Skip to playerSkip to main contentSkip to footer
  • 5/8/2025
Watch former iLearningEngines CFO Farhan Naqvi discuss private capital’s impact on tech innovation at the Bloomberg Invest 2024 Conference in New York. He joins Barclays’ Kristin DeClark to break down today’s investment trends and how smart capital drives growth.

Moderator: Bloomberg’s Mandeep Singh
Sponsor: Mubadala
Topic: Private Equity • Enterprise Tech • Growth Strategy

#SayyedFarhanNaqvi #SayyedFarhanNaqviiLearningEngines
#SayyediLearningEngines

Category

🤖
Tech
Transcript
00:00Good morning. Thank you, Elizabeth. I appreciate that. It's an honor to be here with you. So many
00:04great faces here this morning. My name is Amem Aboubakar, and I'm the Deputy Chief
00:08Communication Officer at Mubadila. For those of you who don't know Mubadila and are familiar with
00:13the organization, we are based in Abu Dhabi, UAE. We're a sovereign wealth fund organization.
00:19We have an AUM of around $300 billion, and most of that, 40% of that is in here within America.
00:26I'm going to look at my notes so I don't forget anything important I want to make sure that I
00:29share with you today. We are here at an exciting time. I was making a joke earlier today with
00:35regards to AI. I remember when we started talking about AI and technology, I was like, what is this
00:39thing that we're talking about? And now I can't seem to have a single conversation without talking
00:43about AI or technology. So clearly we've moved in a big direction. This is an exciting time for all
00:50of us. We're here together today for a particular reason, because it's transforming and it's moving
00:54in a direction that all of us are trying to get our heads wrapped around it. At Mubadila,
00:58we're a big believer when it comes to collaboration. So collaboration for the cause of coming together
01:03to figure out how we can progress together, make transformational change for a positive
01:08outcome. And that's why we're in this room today. Today we're seeing, as many of you know,
01:13so I don't need to state the obvious, we're seeing a booming when it comes to the tech sector,
01:18fueled by AI, and which is already unlocking a lot of human capabilities in ways that many of us felt
01:24unimaginable. Tech is transforming operating models, it's transforming and increasing efficiencies,
01:31boosting innovation, and helping address challenges to humankind. And that's what many of us get driven
01:36by. We are seeing a huge leap when it comes to things like fifth generation technology and cloud
01:41computing, and associated industries, such as digital infrastructure. These present a significant
01:48opportunity for many of us here in the room today. How do we come together to make sure that we're
01:52investing correctly, and how do we come together to make sure that we're all creating value, we're all
01:56having an opportunity to get value creation from it. So coming together today, but just because
02:03we're investing in tech doesn't mean it's one straight line. It's a complex situation. Many of us
02:09are trying to figure out how to invest, where to invest, what are the risks and rewards that come
02:14with that. It's forums like this where we are here today, where we're hoping to unlock some of those
02:19opportunities, unlock some of that thinking, and share learnings. We, as a long-term investor, as a sovereign
02:25investor, have the opportunity to not just be investors, but also partners to many of you. We're investing in
02:32tech for decades, from our early investments in organizations like Global Foundries, which might be
02:36familiar to some of you here in the room today, all the way up to partner, all the way up in terms of creating
02:41organizations like G42, M42, Space42, and our U.S. organization of Powerhouse, which is our MGX organization.
02:52Across the board, what we're seeing today is that we're investing in tech is playing a huge role,
02:58all the way up from early startups, all the way up to entrepreneurs, and where do we fit into that
03:02process? Private capital has a huge role to play across all of these areas, including the Global South,
03:09as sovereign wealth organizations and financial players. We have a unique capability and ability
03:16to not just participate in this progress, but also to tackle some of the biggest challenges that
03:21we're seeing to humankind. Whether we're a sovereign investor with multi-generations views or a private
03:28equity or angel investors, we all play a part in this role. So with that, I'm going to get out of your
03:34way, allow you to hear from this amazing esteemed panel, who will unlock some of these conversations
03:38we've already shared with you. Thank you so much.
