Episode 2: Thinking Through Uncertainty: Technology, Global Dynamics, and the Private Capital Outlook

Kanwar Singh: Today? Sure. And thank you for the intro and the handoff. Let's just maybe level set a little bit with regard to what we're talking about when we talk about AI. Right now, we're talking about generative AI. It's basically these large language models that are processing information and learning off of whatever data we feed them. The future may involve something called artificial general intelligence, which is a self-learning tool that creates all the dystopic movies that we've been exposed to. I think the benefit that we have right now is that we're working with a tool that actually is providing leverage in the system for a lot of us. And the question is, how do we make sure that leverage is a beneficial thing is accurate? One of the things that I've seen that has probably tripped up a few people is that they tried AI when all the noise first came out maybe at the middle to end of last year and it did have a lot of errors and it was a little wonky and it gave you kind of not very human sounding answers.

If you go back and try it now and use it now, what you're seeing is a marked difference. And I would say each month that I use it and I would put myself in a small percentage of what we might call power users, but you're seeing each month, I think almost a logarithmic improvement in the quality and the capabilities of what the AI is doing. And part of that is because again, AI is only as good as the information that can train itself on. You have to constantly be feeding it. And what AI is beginning to do now is actually use its own tools to create information off of which it will train itself, which is really interesting.

When I think about our world in the family office world and the family office advisory world, as well as the sort of investment world, a couple of things to note that I think is relevant. Certainly AI is the dominant narrative. We've seen that just across the four hyperscalers, the big ones we call them MAMA, Microsoft, Amazon, Meta and Alphabet, Google. You're seeing $710 billion of AI capital expenditures just in 2026 alone. That's astonishing. So all of this money's being poured into continuing to grow and increase the capacity that exists with regard to AI. That is fueling a lot of what we've seen as the markets rise this year off of the volatility we saw at the beginning of the year with Operation Epic Fury and the Iran conflict. So you're seeing not only the hyperscalers increasing corporate AI investment across all entities, including ours and all of them has more than doubled.

We're seeing a significant increase in terms of that role. Now what's interesting is on the family office side, 70% of family offices have no investment in AI infrastructure, despite the fact that over 65% say it's a priority. It's probably the key driver of where returns will go. And so I think there's a gap there in terms of how that's being utilized, especially as you think about the way that in the context of what we've talked about, private markets, capital markets, M&A deals, a lot of the private equity firms and institutions are heavily getting involved in leveraging AI for the sourcing of transactions, the analyzing of transactions. But in this situation, despite what we talked about last week where family offices are trying to take more of a greater role, I think they're a little behind the curve in terms of not being as facile with AI.

I'll hand it back to you, Dean.

Dean Peterson: Yeah. Thanks very much, Kanwar. John, I know you're AI user. Tell me how AI is changing the competitive landscape between PE firms and strategic buyers.

Jon Zefi: So what I think you're just generally seeing is that PE firms are broadly supplementing their associate base with the use of large language models to suss out potential opportunities relatively rapidly in verticals that they're interested in deploying capital in. And what you're seeing across various PE firms is the drive towards further leveraging their teams with the use of large language models to run EBITDA calculations, revised projections, competitive landscape determinations, peer reviewed determinations to figure out how to best deploy capital. It's an intrinsic part of every PE firm's model now and it's becoming a obvious tool for increased leverage across PE firms broadly.

Dean Peterson: Sure. Ted, let's ask you, are you seeing AI powered deal sourcing change? Is it changing the flow of opportunities that come across your desk?

Ted Rosen: Yeah, 100%. So look, traditionally investment bankers source deals by cold calling, by attending conferences, by developing relationships. It was a much more relationship based process. What AI is doing is it's completely changing the way investment banking firms are going to be competing and the AI driven apps are compiling information at such a rapid speed that you can now compare and locate sellers in industries and revenue segments and industry segments and financial segments and changes in management, age of the sellers. So from an investment banking perspective, it's a deal changer. From the private equity buyer perspective, they're now going to have the same access to that information that they, in some respects, relied on investment bankers to bring them more of their deals. And I've been seeing this for a long time, but I get a lot of my sellers will come to me and say, "We've been approached directly by a private equity fund." That I think will continue to become more and more common.

