,right, these are basically AI
bots that can execute a task foryou with minimal or zero
oversight.
So they can move, say, a newhire right into your payroll,
your talent retention, withoutyou having to continuously move
(16:44):
all their data from app to appto app.
That's pretty revolutionary,because, as great as the app
economy is, it can be prettyinefficient at times if you have
to constantly move data fromsilo to silo.
Well, what if an agent can dothat for me?
That's the first one.
Digital twins is the second onethat really excites me, where
the idea that you can virtualizean object into a digital
environment and do AI-basedsimulation and scenario
(17:06):
generation means that if you'relimited on budget, you don't
have to sacrifice your researchand development potential.
That's the second one.
And then the third one isphysical AI, which would
basically mean robotics.
How do AI algorithms placed onrobots really allow for turning
it from being hard-coded?
Input A, output B to input A ofthis algorithm allows for
(17:29):
outputs B through infinity.
So those are the three pathwaysI see, and I can talk about any
one that you want to go with,but those are the three.
I think that excite me the most.
Ryan (17:38):
Yeah, People that I spoke with coming out of the most
recent CES in Las Vegas earlier.
Well, I guess it was the end oflast year, the beginning of
this year.
Brian (17:47):
Beginning of this year.
Ryan (17:48):
It's always in January, okay so one of the things that I
heard repeatedly was peoplewere really excited about some
of the robotic applications andyou know just some really cool
stuff that was going on there.
Let me push back on that for asecond, and I maybe I'm just a
bit paranoid about having ahumanoid robot roaming around my
(18:11):
house, so set my mind at ease.
Why is that a?
You know whether it's Optimistor one of these other robotic
companies.
You know why should I be moreoptimistic than I am?
And maybe I'm just I grew upwatching the Terminator and you
know RoboCop and some of that,so set my mind at ease.
Brian (18:33):
Yeah, I think a lot of it is.
It first starts in theenterprise.
So I think, like, let's startwith where it's occurring.
It's occurring in warehouses,it's occurring in the medical
sector, it's occurring in thehospitality sector, where you
have humanoid robots, inparticular, being designed to
fill workforce shortages, soit's helping make things more
efficient.
So, especially if you'rethinking about warehouse workers
moving packages that you mightwant delivered, if you enjoy
(18:54):
that two day delivery, this ishow we keep that in place by
having those workforce gaps met.
So that's the first one, Ithink.
On the consumer side, I thinkone of the big areas to think
about is aging populations.
Who's going to need a consumerside humanoid robot first?
Well, it might be as acaregiving.
We saw that actually at CESthis year.
(19:16):
There was a startup out ofFrance that focused on building
humanoid robots that wereactually anime design inspired,
so something a little bitdifferent than trying to make it
look like an artificial human,to go with something a little
bit more friendly, a little bitmore cartoonish, because they're
trying to put the mind at easeof the potential people that
(19:36):
they're taking care of.
And, as more Americans inparticular, and as we found age,
they rely on technologyincreasingly like they view
smart doorbells as healthcareproducts, because it's an
emergency camera system and itcan also alert people to when
they're you know they needsomething more than they are
currently getting with theircaregiving situation.
Ryan (19:57):
So is there a really good reason that some of the robotic
applications would have thatsort of humanoid form factor?
I mean, I'm perfectlycomfortable to have a Roomba
roaming around or vacuuming myhouse, but I mean, does it have
to walk on two legs and havearms?
I mean, does it have to walk ontwo legs and have arms?
(20:17):
And you know, it seems like arobot or some sort of robotic
application could do thesethings without necessarily
looking or having the form of ahuman.
Brian (20:27):
It's certainly possible, and there's companies like Rich
Tech Robotics that come to mindfrom the show four, where they
build mostly like they have onethat has arms and it looks like
a kind of like a star wars-esque.
Uh, bartender droid, it'scalled adam it works at the
texas rangers ballpark.
It's not quite human, but thearm movements aren't completely
human, because if you're gonnahave a robotic bartender, you
really want to make sure thatit's simulating the human motion
(20:49):
as best it can.
So of course, the joints aregoing to resemble that in
particular, but they haveservice sector and hospitality
sector robots that look morekind of on the r2d2 side, where
it's just carrying trays or ithas arms, but it's a moving like
kind of a tin can in a littlebit um, and so I think it
doesn't necessarily always needto be human-esque in the
(21:09):
performance.
It's just how can you simulatesomething that we expect it to
be human-like?
And so you say you'recomfortable with Roomba.
Well, we saw from Roborock, acompetitor of Roomba, them
developing vacuums with arms onboards so that the arm can move
objects out of the way moreefficiently, and so which is
exactly what a human would do inthis case, because sometimes I
(21:31):
think in the past we go well, Ican just do it.
This is sometimes something thatI find myself saying, where I'm
like I can do that better yeahlike I can vacuum and I'm gonna
get exactly what I want done theway I want it done, because I
have to move the cord out of theway, I have to move all these
things, yeah, um, out of it.
But you put an arm on a roboticvacuum and now you're actually
like, oh, it can actuallycompete with me quite well and I
(21:52):
would.
Ryan (21:53):
I would think maybe having a little bit more intelligence
in that vacuum could help itunderstand, because, quite
honestly, my Roomba usually getslost and we find it in the
laundry room and it's kind ofstranded a few days later Like
where is the Roomba?
And sure enough, it's gottenlost somewhere.
So I do think having some formof intelligence paired with
(22:13):
these robots would be animportant step.
Brian (22:15):
You have to.
And that's where that termphysical AI, which was
introduced on stage at CES byJensen Wong during his keynote
this past January, really comesinto mind, where robotics are
increasingly just the physicalembodiment of AI algorithms and
I think the Roomba examplegetting lost is a good one.
I tend to think about what itcan mean in the medical sector.
(22:35):
So companies like JohnsonJohnson, stryker and Intuitive
Surgical they've been usingrobots for joint replacements
and other applications for years.
Well, even though it soundslike a rinse and repeat process
a knee is a knee, a hip is a hipthat's not really quite true.
Every person's knee isdifferent than than others.
Every joint is individuallymade.
So if you have an AI algorithmthat can better scan and
(22:56):
understand what's going on onboard with the individual
patient, that robot becomes alot more catered for what would
be personalized medicine, whichfits very well into this
longevity and the overallhealthcare theme that's going on
right now too.
Ryan (23:11):
Yeah, it seems like there's a lot of linkages
between AI and healthcare.
The examples that you just gave, or maybe I've heard examples
of surgery related to cancer orsomething and making sure using
AI to somehow make sure thatyou've got clear margins when
you're removing a tumor orsomething like that, instead of
(23:32):
just relying on just, maybe,human input, a higher success
rate using AI.
Is that something else that yousee?
Brian (23:40):
Certainly, I think healthcare is probably one of
the best beneficiaries of AI ingeneral.
There's a company calledTempest AI that I think of a lot
, where they're using AI fordiagnosis, diagnostics as well
as drug discovery.
So them and Illumina are twocompanies that are really leaned
into the AI story to reallypush forward where the
(24:00):
healthcare sector can go.