03:40Thank you, Maima. A quick reminder that this session is on the record, so you are very welcome
03:59to take notes, share insights, take pictures. And if you are avid on social media, please use our hashtag
04:06Bloomberg Invest and tag us on Bloomberg Live. So without further ado, I would like to invite to the stage
04:13our panelists, Kristen Roth-de-Clark, Global Head of Technology Investment Banking at Barclays,
04:20Furhan Nakvi, Chief Financial Officer and Head of Corporate Development for iLearning Engines,
04:27and our moderator, Mandeep Singh, Global Head of Technology Research of Bloomberg Intelligence.
04:32Great. Good morning, everyone. Very exciting to be here. So I lead the tech research at Bloomberg
04:43Intelligence, and I'm going to share a shameless plug right now that we did a $1.3 trillion forecast
04:52for Generative AI by 2032. And this is a report that's available for Bloomberg subscribers. It goes
05:01into the depth of, you know, the different parts of the Gen AI stack, and I would be happy to talk to
05:08anyone about it. But regarding the panel today, I think it's exciting to talk about the role of private
05:17capital in this wave of technological innovation driven by Generative AI, because at the end of the
05:25day, you know, that's what's going to keep this going in terms of, you know, kind of widespread adoption.
05:32So I'm going to start off with Christine. Christine, you lead the IPO market, the tech IPO market at
05:41Barclays. And given what we have seen from, you know, the blowups in 2021 when it comes to the tech
05:48IPOs, especially on the software side, where are we now in terms of the broad IPO market when it comes
05:56to tech companies? Well, so there are a couple of things that happened in the last two years,
06:02primarily. We had a major valuation reset as it relates to technology companies in the public market.
06:09And we had a major shift from a focus on growth at all costs, in some cases, to profitability.
06:19Because of that, I think it, you know, the companies that were sitting in the backlog that were looking
06:24to go public have had to take a bit of a pause and figure out business model fit, make sure that they can,
06:33you know, prove unit economics, prove a path to profitability, get profitable.
06:36And also kind of get a more sober view on what valuation looks like, especially for the
06:45venture-backed businesses that may have raised at, you know, a significant premium to where
06:50they could ultimately get public in today's, in a kind of more normal valuation environment,
06:55if you look over history. So with that, I think we've seen very few technology IPOs over the last
07:02couple of years. Part of that is companies getting ready, getting to scale, getting to a
07:07financial profile that the market would accept, and also getting comfortable with the new valuation
07:12paradigm. I think we'll start to see that improve because we've now been, you know, two years or so
07:19without a number of IPOs. And so a lot of these companies have gotten to significant scale
07:25and will be poised to hit the public market. Today is, you know, we're in a pocket of uncertainty
07:32right now. We're leading up to the election. We have, you know, some time to see kind of when
07:38rates will start to come back down, how many cuts we have this year, if any, and then what that looks
07:43like next year as well. So I think that most companies that are in the backlog have opted to wait
07:49for a little bit more certainty in the macro before they tap the market, given that they've waited
07:54a couple of years. That being said, there have been a few IPOs that have gotten out and they have
07:58strong, you know, strong conviction from the shareholders that are coming into those IPOs
08:04because they're not buying it for a quick pop. They're buying it for, you know, to own the company
08:11long term because the performance of IPOs this year, if you average all the IPOs that have come public
08:16this year across all sectors, up 6%. But it's not exactly like running away. There's no FOMO
08:21chasing IPOs. And so it's really fundamental buyers investing. But I think that we'll start
08:30to see a little bit more momentum next year. Yeah. And we've seen a couple of good ones,
08:37Arm, Rubrik. I'm sure there are others that are there. Farhan, you have an investment background
08:44and you've also done M&A deals. From your lens, you know, given we are talking about AI and tech
08:52being synonymous right now, what is it that makes for a good business model when it comes to these
08:59new crop of AI companies? Well, so I view this as being split into different dimensions. One is
09:10infrastructure level investments. You'll see L&M companies which are building, actually building the
09:16models, garnering a lot of investments. And then there's the other part which is actually building
09:23applications for business use cases. That part is still not as robust. So most of the money that has
09:30been going into the sector has been going into the infrastructure level investment part. As the market
09:36evolves, as the technology evolves, we'll see that shift happening where businesses which are actually
09:42using AI tech to further their business use cases will start getting investments, valuations that they
09:51deserve. And so what valuations do these companies deserve, Christine, given, you know, there isn't a clear
10:00business model. Given, you know, every day you talk about models getting bigger and, you know, there's a lot of
10:07investment a company has to make to train these models. I mean, are they at a point where you feel, okay, when
10:15they go to the public market, investors would be like there is a path to profitability, which wasn't the case
10:21back in 2021. So you think what would help these companies get there? And just what's what's your
10:31perspective in terms of the evolution? Yeah, well, there's a reason why this room is so full. And
10:36Umaima said something in her opening remarks. She said, we're trying to get our hands around it,
10:41right? And I think we're all trying to get our hands around which companies are investable, how to think
10:47about, you know, who's going to win. And so most of the investment is happening at the venture level,
10:53where you can take a portfolio approach, you can buy eight, nine different companies. And, and then
10:59you're not necessarily picking a winner, but you're hoping that one of them is the winner.