So there really is a rapid change in not only how deals are sourced, but in the kinds of deals that are being sourced, because there's a lot more niche businesses that are being acquired like HVAC and MedSpas and specialty manufacturing. So AI is a great tool to locate these fragmented companies and fragmented industries.

Dean Peterson: Yeah. Thanks, Ted. So what we're seeing is we're seeing an expansion of the market of potential buyers, but there's also a lot of uncertainty around AI. John, the panel today is framed around uncertainty, AI, geopolitics, energy, et cetera. From tax respective, what are the capital-

Jon Zefi: Sadin, thanks for that comment. What we're seeing them do right is that they're updating their asset allocations across the board. Obviously AI infrastructure, real assets and energy, those are the sector calls that are largely defensible and have been the focus of many private equity firms as well as family offices that are operating in the space. What they're seeing that may be a problematic area is that they're treating tax potentially as a downstream consideration in all of this. And remember, we're running those models. The single biggest determinant of after tax return on a 10-year private capital portfolio is structural, how the entity is organized, where we source the income, what credits and incentives do we get at the federal, state and local level and how exits are sequenced and what portions of the code can we utilize as a sword as opposed to a shield to maximize the after tax return to our clients.

The shift that's happening now, 2026 is the year that tax architecture becomes a top three investment input alongside the thesis, the underlying investment thesis, and the execution on that investment thesis. Family offices that have made the shift are pulling away from those that haven't and they're seeing an express increase in internal rate of return on investments. A lot of the risks that you'll see in the marketplace, particularly from a tax perspective that we're going to address from a global PE perspective stem from the one big beautiful Bill's international changes and you and I will probably be discussing that a little bit later. We're going to look at tariff durability and pillar two as its impact on cross-border transactions. Anyone who's put together models that are designed on pre- 2024 may now be tax inefficient and have to update those models. So I guess in some, in 2026, tax architecture isn't operational, it's strategy.

The families and institutions that get this right are compounding faster on an after tax basis, which is what everyone wants. They want a 3X multiplier at a minimum on transactional value.

Dean Peterson: Sure. Thanks very much, John. Yeah. I mean, the tax factors are always critical whether you're thinking about your capital structure or going down the road. Now, John, you mentioned Pillar Two. I'd like to bring up a comment about Pillar Two. So Pillar Two, I know this is a fairly sophisticated audience, so you've heard of Pillar Two likely. It's an OECD generated 15% global minimum tax. Now the US has not signed onto Pillar Two and under this administration or maybe future administrations will not. So that's throwing the global tax picture into a bit of flux. In January of this year, there was a side by side agreement adopted by the OECD and the G20 nations to accommodate US multinationals that are affected by both the US tax system as well as pillar two. And they're trying to make an accommodation for those US multinationals so they will not be double taxed essentially by both of those regimes.

So this was seen as a step in a positive direction. We'll see how that actually plays out. That affects many clients operating globally. Pillar two is a story that's going to continue to be told in 26 and 27. So let me go back to talking about AI. So AI infrastructure is the dominant capital story of 2026. Kanwar, where is tax driving for your clients the actual investment case?

Kanwar Singh: It's interesting because there are a number of people who are playing various angles regarding AI and you've certainly seen a lot of headlines about the impact that the AI story has had on investment returns. As of the most recent data that I've got, fully 45% of the S&P 500 is effectively now an AI index and that's a marked change. It wasn't that long ago that you really had in the S&P 500 a true core portfolio of growth and value type stocks. And what's interesting is when you sort of disaggregate it now, if you look at the S&P value index, that's really what the S&P used to look like. I bring that up because a lot of people are chasing a lot of the same deals. And so the hyperscalers where a lot of the activity is, again, the metas, the Microsofts, the Amazons, the Googles, they're getting a lot of attention for our clients and for a lot of folks who are in the space looking at the infrastructure side of the equation and not exclusively just data centers because if Moore's Law continues to do what it does, which is to increase the efficiency and reduce the size and energy demands of processing power, semiconductor chips or new technologies, eventually there might be an overbuild, but it is clear that we are facing in this country a power shortage to supply all of the power needs and the other infrastructure that has been deferred for a long time.