Ryan (24:02):
You mentioned digital twins a minute ago and that
makes me think of all thedifferent industrial
applications and constructionapplications, and it just seems
like that is a huge set ofopportunities.
If you can virtualize somethingand do this sort of digital
twin work.
(24:22):
It seems like it would add someproductivity and some
efficiency.
So can you talk a bit aboutthat?
Brian (24:27):
Yeah, of course, and so when we think about digital twin
, again, this is a digitalrepresentation of it can be a
physical object, it can be anenvironment, it can be the
entire earth, like NVIDIA hasdone.
It can even just be a processthat you're visualizing.
So how does a factory flooroperation move across the board?
What it really does is unlockopportunity to save on physical
(24:49):
capital expenditure, to startbetter planning your allocation
of your resources.
So a company that comes to mindis actually someone like
Siemens, who they've done andthey've partnered on the
industrial applications, theconstruction side.
One of my favorite examples,though, is that they actually
create digital twins for the RedBull racing Formula One car.
(25:10):
So in Formula One, there's acost cap.
You can only spend so much oninnovation in a given, or so
much in a given year.
That includes how much goesinto the development of the car.
Well, if you're in thatmentality, you might start
cutting corners on how muchyou're putting into aerodynamic
testing or the like.
Well, in a digital twinenvironment, what they do is
(25:30):
they take sensors that are onthe car at all times, that feed
back into the digital twinversion so they can experiment
and pilot for a fraction of thecost, exactly what they want to
do and changes they want to makein the car.
So it creates this virtuouscycle of innovation where the
digital car is informing thephysical car, which has sensors
feeding back to the digitalversion of the car.
And that's just one element.
(25:52):
There's a healthcare story toothere, where companies like
Dassault can do it on the humanheart, which allows medical
researchers that maybe mighthave their budgets not be quite
what they used to be.
Right now they can use adigital twin of the heart to
practice surgery over and overagain, because you can't really
test on a cadaver only so manytimes.
And they have found they'veactually pioneered some new
(26:13):
pediatric cardiology surgerytechniques that have saved a few
thousand lives already.
Ryan (26:20):
It strikes me that that is basically kind of like a really
complex video game, yeah, andthe background of a lot of this
is actually in video gamedevelopment, yeah, which is, you
know, I think, is kind.
Of.
I don't know if it's surprising, but it, you know, growing up
we were always told, like, videogames are, you know, a waste of
(26:41):
time or something like that.
But, as it turns out, the samesort of technology that was
really invested in because ofvideo games is what has, it
seems like down the road,resulted in some of the AI that
we're seeing today.
Brian (26:55):
Yeah, some of the stuff that you're seeing, like Epic's
Unreal Engine, some of the stuffthat you're seeing from Unity
Technologies in particular, theyreally start to, I think,
improve the virtualization fieldin a way that we haven't quite
seen.
And what's really cool that youmentioned is Siemens.
The person that they partnerwith on a lot of the
visualizations is Sony, themaker of the PlayStation, so
it's not surprising that gamingexpertise is being leveraged for
(27:17):
these high fidelity versions.
Visualizations you go to gaming, I think of every time.
I always go to like somethinglike iron man and how, when he
first builds his iron man suit,he actually uses a digital twin
in real time of moving out themark pieces of the suit and in
my mind I'm like that was 2008when that movie came out, I
think, and so we're alreadyliving in a future where that
(27:38):
kind of exists to a degree.
You're just putting on a headsetInteresting.
Ryan (27:43):
Well, and then the other digital utility that you kind of
linked in with this was cloudcomputing, and it seems like
none of this would really bepossible without that cloud
ecosystem.
Can you talk a little bit aboutthat?
Brian (27:57):
Yeah, I think in a lot of ways, we've talked so much
about AI rightfully so but weoverlook how important cloud
computing is in terms ofcreating the infrastructure that
enables us in the first place.
When we think about AI whatenables artificial intelligence
we tend to go right to thesilicon and the sensors.
We go, okay, these are the mostadvanced chips and processing
(28:17):
units, but think about chipsalmost as the brain.
Data is the language throughwhich those brains communicate.
Well, who's infrastructuringthat data?
What's doing that?
Well, that's cloud computing.
So that is what your Amazons,your Googles, your Microsofts,
your Oracles are all doing totry and enable this revolution,
and we expect by 2030, based onAI demand alone that it will be
(28:40):
about $2 trillion in spend oncloud or in terms of total
revenue for the space.
So this is a market We've beentalking in billions for this
entire podcast so far.
Now we're mentioning trillionsbecause that's where cloud
computing sits in terms of itsimportance to enabling the
revolution that we're seeingright now.
Ryan (28:57):
Yeah, and all of the massive capital investments that
have taken place.
I mean, there's still plentythat will take place in data
centers and all these hugecomplex warehouses full of GPUs
and server farms and things likethis, and all of that is
(29:20):
operated by these cloudproviders.
Is that correct?
Brian (29:23):
That is correct, and then they're also making deals
directly with some othercompanies that help support GPU
compute as a service.
So something that we've beenwatching really closely is how
does a company like CoreWeave,who has deals with NVIDIA, start
to better partner with some ofthe cloud providers in terms of
dialing up that compute as aservice overall?
And so one of the questionsthat we're financially asking is
(29:44):
is that more of a software as aservice or is it actually
closer to something like aninfrastructure as a service
where it's really pioneering,allowing for the operation of
the cloud as a whole?
Probably right now I'd leanmore to software as a service,
but ask me in a few months andmaybe that shift changes as we
start to really understand howthe layers of cloud are evolving
(30:04):
alongside AI, because that'sthe thing we tend to think about
with cloud, and why I feelsometimes it gets overlooked is,
I think it gets treated as thisstatic innovation, that it's
not as important anymore, butit's vital, it's critical and it
has to evolve alongside theneeds and demands of data center
, as well as power and energydemands.
Ryan (30:22):
Yeah, and glad you brought that up, because I think that's
another important point when wethink about, whether it's
robotics or all these thingstake electricity.
They take a massive amount ofpower.
All the data centers take amassive amount of power.
And I'm just curious on yourthought.
I know you're not an energy guyper se, but how do you think,
10 years from now, how are wegoing to be able to scale up the
(30:46):
power resources in order toprovide enough electricity?
Brian (30:50):
Well, it was funny, I actually did.
My first clients when I firststarted my career were actually
in nuclear energy and naturalgas, and so what I find
interesting is about that wasover a decade ago.
It was a very differentconversation where the terms all
of the above energy was notbeing used as frequently.
It was a term that existed, butit wasn't used in the media as
much.
(31:10):
And so what I find with thedemands is you'll see a probably
135% increase in data centerpower demand from AI alone by
the end of the decade, and soyou've seen a lot of energy come
on board from solar and wind,which are really incredible
renewables, but working from thenuclear side.
Nuclear is always on.
It's higher in terms of energydensity, so you need less
(31:32):
uranium to produce as more interms of terawatts of energy
than solar and wind can do.
But it's also costly and ittakes time.
So there's a few ways to do it.