11:03And in the other investment that's happening in a more meaningful way, it's because it's kind of like
11:09winner agnostic is on the chip side, right? So we've seen a number of these chips companies really
11:13take off with the explosion of, of, of AI and these technologies. And so I think to get to the
11:21point where you're not just taking a portfolio approach and it's the early stage investments
11:27happening, we need to see those exits start happening. And we haven't seen that yet. So we
11:32haven't seen really a strong strategic bid for these venture backed AI software businesses,
11:38because I think it's too early to tell who the winners are going to be. And so we need to see
11:46that play out a little longer. And once we start seeing a strategic bid for some of these businesses,
11:51it'll give, I think investors a little bit more confidence in terms of what areas and who the
11:56winners are. And just to hone in on that. So is there a difference in positioning for venture
12:01backed versus the private equity guys here? Well, yeah. So the venture backed businesses are very
12:06early, right? And so, um, it's concepts and then it's proving out market fit. It's proving out,
12:12you know, customer, uh, adoption. I think in your case, right, a big part of the business is
12:17showing when the technology, when, when you can't totally get your arms around the technology
12:22differentiation, it's who are your customers? What's the net retention look like? And proving out the,
12:27um, kind of end market more than it is what, what are the specifics of the technology? And so that's
12:33the part that a lot of the venture investments that are happening are kind of pre, um, proof
12:37points like that. Private equity, um, typically are looking for businesses that have a proven path,
12:44you know, proven profitability. It's something that they're just taking, you know, investing to,
12:49to scale or to improve the, um, financial, the, the profitability of the business. And so that's,
12:55that will happen at a much later stage. Or they're buying some of these technologies to put
12:59into their existing portfolio companies, but that's more of a strategic buy. And we haven't
13:04seen that yet. Right. And, and to dovetail, uh, to that, uh, for Han, uh, I mean, the market has
13:10been very good at sniffing out, you know, where the next wave of, uh, AI is going from chips to power
13:18shortages to data center demand and upgrades from your lens. Do you care about these on a day to day
13:25basis, uh, at your company or when you are looking, uh, talking to investors? Well, so, uh,
13:34don't, don't take this as a humble brag, but we are one of the few companies which have been able to
13:38actually build out business use cases. So, uh, moving away from the infrastructure part to actually
13:46building use cases, which clients of ours can, can deploy. So take, to take a step back, we are an
13:53enterprise AI platform for learning automation and workflow automation. We enable our clients to
14:00build and deploy AI apps at scale. So our clients in oil and gas manufacturing, they're using our
14:07platform to build and deploy AI apps for their own use. Right. They're driving efficiencies right now,
14:15and they're moving to the phase where they'll start, uh, generating revenue pools or products which can,
14:22which can contribute to their revenue top lines, uh, moving away from just efficiency focus that is
14:28being used by AI apps right now. Yeah. I mean, uh, so, uh, I, I guess when it comes to how companies are
14:37shifting, like it's not going to happen all of a sudden that, you know, you upgrade everything or you, uh,
14:45pivot to this new way of doing things, uh, like from a market timing perspective, how do you think,
14:51Christine, companies are going to say, okay, this is the right time for us to go public because there
14:57is this nice tailwind when it comes to enterprise spending or just, you know, how, uh, like the
15:05investors are thinking about these companies. Like, is there a sentiment that you track or any
15:10thing that helps you determine from a market timing perspective? Well, I mean, one thing we've
15:16learned from 2020, 2021 IPO classes is how important scale is, um, because the companies that were smaller
15:24scale, less liquidity in the public market, um, were impacted much more meaningfully than those that had
15:30liquidity. And that's just because it was kind of like the top 10 investors, one or two start to sell,
15:35and you see the stock, you know, impacted meaningfully, and then everyone runs for the
15:40hills. And then you have a number of orphan stocks, even if the companies are performing,
15:43it's hard to get, it's hard to have the confidence as an investor to get back in.