There are investments that can generate with contracted revenues with AA minus, double A rated off takers, long streams of revenues that can generate as much as the 10% rate of return, which because at least for the foreseeable future of the next several decades, it's going to be return of capital. You're looking at that as being an after tax return and people forget that the long term average return of the stock market is seven to 10% pre-tax. So a 10% after tax return could be exciting. I'm not suggesting that's a promise or a guarantee, but you're hearing this kind of theme come out more and more. So the question is, especially when you look at the fact that so many dollars have chased sort of the parabolic growth that you've seen in some of these high sexy stocks, there is an increasing desire to look for other avenues that can generate returns without reliance on exclusively just the sort of the meme stocks or the stocks that have gone parabolic.

Dean Peterson: Sure.

Kanwar Singh: Infrastructure.

Dean Peterson: Thanks for that, I want to go back to something that you mentioned within that answer. So I know this is a bit of a boring topic, but energy. So energy is having a moment right now, whether it's driven by AI, geopolitics, and the realization that the grid is not built for what's coming. So can I get your thoughts? Ted, let's start with you on how you see energy being an investment in 2026 or 2027.

Ted Rosen: So from a legal perspective, we have an enormous data center practice and it touches real estate, it touches corporate, it touches IP and energy obviously is the power that drives the AI boom. So the question becomes in AI, sorry, in energy acquisitions and in acquisitions of anything that's related to AI infrastructure. We're seeing very high multiples, but in diligence we're concerned about customer concentration, supply chain issues, Taiwan, Nvidia, geopolitical risks. And then state and local risks, people are starting to say, "We don't want these giant power centers and giant data centers in our neighborhoods." So it's really a complicated issue and I think Camar's right, it will resolve itself over time, but there are certainly concerns that revenue for AI is outpacing cost and that infrastructure booms historically overshoot their projections, which is what you see in a lot of these major companies. But in my lower and middle market world, if you touch on AI and your business, whether it's an HVAC contractor or a construction company, you're a hot asset and we're seeing 20 to 30 times multiple EBITDA.

So anything that connects to AI is hot and anything that doesn't have an AI component or isn't adjusting quickly enough to AI is going to fall behind.

Dean Peterson: Yeah, absolutely. John, what are your thoughts on energy and how we're looking in 26 and 27?

Jon Zefi: So I think we're started the conversation by focusing on what the hyperscalers are doing. So I think we should spend just a second there and then we'll put it in context of what's driving some of the tax benefits around the deployment on the data center side. So we have $700 billion in data center projects projected to take effect in 2026 alone. So let's just pause there Amazon has indicated their spend on data centers will be around 200 billion. Google between 175 and 185 billion Meta between 115 and 135 billion. Valor Capital led along us Colossus two investment now for somewhere around 35 billion that capital was called at the end of 2025 into 2026. A few things are driving that. First, bonus depreciation on data center build outs is a huge driver. We have 100% bonus depreciation under the one big beautiful bill. You overlay on top of that cost segregation studies that reclassify components into five and seven year property from 39 year property.

That can drive year one deductions of 60 to 80% of the building's depreciable basis on a purpose built data center. Think about that. You're deducting potentially 80% of the building's depreciable base in year one. Second area that's a key driver here that we see many of our clients focused on are the energy credits, both post the O triple BA and the IRA remnants. So two sections of the code that I'll highlight, Section 48 ITC and 45 PTC, those are the frameworks that survive the one big beautiful bills modifications. They still apply to behind the meter generation battery storage and certain grid interactive assets. Stack that with bonus depreciation and you're looking at after tax IRR of between 200 and 400 basis points above pre-tax numbers. The third area that we see and a significant component, remember that this is a real estate play. So there's the real estate structure going on here.