I know this is something thatyou talked about a lot and it's
and it makes me happy to seethat this conversation is
happening again which is you'reseeing hyperscalers Microsoft,
(31:52):
google and Amazon engaged inthis capacity of saying well,
can we just dial back on nuclearpower plants that maybe have
shut down, which there were alot of closures a decade ago or
so?
There's not a lot more comingonline.
So we're trying to dial back onthe power that we have now and
we're trying to think about well, what does that look like in a
(32:13):
decade from now?
Well, there's a play of is itgoing to be these large power
plants or is it going to besomething like a small modular
reactor, which there's a lot ofpatents for them?
There's a lot of deals outthere, but they're probably not
commercially ready for anotherdecade or so.
And that's just fission.
There is the companies that areworking towards fusion, which
we've had some breakthroughs infrom the Department of Energy a
(32:35):
couple of years ago.
With ignition, that workforcehas increased by 200% in the
private sector, so clearly thatthere's more interest than ever
that maybe fusion's closer.
I think fusion requires AI andquantum to get dialed up a bit
more for it to become viable andit's probably something that's
more of a 2040s technology.
So what do you do in theinterim?
Well, it's probably an all ofthe above energy strategy of
(32:57):
renewables, nuclear and,probably to a degree, gas in the
mix.
Ryan (33:00):
Yeah, it does seem like it's not something that we can
wait until 2040 for the fusiontechnology that's always just on
the horizon and we never quiteget to and hopefully some of the
breakthroughs that people havetalked about become realized and
they can actually.
You know some of them.
There's some pretty massivespending that's taking place to
(33:22):
advance fusion technology around, you know, in Massachusetts and
some other parts of the country, so that's encouraging.
But you know to your point.
I do think there has to be somespending over the next five
years to actually increase theelectricity capacity in the US
to compete globally, especiallywith China.
Brian (33:43):
Yeah, I think that's the biggest thing to focus on is
really how do we build up thiscapacity and I think some of it
too is.
And what I find fascinating inthis story was mentioning the
hyperscalers making these dealsthemselves.
We saw a shift with AI ingeneral, with the headwinds that
were in the merger andacquisition market where they
couldn't really acquire asfrequently in the last few years
(34:07):
Companies across the board.
They got into the investormindset where Microsoft invested
heavily into open AI.
You saw Amazon heavily investin AI companies.
Well, now they're putting on anew hat, which is we're power
purchasers, which I think isvery fascinating.
Which how does that translateinto it?
It might be a competitionthat's either driven by
(34:27):
geopolitical competition or itmight be just this is the table
stakes and required for thefuture to bet on it.
So it'll be a complex landscapefor sure.
Ryan (34:36):
It is complex and it adds a level of complexity and
difference to the traditionalasset-light business model of
many technology companies.
If they're having to makecapital investments to a degree
that they have never made before, that transforms the sort of
business that they're operating,and so time will tell to see
how that plays out financiallyand seeing what sort of return
(34:59):
on investment these companieshave with those capital
investments.
But I do think that they'rekind of in a position where they
don't have a choice.
To some degree they have tomake those investments if they
want to be able to advance thetechnology and compete and have,
if they want to be the winnerin 10 years from now.
Yeah, okay.
(35:20):
Well, brian, this has been agreat conversation.
Appreciate you coming on thepodcast again.
I do have a final question foryou.
I want you to put on your youknow futurist cap, okay, okay,
so we've talked about a lot ofdifferent technologies.
What is you kind of thinkforward, looking into the future
(35:41):
, a decade?
Is there any specific thingthat we haven't talked about
that you think is going to be?
You know that maybe fewerpeople are focused on right now
that, looking back a decade fromnow, we're going to say, wow,
that was an amazing advance andamazing innovation and nobody
really saw it coming.
Brian (35:59):
I think it's one of the things that we talked about.
I'm happy we already talkedabout robotics, quantum and
nuclear, because those are thetable stakes, ones that are
quite popular.
But I did mention one verybriefly in the visualization,
one which is, I think, smartglasses are oftentimes
overlooked, where there's veryfew ideas, where you already had
the start and stumble about adecade ago with smart glasses,
(36:20):
of where they're going.
But with AI capabilities onboard, translation capabilities,
augmented reality technologygetting to a point where it's
able to better be customized toa comfortable form factor that
we're seeing, I think we mightbe in a decade seeing a point
where there's a greaterproliferation of smart glasses
on the market than we expected,because I think it's the only
(36:42):
device, by a form factorcomparison, that might be able
to compete with a smartphone.
I don't think it replaces thesmartphone outright in a decade,
but I think it's the onlydevice that could possibly
compete with it.
In a way.
It will come down to certainlylooking at how many more models
come on board.
We're already seeing Oakley isnow going to come out with their
version to compete with Meta'sRay-Ban, or I should say
(37:05):
complement Meta Ray-Bans,because Essilor Lungsotica owns
all those brands, but I thinkthat's one that gets a little
bit overlooked, and I'm excitedto see a lot more in terms of
that innovation in a decade.
Ryan (37:15):
So are those joint ventures typically that take
place where it's, you know, thesunglasses company and the
technology company are kind ofworking on a product together.
Or do you see, you know, Appleor Meta or one of these
companies trying to kind of puttheir own glasses out there?
Brian (37:31):
I think right now they're leaning into the partnership
approach, because part of whenyou get into something that has
a consumer implication like thatis you want it to look cool and
I think, like I've seen metaRay-Bans in the wild and they
look like Ray-Bans, they looklike a good sunglasses model.
So I think that there's a lotof strength in that partnership
approach.
So, as much as it's going to bea technology that maybe
(37:55):
redefines translation,communication and computing in
different ways, it might also bea fashion statement which is
pretty powerful in a lot of ways.
So there's a cultural elementto it which is important,
because technology doesn't existin a vacuum.
It's adopted by people at theend of the day.
Ryan (38:11):
That is a really interesting insight and we'll
look forward to seeing yourtranslation example.
I think is really interesting,as somebody who travels
internationally quite a bit,having the ability to have
real-time translation comingfrom your glasses into your ear,
instead of having to hold up aphone or something like that and
say, you know, speak into thisso I can get the translation.
(38:31):
Yeah, that seems like a reallyinteresting application.
Brian (38:35):
Yeah, I think it would be useful.
As someone who also travelsinternationally quite a bit I
that is probably like the mostin demand.
As much as I love learning newlanguages, it's really hard, and
so having something that can atleast assist that and put that
at ease is incredible.
Ryan (38:49):
All right.
Well, brian Komsky, thanks forjoining us again on the First
Trust ROI podcast, and thanks toyou as well for joining us.
We will see you next time.
Hi, welcome to this episode of the First Trust ROI
podcast.
I'm Ryan Isakainen, etfstrategist at First Trust.
Today I'm joined by BrianKomsky, senior Director of
Innovation and Trends at theConsumer Technology Association.
Brian is an expert ontechnology and that's what we're
going to discuss.
We'll talk about cybersecurity,robotics, artificial
intelligence, cloud computingand a variety of other topics.
(00:31):
Thanks for joining us.