15:47So scale matters a lot. Um, and I think we're going to, it's going to be some time before we see
15:53really scaled assets and AI. I mean, a lot of these companies are growing very quickly,
15:58so maybe it's not as long as we've seen historically in other trends, but I think it will take some time for
16:03these businesses to get to scale and have the proof points around unit economics. I mean,
16:08you've got scale growth, profitability, all the things that really do matter. Um, but that's what,
16:14that's what the market needs to see and scale. We're talking a revenue scale of 250 to 300 million,
16:22at least in the year that you're selling off of, which is the forward year is really, I think the,
16:26um, the benchmark right now for investors to participate, at least on the software side,
16:32would you agree? Absolutely. Couldn't agree more with that. So it took us about 400 million dollars
16:38to be able to start thinking about putting the IP process in motion. Yeah. Yeah. Anything you can
16:43share around the state of tech unicorns? Um, well, the interesting thing is a lot of the companies that
16:51are private in our backlog, kind of street wide to go public, a lot of them are software companies. And
16:57we've seen in the public market, um, a number of companies have challenges with, you know, um,
17:04elongated sales cycles. We've seen vendor consolidation. And so I think that if you look
17:09at, you know, the, the companies in the backlog that are unicorns, um, a lot of the software companies
17:15are going to wait a little bit to have investors have more confidence in the enterprise spending
17:20towards software in general, or be able to show metrics that they've been able to withstand either price
17:25increases. They have net retention numbers that are significantly better than some of the companies
17:29that have been challenged in the public market. Um, as it relates to unicorn valuations, I think that,
17:37you know, that was something that was such a focus because there was a competition for
17:41engineering and sales talent, especially in the Valley. Um, I think that has,
17:47uh, there's a lot of, a lot more sophistication around what unicorn status really means from a lot
17:53of the employees. And that is, um, you know, cause some of these companies would have unicorn status,
17:58but really have ratchets or some sort of synthetic instruments that got them to the unicorn status,
18:02but didn't necessarily persist. Um, if they were to go public, you know, uh, when they were to kind
18:08of get valued for an IPO. So that's the hype around unicorn status, I think has come down pretty
18:13significantly in the last couple of years. And, and Farhan, for a company with 400 million in revenue,
18:19like, how do you think about tech spending? I mean, in general, I mean, you're the CFO. So when you talk to,
18:26uh, your, uh, you know, tech lead or your CIO, uh, how is it that you're thinking about, uh, investing
18:35in your business to upgrade and, and just overall the tech stack that you currently have?
18:40So one, one thing that we've realized, uh, very pertinently over time is data is the lifeblood of
18:51any AI. Uh, data is what breaks or makes models. Uh, and one thing that I think we did well was
19:01identify this early on and we put in motion and process where we are actually procuring data.
19:05data. So a lot of our spend is focused on data purchases to, to make sure that the, the models
19:12that we built ourselves are actually good, good, uh, to be deployed at our client's end, right?