Data centers are increasingly being held in reach structures and qualified opportunity funds, which we can talk about where the underlying real property qualifies under the qualified opportunity fund structure. You can locate a data center in a QOF and have it housed there or in a REIT. So the choice between an operating company, a REIT and a QOF wrapper materially changes the tax profile. The fourth area that's driving a lot of transactional activity is the state and local side of the equation. You're seeing states like Texas, Virginia, Arizona, Iowa, Ohio, and Tennessee are competing on data center incentives, Texas being at the forefront of this. There are sales tax exemptions on equipment purchases. Obviously this is a huge build out, property tax abatements, electricity tax breaks. The state package can be eight to 15% of the total project cost, which is important. The one thing that I just want to highlight for folks, what we're seeing just on a day-to-day basis and our IRS controversy practice has informed me of is that the IRS is already scrutinizing aggressive cost seg positions on AI data centers.

The technology lifecycle of three to five years for compute is not the same as the building lifecycle of 39 years. So sloppy classification in your cost segregation studies will get challenged, it will get reviewed. So I guess the important takeaway here is that an AI data center is not a building. It's a portfolio of assets with different lives, different tax treatments and different tax attributes with different exit profiles. Treating it as a monolithic real estate investment leaves seven to nine figures on the table, which none of our clients can afford and has to be a key consideration for most of our clients.

Dean Peterson: Yeah, definitely. So you're seeing data centers, is it creating any new tax rules? Is it forcing tax law to evolve so the IRS can get its clause into companies that are using AI data centers?

Jon Zefi: I think what you're seeing on the state and local front is a more proactive response to the desire to garner more and more of the capital investment in their state and local jurisdictions. And so the economic incentive groups in the various states have been extraordinarily proactive and flexible in structuring their incentives to encourage data centers to migrate to certain jurisdictions, Texas being at the forefront of this. I think that's where you're seeing it. I think the constraint on the federal side is that many taxpayers probably engage certain advisors to take the greatest advantage possible from a tax perspective without spending enough time realizing that because of the order of magnitude of the dollars spent, this is going to be a hot button issue for review by the IRS. And so spending a little bit of time maybe taking a more conservative approach on cost said positions would behoove clients to avoid any potential future IRS controversy down the road when and if the data center gets scrutinized based on the positions taken on the tax returns.

Dean Peterson: Sure. Well, thanks for that, John. Let's take it back to the client and let's talk about AI as an investment theme. Kanwar, as a family office advisor, how should families think about AI as an investment theme and where is the real money being made?

Kanwar Singh: We touched upon this a little bit earlier, but I want to give you a couple of stats that I think will be helpful. There's been a lot of talk about the fact that markets have certainly run and rallied quite a bit off of the lows that we hit following Operation Epic Theory, and a lot of that has been driven by the whole AI story, right? Underneath that is a belief that earnings are projected to grow at a really rapid pace and earnings or visions have gone upward and upward upward. Now what's interesting is we've seen the emergence of this circular economy in the AI world. And by that I mean there were articles not too long ago about the fact that you saw OpenAI making a very substantial investment in Nvidia, which then made a substantial investment in Oracle, which turned around and made a substantial investment back in Nvidia.

So they each are owning each other's stock. Now what people don't realize is that a good portion, 34% to be precise of the hyperscalers income and earnings growth is attributable to other income, which frankly is the right up and the value of the investments in these other companies. And in fact, specifically nine percentage points of earnings growth that we've seen, the 23% that we're talking about in terms of earnings growth, 9% of that 23% is simply the circular economy and the writeup and the other income levels.