So, looking around the world,there's quite a bit of stress,
there's quite a bit of conflict,and one of the things that
strikes me is that the newfrontier of war isn't
necessarily kinetic war, butmore information and cyber war,
a lot of potential cyber threatsfrom some of the players on the
(00:54):
geopolitical stage.
So my question for you, brian,is, as you kind of think about
cybersecurity quite a bit andsome of the related technologies
, how do you expect that willactually impact spending on
cybersecurity from companies,from nations?
Do you think that'll have animpact?
Brian (01:14):
Certainly, and part of the experience that I like to
draw upon and think about is Iused to do some consulting with
the US Coast Guard, and one ofthe areas that was a major focus
of my time there was buildingout new policies around
cybersecurity and cyberoperations, because they
recognized that we were movingfrom three traditional theaters
or kinetic theaters of war, land, sea and air into a fourth
(01:35):
theater that was equallyimportant, which was the
cyberspace, and so you alreadysaw that investment going in
from the US government side, andthat's something that's
definitely replicated throughoutthe world, where you're seeing
budgets increase in the EU,you're seeing those budgets
increase in APAC region, and socertainly governments are
recognizing that they need toplay a role in amping up
(01:55):
cybersecurity from the idea thatthere are nation states that
might attack them in that regard.
But one of my favorite stats isonly 25% of cyber attacks are
state sponsored.
Most of them about 75% arefinancially motivated, which
means that they're notnecessarily targeting a nation
state.
They might be targeting abusiness in this case, and if
(02:17):
it's such a lucrative illiciteconomy that emerges, then
companies need to make sure thatthey're spending in order to
protect from themselves againstit.
So we expect about like $212billion in information security
spending this year alone perGartner.
Ryan (02:31):
So is that ransomware sort of attacks?
What's the financial motivewhen you're talking about
financial motives, is itransomware?
Is it corporate espionage?
What sort of financial motivesare we talking about?
Brian (02:42):
It can be a bit of both Corporate espionage, for sure
but that usually means thatyou've got to pull back a layer
and there might be a nationstate behind that.
But usually it's ransomwarewhich has a lot of payments.
And if you look at the statsgoing in from 2023 to 2024, you
saw a 5x increase in the mediumransomware payout, which, when
we think about what thosepayouts look like, it's not just
(03:03):
oh, here's the money that I'mspending in millions of dollars
to get my data back, but you'vehad an impact to your operations
as a business, which means,uh-oh, I have more costs that
are associated with it.
So when you think about payingout ransomware, it's not
actually just a one-time onlypayment.
It's usually a three-year costwindow that you're paying out,
(03:23):
where you're paying 66% of yourtotal cost from the impact the
first year and then another 22%second year, and then you finish
out with that 11% in your thirdyear.
Ryan (03:31):
So explain that a little bit more.
The cost is spread out becausethe cyber criminals come back
and say, okay, you paid me once,you need to pay me again.
Or am I missing something?
Brian (03:42):
It's more about the cost and impact of operations when
you're losing all thatopportunity costs.
Opportunity costs and that's oneof the most important things
where a lot of companies areconsidering.
They're like, well, what if Ijust have backups right, like I
can beat the hackers?
I can beat people.
We're going to try and hold mydata ransom by having more and
more data in different storageareas, in different off premise
(04:04):
orpremise or in maybe differentcloud environments.
But those hackers are gettingsmarter, which means that
they're targeting backups moreand more.
That's one of the weedingtrends that we're seeing.
So what that does is it holdscompanies to account, saying
that you actually need to makesure that you're budgeting for
cybersecurity in this instance,because it's really a
requirement for basically modernenterprise operation.
Ryan (04:26):
Yeah, you can't afford to not spend, yeah, so what's?
We've talked about this before,but if you could discuss a
little bit more on the linkagebetween cybersecurity and the
need for cybersecurity and sortof this new frontier of
artificial intelligence, how arethey linked?
Brian (04:43):
Oh, they're certainly linked.
I mean, I think a lot of it is.
Cybersecurity's threatlandscape is already under
constant evolution.
It's always changing.
That's where solutions have hadto meet and match up to it.
But what artificialintelligence does is rapidly
increase that evolution timelineand lifecycle.
So when we think about whatartificial intelligence does, it
increases the speed ofransomware attacks.
(05:05):
We've seen an 1800 or sopercent increase in distributed
denial of service attacks.
That's usually bot generatedand we even see the idea that
phishing attacks which usuallyyou know it when you see it
right, when you get a bad email,you can tell that this isn't
addressed to me those emails aregetting a lot more
sophisticated.
If you haven't noticed thatthis isn't addressed to me,
those emails are getting a lotmore sophisticated, if you
haven't noticed, and a lot ofthat is it's turning large-scale
(05:27):
phishing attacks into spearphishing, which means they're
much more targeted to theindividual.
So you have a threat landscapefor artificial that is driven by
artificial intelligence that isgetting much more complex and
rapid.
And if you look at that fromthe negative side, what does
that mean?
In terms of the positive side?
Well, you can leverageartificial intelligence towards
maybe being better at automatedthreat response in particular.
Ryan (05:56):
So there's both sort of two sides of that AI coin when
it comes to its relationship tocybersecurity.
You've got the cybercrime side,where they're getting more
efficient and better at theirjobs, and then the response is
also getting better by usingsimilar types of resources.
Brian (06:06):
Exactly Like in anything in cybersecurity space, there's
always what you would call redteam or blue team.
Right, the red team that'sengaging in the hacking activity
, blue team is your defensiveunit.
Well, in this case, you'reprobably seeing a future where
artificial intelligence leads tored teams being made of bots
and blue teams being made ofbots in return, which then
effectively helps if you'redefending against an attack.
(06:27):
Triaging you get, so that yourhuman resources, which are
limited, because if you were totake every single certified
professional in the world andtry to fill out every position,
you'd only fill two-thirds ofthem.
That's how much of a gap thereis in terms of qualified
professionals on the cyber front.
So you really need thoseautomated detection units, and
there's companies likeCrowdStrike, sentinel One,
(06:49):
darktrace, that are really goodat these automated threat
responses.
They're very much AI firstcybersecurity companies at this
stage, and so what that does isit allows you to triage
effectively down so that yourhuman responders can respond to
the major breakthroughs andbreaches that require a little
bit more of that humancreativity side.
Ryan (07:10):
One of the things that I've begun to hear a little bit
more about and I think mostagree that it's still on.
Think of the relationshipbetween quantum computing and
cryptography and the ability toprotect, you know, digital
(07:31):
assets.
As a result of that, If you cancrack someone's password almost
instantaneously, all of asudden the digital world that we
live in becomes a verydangerous place.
So what's kind of coming on thehorizon related to quantum
computing and cybersecurity?
Brian (07:49):
Yeah, and we're talking about quantum computing.
It feels like it's very faraway and, let's be honest,
commercially viable.
Quantum computers in terms ofbeing widespread and in use is
probably a 2030s innovation.
But if you're going to be asavvy business or savvy
government in in this case tooright, because we're talking
about there's interest inincreased spend by governments
as well as enterprises you haveto start thinking about now what
(08:12):
the quantum landscape lookslike.
Quantum computing can reallybreak through traditional
encryption quite fast and quiteeasily.