19:20Other than that, uh, your, your spend, uh, on, on infrastructure, uh, on, uh, on compute power,
19:32that, that follows the trends that you'll see across the market. Like there are not going to be
19:38a lot of variations between players on how much are they able to squeeze or increase their spend on
19:44that. So your infrastructure guys and coming into you and saying, oh, there's supply shortages that
19:50you got to pay up for the GPUs. They always are. I don't get surprised by that anymore.
19:56Uh, but, uh, yeah, so that's a constant. Uh, and that is something that we always have to factor in
20:02when we think about where do we actually make the investments. Uh, but what I'm saying is that is
20:08consistent across any AI company. Yeah. Is, is there a feeling right now that you are going to be left
20:14behind in AI if you don't allocate more capital or no? I mean, yours, I really hope my clients have that
20:21feeling if they don't have it yet, but, uh, so there's definitely excitement. Uh, people want to
20:29try out the technology. Uh, they want to understand what this has to offer and then taking baby steps in
20:35terms of, uh, adoption. So they'll make few investments, try it out. And if that works for their business,
20:42they'd expand the use case. Yeah. That's how we've seen this spanning out. Yeah. And, and Christine,
20:47when you're looking to do TAM analysis for the companies you're looking to take public,
20:54how are you even thinking about a TAM for something like Jenny? Are you throwing trillion
20:58dollar numbers or you're more realistic here? Yeah. Over time. I mean, this is a massive decades
21:04long, um, shift in innovation and, and huge tailwinds. I think it's hard to put a number on it. Um,
21:12I would say just anecdotally kind of adding to what Farhan just said, anecdotally,
21:16and talking to some of the large cap clients that I cover, we cover, um, there are two ways,
21:22the kind of two ways they're approaching the efficiency piece of AI. One is top down. What
21:27are the, what are the point solutions that we can plug into our existing processes to improve
21:31efficiency? That's, that's an easy sell, right? Because call centers, whatever it might be,
21:36anywhere you can add automation and improve margins very quickly. That's an easy sale. I think what you
21:42said at the beginning, Farhan, when you mentioned, um, it's more than just the efficiency. It is
21:48actually the innovation piece of it and accelerating innovation. That's the piece that, um, I think
21:55investors and companies are trying to figure out because they don't want to be left behind because
22:00they don't, they're worried now. Even these large incumbents are worried about new entrants scaling so
22:06rapidly because of the, um, because of the technology advancements. And so that to me is where the,
22:12you know, the board meetings are. What do we need to be worried about? Um, because the easy sale is on
22:18efficiencies. Yeah. And, uh, are you seeing efficiency, uh, Farhan, in, in the OPEX savings that, you know,
22:25uh, folks are able to... In our organization? Yes. But no, she's, she's absolutely right.
22:30Uh, the focus right now is on capturing efficiencies, uh, across the cost buckets. But once a client has,
22:40has done that for a bit, they've very quickly realized that they can actually use and leverage
22:45the same tech for enhancing their, their product portfolio. They can, they can build newer products,
22:52uh, which adds directly to their top line. So the, the coverage area where they can deploy this
23:00technology expands from just being focused on capturing efficiencies to enhancing revenue,
23:06you know, enhancing top lines. Yeah. I mean, right now we are at a stage where we have generalized
23:11large language models that can, you know, do everything, but everyone is expecting there
23:17will be more specialization when it comes to LLM. So how do you think it will be disruptive to your
23:23area of, uh, operations and some learning? I mean, we, I think education is where it all started,
23:30you know? Yes. And, and so, uh, maybe, uh, you know, spend a couple of minutes on that.
23:35Uh-huh. So we, uh, we, as a company, uh, we do not use LLMs. We have built SLMs or small language models
23:42which are specific to certain sectors and that is what we deploy for our clients. So, uh, LLMs,
23:49like, and that is what drives the actual, uh, usage for those clients. If you start with an LLM,
23:58uh, a generic LLM and try to deploy that for a client in the oil and gas space and then also the, uh,
24:05same LLM for a client in the medical or insulin space, it's, it's a lot of work, uh, and the results
24:11are not as effective. Right? So the approach which we've realized over time works is you create smaller
24:19language models, uh, which are vertical specific and in some cases client specific and deploy them.