Dean Peterson: No, that's very interesting. So

Kanwar Singh: If we start to back out that 9% paper gain in the value of each other's investments and look at what's actually driving the growth and you apply whatever multiple you want, you might suggest that perhaps we're looking at markets that may have front loaded a lot of the future growth and returns, which is why I was speaking earlier about the fact that from our perspective, from the family office perspective, a lot of our attention is going to the harder assets that we know are likely going to be involved and there are a variety of vehicles. John did a great job of talking about whether there are REIT structures, evergreen wrappers, or otherwise, where you access to a variety of different opportunities to participate in the infrastructure underneath the AI economy, right? So whether that's power generation, be it renewable power, natural gas or otherwise, be it the data centers, although I think that we've all talked about the fact that being a little bit cautious about not biting off more than you can chew there and recognizing that technology will make that more efficient is probably a good idea, but then there's pipelines and there's other forms of infrastructure that are all sort of at play here.

I think there's a lot of reason to be cautious about the broader traditional real estate space, especially if you think forward to where AI is going, if we actually hit artificial general intelligence, office buildings like the offices surrounding me where I am right now could come under further pressure and they've already felt that. And so picking your spots is going to be relevant for us. It has been largely more of the harder assets in the infrastructure space or looking at private businesses where multiples, and this is sort of the premise of this conversation we've been having, multiples in the private markets are substantially lower than what you're finding right now in the public markets. And so why pay that 30X multiple and perhaps you can get something at a more reasonable multiple rate in private markets, especially if you look a litle bit down market from the mega cap deals where there isn't as much activity.

In fact, there's a lot less activity happening in the private markets at the sub, sort of the middle market and lower middle market levels. You find yourself with the opportunities to buy something at a lot more reasonable multiple.

Dean Peterson: Great. I'd like to drill down on something you mentioned, Ted, and I think this is right up your alley. From the deal side, are you seeing AI infrastructure assets command premium valuations and how's that changing your diligence or is it?

Ted Rosen: Yeah. So the answer is yes, it's increasing their high value assets and I think we talked about it earlier, but I think I would like to explain how from a legal perspective it's changing our due diligence. So at Akerman, we're extremely excited and optimistic on how AI will fuel growth for the firm and part of that is in the way we do due diligence. I was recently successful in pitching a deal where there's 250 leases involved and we can, in the old days, I would have three associates working on a week for a week summarizing those leases and now I can put it through Harvey AI and in a matter of minutes you can get a complete summary of the documents.

Dean Peterson: And

Ted Rosen: Ken,

Dean Peterson: When you say the old days, do you mean last year?

Ted Rosen: Yeah, like last year- Two weeks ago,

Jon Zefi: Dean.

Ted Rosen: Yeah. Personally, I can barely turn on my TV so I'm a little bit reliant on my associates now to help me with the AI, but we actually have mandatory training for AI and I think just as a global comment, you either learn and grow at this or you really fall behind. So in terms of due diligence, whether it's infrastructure assets or any other assets, due diligence is going to be faster, more efficient and more accurate utilizing AI. So we're super excited about it.

Dean Peterson: Yeah, that's great. So Ted, let's keep going with you for a minute. How does AI change your outreach for your firm, for your practice, to your potential targets?

Ted Rosen: Yeah, I think a lot of the tools that the investment bankers now have at their disposal we do as well. So I think we'll be able to do a lot more targeted outreach to potential clients. So that's really valuable. I think everyone is always talking about legal fees and legal fees are too high. So the question becomes,

Is AI going to make us more efficient? Is it going to make us better lawyers? Will it create more business and will we be cutting staff? And the answer to those things we believe is that we won't cut staff. We believe that we'll get more business. We think we're trying to be at the forefront of using AI to help our attorneys do better documents. I'm going to say one comment that I really want to make for the audience and that is when you're using AI tools yourself, whether you're a lawyer or a non-lawyer or a young lawyer, don't rely on it. I mean, you have to really be careful. So I'm seeing young lawyers utilizing AI tools to send me documents with comments that are not market. So we have a client going out to market, so we're getting back NDAs. So NDAs, I've done hundreds of them.

We know every comment, but they're coming back with new comments that are clearly AI generated because I've never seen them before and they're like, "We want to know when you're not soliciting an NDA who your senior management is before they even look into my data room." I'm like, "No, it's not market." So I think you have to be really careful as a client to make sure that or someone using these tools to make sure that you're using them properly, efficiently and smartly and still get the right people reviewing them. I can do a will theoretically, but I'm not competent to do my own will. So don't just rely on a form that you get on Claude.