So what you have to start doingis build what's called post
quantum cryptography.
So how do you start buildingand planning for that future?
Well, that's already happeningright now.
There's contests from groupslike the National Institutes of
Standards and Technology thatare really asking for companies
(08:33):
to put forward tools that canbuild and reflect this quantum
future.
How do they do that?
Well, oftentimes they're usingsimulated quantum environments
that use a cloud infrastructure.
So think like AWS or Azure, andyou can use AI to actually
simulate a quantum environment.
And while this sounds like it'sfar off, they've already found
(08:55):
two companies that have actuallybeen able to meet some
post-quantum cryptographystandards.
So you're already seeing someof the solutions be built now,
which should be a sigh of relieffor a lot of people Because,
again, the 2030s are comingsooner than we think.
We're already halfway into the2020s, so we'll be at a time of
quantum computers faster than weknow it.
Ryan (09:14):
Yeah, I've heard stories and you can maybe tell me if
this is actually happening or ifmaybe I'm worried about nothing
happening, or if maybe I'mworried about nothing, of cyber
criminals actually getting ahold of data that is protected.
It's encrypted data With thethought of later on, once I've
(09:38):
got quantum, I can break thatcode and actually have access to
the data.
So, even though you haveencrypted data that maybe even a
nation is stealing to store forlater, once they have access to
that quantum computing, is thatsomething that's actually
happening or am I worried aboutnothing?
Brian (09:53):
I would imagine that's probably a strategy that people
are deploying, so I don't thinkit's a worry out for nothing
Seems like a good.
Ryan (09:59):
hopefully I haven't given all the cyber criminals good
ideas now on this podcast.
Brian (10:03):
That is a pretty good idea.
Maybe you have a side careerthat I don't know about.
Ryan (10:07):
Okay.
So the bottom line is we're ata point now where every company,
every government agency,individuals everyone needs this
as a product or service, and itseems to me that the spending
has become non-discretionary inthis and your CTA.
(10:30):
You guys have come up with areally good term.
You call it digital utilities.
Can you talk a bit about whyyou call it digital utilities?
Brian (10:35):
Yeah, of course.
So we view cybersecurity, cloudcomputing and AI robotics as
the new digital utilities.
So the way that we view it islike water or electricity that
are required.
Say, you want to open abusiness, you sure want running
water and electricity to poweryour building and allow for good
facilities, right?
Well, in this case, we viewcybersecurity, cloud and AI as
(10:56):
all being part of this new waveof digital utilities, which are
requirements for any modernenterprise to operate in an
increasingly digital world.
So what they offer are what wewould say are three S's Security
, straightforward, right?
You want to make sure that yourdata is secure, that your
operations are operating in away that's not going to be
hacked anytime soon.
Cloud offers scalability theability of, say, a small
(11:19):
business to have greater reachthrough e-commerce tools, having
cloud infrastructure to hostall of their data from all of
their not just their customers,but also their employees.
And the last one is AI androbotics offer simulation.
We produce so much data on aday-to-day basis about 1.7
quintillion bytes, which is 1.7,followed by 18 zeros, which is
(11:41):
a number that's very hard tofathom.
Seems like a lot of data.
Seems like a lot of data.
Seems like a lot of data andit's impossible for humans to be
able to process on a givenbasis.
So you want AI robotics tosimulate human productivity as
much as you can.
So security, scalability andsimulation Okay.
Ryan (11:58):
So the CTA, that's, the Consumer Technology Association,
is, I believe, the largesttrade group for technology in
North America.
Is that correct?
That's correct.
You put on the ConsumerElectronics Show, or the CES.
Yep Is what it's called now,yep the.
Brian (12:14):
CES because about 40% of our exhibitors are doing
enterprise products.
Ryan (12:18):
So it's no longer just consumer.
Brian (12:19):
It's not just consumer and if you think about that,
that line's been blurring for along time.
This idea of well, aninnovation in the enterprise
will come to consumers.
I think 5G in particular, whereit was adopted by businesses
for rapid scale deployment ofdigital operations.
Then it really came to yourphones and got better
connectivity over time.
That's one example of it, butsometimes it might go the other
(12:42):
way.
Where smart glasses, forexample, you have meta-ray bands
that are quite popular withconsumers.
Well, there's enterpriseapplications in the long run,
like augmented reality onglasses, helping surgical
applications.
So it might be where you seethe popularization in the
consumer landscape actuallyinfluence the enterprise.
So that line only got blurred,I think, more during the
(13:02):
pandemic, where you had workfrom home tools, so common
consumer device, like an earbudsor AirPods, anything that you
can think of.
Well, those are also workdevices now, because it's how
many people are taking calls onthem and using them as an
enterprise product.
So it's really nice to be atthe nexus of it all at CES.
Ryan (13:32):
And CTA has an index, a thematic index program, which is
why we have First Trust, has arelationship with CTA and NASDAQ
, who create some of the indexesthat our funds track, and for
First Trust, one of the thingsthat's really valuable in that
relationship is sort of the edgethat you have in understanding
technology and understandingsome of the companies.
So could you talk a little bitabout that, the edge that you
have at somebody who's veryclose to the technology industry
?
How does that make you betterable to identify companies that
(13:55):
might make it into one of theseindexes?
Brian (13:57):
Yeah, well, I think it comes with some of it is the
Roots for our trade association.
So we represent over 1200members, so we have this ability
to see companies from yourlarge hyperscalers, your Amazon,
google's, but the majority ofour membership overwhelming,
like around 80% are startups andsmall businesses, so we see
that early stage ideations ofnew technologies occur, so
(14:18):
that's allows us to see theentire technology lifecycle.
You have CES, which is a globalshow.
About 40% of our attendees areactually from abroad, so they're
from APAC Europe.
Across the board.
We're able to not hold to justbeing a North American trade
association.
We can take a very global viewto innovation, which I think is
quite important when you thinkabout how technology diffuses
(14:39):
across the world.
But then I think one of themost important things about this
and this is why we love havingour partnership with NASDAQ when
we develop indexes NASDAQ is anincredible institution when it
comes to finance andquantitative measures.
They've been doing the indexbusiness.
They've been doing thefinancial business for decades.
We let them focus on that.
(15:01):
That's their expertise whichallows us and frees us as a team
myself included to focus on thetechnology itself and really
spend our time looking at patentportfolios, research and
development budgets, merger andacquisition activity in the tech
field.
Or our big one is calculatingthematic revenue, which is, how
much revenue do you derive froma given theme?
(15:21):
That's usually a prettyintensive manual calculation
because you're having a humancomb through financial
statements and take theirexpertise of a technology and
say, okay, this is how much theyget from cybersecurity or this
is how much they get fromrobotics.
So I think that's really whatsets it up is we can lean on the
partnership at the end of theday to let us be CTA and be
(15:43):
technology experts.
Ryan (15:44):
Yeah, so you're not.
I'll paraphrase what I justheard you say you're not
necessarily just looking at thefinancial statements, but
there's a little bit more.
There is the quantitativeaspect to it, but there's also a
qualitative understanding ofthe companies.
Brian (15:56):
Is that fair?
Exactly, I think that's a veryfair assessment.