24:24That's what drives the usage. That's, that's what drives the ROI for the clients in the end.
24:29Yeah. Would you agree with that? Yeah, I would say in the quality of the data that they have is
24:33the really the, and back to your point on data matters and the quality of the data that they're
24:38putting into these models is really what is going to drive the, the best results.
24:44Yeah. And, uh, so we do a CIO survey every six months at Bloomberg intelligence. And one of the
24:51things that has come out for the last two times you've done this survey is the dependency on
24:56hyperscalers, the cloud service providers, you know, how much sway they have when it comes to the
25:02compute, the development of large language models. How are you advising, you know, young companies that
25:09you're working with, uh, to take public in terms of, you know, their positioning in, in these markets
25:15where the hyperscalers have so much dominance? Well, I think, I mean, the reliance on hyperscalers
25:21is massive and that's no question. I think the best advice for young companies is to have
25:25diversification of partners. And I'm sure you guys are seeing that too. Like it is, it's, it's, it's like
25:31any, I mean, it's like any boom in technology. It's, you know, we saw this with chips back in the day
25:37when phones were first coming out. Like any reliance on one particular vendor is going to put you at a
25:42disadvantage and over time, especially from an investment standpoint. So having a diversification
25:46of vendors is the, is the safest way to kind of scale. Yeah. A lot of the value in this whole
25:56chain has been going to the hyperscaler so far. Yeah. And, and, uh, do you think from your perspective,
26:02uh, the reliance is not going to go away when it comes to the IP that you want to maintain versus
26:09what you do on cloud? It will take time. So, uh, we are totally dependent on our clients'
26:15choices as to which cloud environment do they want to deploy on. Right. Uh, and that, that dynamic,
26:23if it shifts, it will take a long time. Right. And, and so, Christine, uh, we have three minutes
26:29left, uh, perhaps, uh, you know, you can sum it all up in terms of your expectations, given you
26:36mentioned this is an election year, so, you know, things may take a while to play out, uh,
26:41maybe have a 12 month outlook or, uh, something along those lines. Look, I think, um, I'm, I'm bullish.
26:48Um, I'm, I'm optimistic probably by nature, but I'm bullish on, uh, kind of a post-election. It's hard
26:54because we have a smaller window, but if we start to see some rate cuts and investors continue to get
27:00more confidence that we're moving in that direction, it makes the funding environment easier
27:05for all of our clients. We're seeing a shift in the way that public companies are financing
27:11themselves. One, one interesting shift recently is, um, refinancing straight debt with convertible,
27:18uh, with convertible paper because they're able to have pretty significant interest expense savings.
27:23And so there are different financial markets that public companies are using or financial
27:27instruments that public companies are using right now, um, that are working, but will continue to
27:32get more, um, you know, more attractive as rates start to come down and we have certainty around the
27:39political environment. Um, and then we start to see, you know, the, the IPOs start, you know, come out.
27:46Once we have confidence in the secondary market, obviously it makes new issues easier. And then you
27:51have more confidence in the private market, right? So it's easier, it's easier to put money to work
27:55in the private market when you see exits in the public market or strategic activity. And that is from
28:03in AI and software and chips and kind of every single subsector within technology.
28:07Uh, so we are not going back to the zero interest rate era, I think, but it'll take time for these
28:19things to stabilize. Uh, as Kristen mentioned, election year plays a huge role into how, how these
28:26things went out, how investors look at it. So from my perspective, it's just wait and watch now.
28:32And, and from the enterprise side of the equation, given you're running a company enterprise side is
28:38not as dependent on this, especially, uh, when it comes to technology spending, uh, and for, for
28:45products like this or for technologies like these, the sale cycle and the spend cycle is fairly well set.
28:52So it doesn't change dramatically, uh, in the election year, you would see some tailwinds, but not, not a lot.
29:00So it won't have any impact on your network. I wouldn't say, I hope not.
29:05I wouldn't say, uh, any impact, but not, not a significant impact.

Recommended