I think qualitative is reallywhere our bread and butter is.
It's what we're leaning into.
Ryan (16:02):
Okay, so we've talked a bit about cybersecurity.
I want to talk a little bitabout what everyone's been
talking about over the lastcouple of years, and that's
artificial intelligence, andagain we touched on it a bit.
But as you kind of look atwhat's coming down the road over
the next year or two, what isit that you find to be kind of
most exciting about AI?
Brian (16:23):
Yeah, well, I think there's really three frontiers
of AI innovation that I'mexcited about.
The first is agentic AI.
So when we think about agents,right, these are basically AI
bots that can execute a task foryou with minimal or zero
oversight.
So they can move, say, a newhire right into your payroll,
your talent retention, withoutyou having to continuously move
(16:44):
all their data from app to appto app.
That's pretty revolutionary,because, as great as the app
economy is, it can be prettyinefficient at times if you have
to constantly move data fromsilo to silo.
Well, what if an agent can dothat for me?
That's the first one.
Digital twins is the second onethat really excites me, where
the idea that you can virtualizean object into a digital
environment and do AI-basedsimulation and scenario
(17:06):
generation means that if you'relimited on budget, you don't
have to sacrifice your researchand development potential.
That's the second one.
And then the third one isphysical AI, which would
basically mean robotics.
How do AI algorithms placed onrobots really allow for turning
it from being hard-coded?
Input A, output B to input A ofthis algorithm allows for
(17:29):
outputs B through infinity.
So those are the three pathwaysI see, and I can talk about any
one that you want to go with,but those are the three.
I think that excite me the most.
Ryan (17:38):
Yeah, People that I spoke with coming out of the most
recent CES in Las Vegas earlier.
Well, I guess it was the end oflast year, the beginning of
this year.
Brian (17:47):
Beginning of this year.
Ryan (17:48):
It's always in January, okay so one of the things that I
heard repeatedly was peoplewere really excited about some
of the robotic applications andyou know just some really cool
stuff that was going on there.
Let me push back on that for asecond, and I maybe I'm just a
bit paranoid about having ahumanoid robot roaming around my
(18:11):
house, so set my mind at ease.
Why is that a?
You know whether it's Optimistor one of these other robotic
companies.
You know why should I be moreoptimistic than I am?
And maybe I'm just I grew upwatching the Terminator and you
know RoboCop and some of that,so set my mind at ease.
Brian (18:33):
Yeah, I think a lot of it is.
It first starts in theenterprise.
So I think, like, let's startwith where it's occurring.
It's occurring in warehouses,it's occurring in the medical
sector, it's occurring in thehospitality sector, where you
have humanoid robots, inparticular, being designed to
fill workforce shortages, soit's helping make things more
efficient.
So, especially if you'rethinking about warehouse workers
moving packages that you mightwant delivered, if you enjoy
(18:54):
that two day delivery, this ishow we keep that in place by
having those workforce gaps met.
So that's the first one, Ithink.
On the consumer side, I thinkone of the big areas to think
about is aging populations.
Who's going to need a consumerside humanoid robot first?
Well, it might be as acaregiving.
We saw that actually at CESthis year.
(19:16):
There was a startup out ofFrance that focused on building
humanoid robots that wereactually anime design inspired,
so something a little bitdifferent than trying to make it
look like an artificial human,to go with something a little
bit more friendly, a little bitmore cartoonish, because they're
trying to put the mind at easeof the potential people that
(19:36):
they're taking care of.
And, as more Americans inparticular, and as we found age,
they rely on technologyincreasingly like they view
smart doorbells as healthcareproducts, because it's an
emergency camera system and itcan also alert people to when
they're you know they needsomething more than they are
currently getting with theircaregiving situation.
Ryan (19:57):
So is there a really good reason that some of the robotic
applications would have thatsort of humanoid form factor?
I mean, I'm perfectlycomfortable to have a Roomba
roaming around or vacuuming myhouse, but I mean, does it have
to walk on two legs and havearms?
I mean, does it have to walk ontwo legs and have arms?
(20:17):
And you know, it seems like arobot or some sort of robotic
application could do thesethings without necessarily
looking or having the form of ahuman.
Brian (20:27):
It's certainly possible, and there's companies like Rich
Tech Robotics that come to mindfrom the show four, where they
build mostly like they have onethat has arms and it looks like
a kind of like a star wars-esque.
Uh, bartender droid, it'scalled adam it works at the
texas rangers ballpark.
It's not quite human, but thearm movements aren't completely
human, because if you're gonnahave a robotic bartender, you
really want to make sure thatit's simulating the human motion
(20:49):
as best it can.
So of course, the joints aregoing to resemble that in
particular, but they haveservice sector and hospitality
sector robots that look morekind of on the r2d2 side, where
it's just carrying trays or ithas arms, but it's a moving like
kind of a tin can in a littlebit um, and so I think it
doesn't necessarily always needto be human-esque in the
(21:09):
performance.
It's just how can you simulatesomething that we expect it to
be human-like?
And so you say you'recomfortable with Roomba.
Well, we saw from Roborock, acompetitor of Roomba, them
developing vacuums with arms onboards so that the arm can move
objects out of the way moreefficiently, and so which is
exactly what a human would do inthis case, because sometimes I
(21:31):
think in the past we go well, Ican just do it.
This is sometimes something thatI find myself saying, where I'm
like I can do that better yeahlike I can vacuum and I'm gonna
get exactly what I want done theway I want it done, because I
have to move the cord out of theway, I have to move all these
things, yeah, um, out of it.
But you put an arm on a roboticvacuum and now you're actually
like, oh, it can actuallycompete with me quite well and I
(21:52):
would.
Ryan (21:53):
I would think maybe having a little bit more intelligence
in that vacuum could help itunderstand, because, quite
honestly, my Roomba usually getslost and we find it in the
laundry room and it's kind ofstranded a few days later Like
where is the Roomba?
And sure enough, it's gottenlost somewhere.
So I do think having some formof intelligence paired with
(22:13):
these robots would be animportant step.
Brian (22:15):
You have to.
And that's where that termphysical AI, which was
introduced on stage at CES byJensen Wong during his keynote
this past January, really comesinto mind, where robotics are
increasingly just the physicalembodiment of AI algorithms and
I think the Roomba examplegetting lost is a good one.
I tend to think about what itcan mean in the medical sector.
(22:35):
So companies like JohnsonJohnson, stryker and Intuitive
Surgical they've been usingrobots for joint replacements
and other applications for years.
Well, even though it soundslike a rinse and repeat process
a knee is a knee, a hip is a hipthat's not really quite true.
Every person's knee isdifferent than than others.
Every joint is individuallymade.
So if you have an AI algorithmthat can better scan and
(22:56):
understand what's going on onboard with the individual
patient, that robot becomes alot more catered for what would
be personalized medicine, whichfits very well into this
longevity and the overallhealthcare theme that's going on
right now too.
Ryan (23:11):
Yeah, it seems like there's a lot of linkages
between AI and healthcare.
The examples that you just gave, or maybe I've heard examples
of surgery related to cancer orsomething and making sure using
AI to somehow make sure thatyou've got clear margins when
you're removing a tumor orsomething like that, instead of
(23:32):
just relying on just, maybe,human input, a higher success
rate using AI.
Is that something else that yousee?
Brian (23:40):
Certainly, I think healthcare is probably one of
the best beneficiaries of AI ingeneral.
There's a company calledTempest AI that I think of a lot
, where they're using AI fordiagnosis, diagnostics as well
as drug discovery.
So them and Illumina are twocompanies that are really leaned
into the AI story to reallypush forward where the
(24:00):
healthcare sector can go.
Ryan (24:02):
You mentioned digital twins a minute ago and that
makes me think of all thedifferent industrial
applications and constructionapplications, and it just seems
like that is a huge set ofopportunities.
If you can virtualize somethingand do this sort of digital
twin work.
(24:22):
It seems like it would add someproductivity and some
efficiency.
So can you talk a bit aboutthat?
Brian (24:27):
Yeah, of course, and so when we think about digital twin
, again, this is a digitalrepresentation of it can be a
physical object, it can be anenvironment, it can be the
entire earth, like NVIDIA hasdone.
It can even just be a processthat you're visualizing.
So how does a factory flooroperation move across the board?
What it really does is unlockopportunity to save on physical
(24:49):
capital expenditure, to startbetter planning your allocation
of your resources.
So a company that comes to mindis actually someone like
Siemens, who they've done andthey've partnered on the
industrial applications, theconstruction side.
One of my favorite examples,though, is that they actually
create digital twins for the RedBull racing Formula One car.
(25:10):
So in Formula One, there's acost cap.
You can only spend so much oninnovation in a given, or so
much in a given year.
That includes how much goesinto the development of the car.
Well, if you're in thatmentality, you might start
cutting corners on how muchyou're putting into aerodynamic
testing or the like.
Well, in a digital twinenvironment, what they do is
(25:30):
they take sensors that are onthe car at all times, that feed
back into the digital twinversion so they can experiment
and pilot for a fraction of thecost, exactly what they want to
do and changes they want to makein the car.
So it creates this virtuouscycle of innovation where the
digital car is informing thephysical car, which has sensors
feeding back to the digitalversion of the car.
And that's just one element.
(25:52):
There's a healthcare story toothere, where companies like
Dassault can do it on the humanheart, which allows medical
researchers that maybe mighthave their budgets not be quite
what they used to be.
Right now they can use adigital twin of the heart to
practice surgery over and overagain, because you can't really
test on a cadaver only so manytimes.
And they have found they'veactually pioneered some new
(26:13):
pediatric cardiology surgerytechniques that have saved a few
thousand lives already.
Ryan (26:20):
It strikes me that that is basically kind of like a really
complex video game, yeah, andthe background of a lot of this
is actually in video gamedevelopment, yeah, which is, you
know, I think, is kind.
Of.
I don't know if it's surprising, but it, you know, growing up
we were always told, like, videogames are, you know, a waste of
(26:41):
time or something like that.
But, as it turns out, the samesort of technology that was
really invested in because ofvideo games is what has, it
seems like down the road,resulted in some of the AI that
we're seeing today.
Brian (26:55):
Yeah, some of the stuff that you're seeing, like Epic's
Unreal Engine, some of the stuffthat you're seeing from Unity
Technologies in particular, theyreally start to, I think,
improve the virtualization fieldin a way that we haven't quite
seen.
And what's really cool that youmentioned is Siemens.
The person that they partnerwith on a lot of the
visualizations is Sony, themaker of the PlayStation, so
it's not surprising that gamingexpertise is being leveraged for
(27:17):
these high fidelity versions.
Visualizations you go to gaming, I think of every time.
I always go to like somethinglike iron man and how, when he
first builds his iron man suit,he actually uses a digital twin
in real time of moving out themark pieces of the suit and in
my mind I'm like that was 2008when that movie came out, I
think, and so we're alreadyliving in a future where that
(27:38):
kind of exists to a degree.
You're just putting on a headsetInteresting.
Ryan (27:43):
Well, and then the other digital utility that you kind of
linked in with this was cloudcomputing, and it seems like
none of this would really bepossible without that cloud
ecosystem.
Can you talk a little bit aboutthat?
Brian (27:57):
Yeah, I think in a lot of ways, we've talked so much
about AI rightfully so but weoverlook how important cloud
computing is in terms ofcreating the infrastructure that
enables us in the first place.
When we think about AI whatenables artificial intelligence
we tend to go right to thesilicon and the sensors.
We go, okay, these are the mostadvanced chips and processing
(28:17):
units, but think about chipsalmost as the brain.
Data is the language throughwhich those brains communicate.
Well, who's infrastructuringthat data?
What's doing that?
Well, that's cloud computing.
So that is what your Amazons,your Googles, your Microsofts,
your Oracles are all doing totry and enable this revolution,
and we expect by 2030, based onAI demand alone that it will be
(28:40):
about $2 trillion in spend oncloud or in terms of total
revenue for the space.
So this is a market We've beentalking in billions for this
entire podcast so far.
Now we're mentioning trillionsbecause that's where cloud
computing sits in terms of itsimportance to enabling the
revolution that we're seeingright now.
Ryan (28:57):
Yeah, and all of the massive capital investments that
have taken place.
I mean, there's still plentythat will take place in data
centers and all these hugecomplex warehouses full of GPUs
and server farms and things likethis, and all of that is
(29:20):
operated by these cloudproviders.
Is that correct?
Brian (29:23):
That is correct, and then they're also making deals
directly with some othercompanies that help support GPU
compute as a service.
So something that we've beenwatching really closely is how
does a company like CoreWeave,who has deals with NVIDIA, start
to better partner with some ofthe cloud providers in terms of
dialing up that compute as aservice overall?
And so one of the questionsthat we're financially asking is
(29:44):
is that more of a software as aservice or is it actually
closer to something like aninfrastructure as a service
where it's really pioneering,allowing for the operation of
the cloud as a whole?
Probably right now I'd leanmore to software as a service,
but ask me in a few months andmaybe that shift changes as we
start to really understand howthe layers of cloud are evolving
(30:04):
alongside AI, because that'sthe thing we tend to think about
with cloud, and why I feelsometimes it gets overlooked is,
I think it gets treated as thisstatic innovation, that it's
not as important anymore, butit's vital, it's critical and it
has to evolve alongside theneeds and demands of data center
, as well as power and energydemands.
Ryan (30:22):
Yeah, and glad you brought that up, because I think that's
another important point when wethink about, whether it's
robotics or all these thingstake electricity.
They take a massive amount ofpower.
All the data centers take amassive amount of power.
And I'm just curious on yourthought.
I know you're not an energy guyper se, but how do you think,
10 years from now, how are wegoing to be able to scale up the
(30:46):
power resources in order toprovide enough electricity?
Brian (30:50):
Well, it was funny, I actually did.
My first clients when I firststarted my career were actually
in nuclear energy and naturalgas, and so what I find
interesting is about that wasover a decade ago.
It was a very differentconversation where the terms all
of the above energy was notbeing used as frequently.
It was a term that existed, butit wasn't used in the media as
much.
(31:10):
And so what I find with thedemands is you'll see a probably
135% increase in data centerpower demand from AI alone by
the end of the decade, and soyou've seen a lot of energy come
on board from solar and wind,which are really incredible
renewables, but working from thenuclear side.
Nuclear is always on.
It's higher in terms of energydensity, so you need less
(31:32):
uranium to produce as more interms of terawatts of energy
than solar and wind can do.
But it's also costly and ittakes time.
So there's a few ways to do it.
I know this is something thatyou talked about a lot and it's
and it makes me happy to seethat this conversation is
happening again which is you'reseeing hyperscalers Microsoft,
(31:52):
google and Amazon engaged inthis capacity of saying well,
can we just dial back on nuclearpower plants that maybe have
shut down, which there were alot of closures a decade ago or
so?
There's not a lot more comingonline.
So we're trying to dial back onthe power that we have now and
we're trying to think about well, what does that look like in a
(32:13):
decade from now?
Well, there's a play of is itgoing to be these large power
plants or is it going to besomething like a small modular
reactor, which there's a lot ofpatents for them?
There's a lot of deals outthere, but they're probably not
commercially ready for anotherdecade or so.
And that's just fission.
There is the companies that areworking towards fusion, which
we've had some breakthroughs infrom the Department of Energy a
(32:35):
couple of years ago.
With ignition, that workforcehas increased by 200% in the
private sector, so clearly thatthere's more interest than ever
that maybe fusion's closer.
I think fusion requires AI andquantum to get dialed up a bit
more for it to become viable andit's probably something that's
more of a 2040s technology.
So what do you do in theinterim?
Well, it's probably an all ofthe above energy strategy of
(32:57):
renewables, nuclear and,probably to a degree, gas in the
mix.
Ryan (33:00):
Yeah, it does seem like it's not something that we can
wait until 2040 for the fusiontechnology that's always just on
the horizon and we never quiteget to and hopefully some of the
breakthroughs that people havetalked about become realized and
they can actually.
You know some of them.
There's some pretty massivespending that's taking place to
(33:22):
advance fusion technology around, you know, in Massachusetts and
some other parts of the country, so that's encouraging.
But you know to your point.
I do think there has to be somespending over the next five
years to actually increase theelectricity capacity in the US
to compete globally, especiallywith China.
Brian (33:43):
Yeah, I think that's the biggest thing to focus on is
really how do we build up thiscapacity and I think some of it
too is.
And what I find fascinating inthis story was mentioning the
hyperscalers making these dealsthemselves.
We saw a shift with AI ingeneral, with the headwinds that
were in the merger andacquisition market where they
couldn't really acquire asfrequently in the last few years
(34:07):
Companies across the board.
They got into the investormindset where Microsoft invested
heavily into open AI.
You saw Amazon heavily investin AI companies.
Well, now they're putting on anew hat, which is we're power
purchasers, which I think isvery fascinating.
Which how does that translateinto it?
It might be a competitionthat's either driven by
(34:27):
geopolitical competition or itmight be just this is the table
stakes and required for thefuture to bet on it.
So it'll be a complex landscapefor sure.
Ryan (34:36):
It is complex and it adds a level of complexity and
difference to the traditionalasset-light business model of
many technology companies.
If they're having to makecapital investments to a degree
that they have never made before, that transforms the sort of
business that they're operating,and so time will tell to see
how that plays out financiallyand seeing what sort of return
(34:59):
on investment these companieshave with those capital
investments.
But I do think that they'rekind of in a position where they
don't have a choice.
To some degree they have tomake those investments if they
want to be able to advance thetechnology and compete and have,
if they want to be the winnerin 10 years from now.
Yeah, okay.
(35:20):
Well, brian, this has been agreat conversation.
Appreciate you coming on thepodcast again.
I do have a final question foryou.
I want you to put on your youknow futurist cap, okay, okay,
so we've talked about a lot ofdifferent technologies.
What is you kind of thinkforward, looking into the future
(35:41):
, a decade?
Is there any specific thingthat we haven't talked about
that you think is going to be?
You know that maybe fewerpeople are focused on right now
that, looking back a decade fromnow, we're going to say, wow,
that was an amazing advance andamazing innovation and nobody
really saw it coming.
Brian (35:59):
I think it's one of the things that we talked about.
I'm happy we already talkedabout robotics, quantum and
nuclear, because those are thetable stakes, ones that are
quite popular.
But I did mention one verybriefly in the visualization,
one which is, I think, smartglasses are oftentimes
overlooked, where there's veryfew ideas, where you already had
the start and stumble about adecade ago with smart glasses,
(36:20):
of where they're going.
But with AI capabilities onboard, translation capabilities,
augmented reality technologygetting to a point where it's
able to better be customized toa comfortable form factor that
we're seeing, I think we mightbe in a decade seeing a point
where there's a greaterproliferation of smart glasses
on the market than we expected,because I think it's the only
(36:42):
device, by a form factorcomparison, that might be able
to compete with a smartphone.
I don't think it replaces thesmartphone outright in a decade,
but I think it's the onlydevice that could possibly
compete with it.
In a way.
It will come down to certainlylooking at how many more models
come on board.
We're already seeing Oakley isnow going to come out with their
version to compete with Meta'sRay-Ban, or I should say
(37:05):
complement Meta Ray-Bans,because Essilor Lungsotica owns
all those brands, but I thinkthat's one that gets a little
bit overlooked, and I'm excitedto see a lot more in terms of
that innovation in a decade.
Ryan (37:15):
So are those joint ventures typically that take
place where it's, you know, thesunglasses company and the
technology company are kind ofworking on a product together.
Or do you see, you know, Appleor Meta or one of these
companies trying to kind of puttheir own glasses out there?
Brian (37:31):
I think right now they're leaning into the partnership
approach, because part of whenyou get into something that has
a consumer implication like thatis you want it to look cool and
I think, like I've seen metaRay-Bans in the wild and they
look like Ray-Bans, they looklike a good sunglasses model.
So I think that there's a lotof strength in that partnership
approach.
So, as much as it's going to bea technology that maybe
(37:55):
redefines translation,communication and computing in
different ways, it might also bea fashion statement which is
pretty powerful in a lot of ways.
So there's a cultural elementto it which is important,
because technology doesn't existin a vacuum.
It's adopted by people at theend of the day.
Ryan (38:11):
That is a really interesting insight and we'll
look forward to seeing yourtranslation example.
I think is really interesting,as somebody who travels
internationally quite a bit,having the ability to have
real-time translation comingfrom your glasses into your ear,
instead of having to hold up aphone or something like that and
say, you know, speak into thisso I can get the translation.
(38:31):
Yeah, that seems like a reallyinteresting application.
Brian (38:35):
Yeah, I think it would be useful.
As someone who also travelsinternationally quite a bit I
that is probably like the mostin demand.
As much as I love learning newlanguages, it's really hard, and
so having something that can atleast assist that and put that
at ease is incredible.
Ryan (38:49):
All right.
Well, brian Komsky, thanks forjoining us again on the First
Trust ROI podcast, and thanks toyou as well for joining us.
We will see you next time.
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!