Dean Peterson: Yeah, that's right. In my practice one of the things we've implemented are additional levels of review for anything that's going out to a client. We want to make sure that any advice we're giving has been reviewed by a partner or a director, not just something generated by AI and directly sent to the client, that just creates way too much risk. So we're all being very careful with that. Now, John, let me continue with this point. Go ahead, Ted.

Ted Rosen: Dean, just one more thing. So you raised a really important point and they're calling it governance, but they're calling it AI governance. You need to build internal controls in a company, whether it's a legal department or non-legal department to make sure that a salesperson isn't just amending contracts without it being reviewed by the legal department and it's happening and it's going to happen because when it comes out of ChatGPT, it looks right, but that's not good enough. But it's not good enough. And that's where the danger is in all this. And that's why my job is pretty secure for a few more years.

Dean Peterson: Well, John, I saw you vigorously nodding your head over a few of these things. So I have a question for you. So how is AI changing the tax practice itself and what does that mean for our audience?

Jon Zefi: Dean, I'll get to that in a moment, but I want to address a couple of other points that Kenwar raised and Ted raised just on premium valuations and due diligence really briefly. So getting back to Kanwar's comment, the honest summary is premiums are real. They're justified for power de- risk hyperscaled anchored assets, but they're increasingly more speculative as you move down the stack to things like neoclouds. So specialized neocloud providers face a tremendous amount of emerging pressure right now. They previously capitalized on chip shortages, but improving access to advanced processors is reducing that competitive advantage for many of the NeoCloud folks. Most least their data center capacity from larger operators and their long-term business models are being reassessed. So we're seeing GPU as a service vehicles that were underwritten in 2023 and 2024 on a thesis of permanent H100 scarcity are now being exposed in the marketplace.

So you should raise a couple of points that if you're seeing fund structures or deal models built on GPU residual value holding flat for five years, you need to push back really hard on that presumption. From our perspective, due diligence has effectively become a hybrid of project finance, utility regulation and partnership taxation, which is good news for Ted on a billable hours front, I think. But if you think about it, what we're looking at, we just mentioned it a moment ago and Kenworth reiterated my point, read eligibility for the real estate layer, key consideration as part of due diligence. Qualifying income tests when GPU as a service revenue gets blended into a transaction. State level apportionment when a single platform spans multiple jurisdictions or you're comparing Texas versus Virginia versus Arizona on an incentive play partnership level allocation on joint venture structures and deals critical.

Sales and use tax on equipment, extraordinarily critical. We mentioned bonus depreciation interplays with the 163J limitation on highly leveraged builds in our first session. IRA energy credit transferability where renewables are stapled to the load. This is where deal economics quickly leak out and this is where we need to be ahead of the curve in sort of our tax due diligence and structuring. So we keep bouncing into this reiteration of the fact that tax is part of a three-legged stool here. It's investment thesis, it's deployment of capital, and then it's tax structuring that overlays and takes advantage of all the opportunities that are presented today. You asked the second question, which is how are we leveraging AI in our tax practice? And obviously Ted mentioned it moments ago. It's a similar thesis. Yes, we want folks to leverage large language models, but we cannot forget the human oversight, human intuition, human experience and practical application to join venture deals that we've worked on for the last 20 years are critical to this.

So relying solely on large language models to structure a transaction or analyze the economics of a waterfall without human intervention and oversight is a mistake and you need to avoid those. So AI governance is a key consideration. We had Eisner feel that the human factor here is critical and that our human capital resources should be leveraged to their greatest benefit, which is their expertise and understanding in my particular area of tax and we should deploy that knowledge and know how to extend the capabilities of individuals assuming that we have appropriate corporate governance and AI governance in place.

Dean Peterson: Yeah, that's right. So AI is a great research tool. It's an amazing research tool from what we've seen and used properly can really help create a lot of efficiencies and some great answers, but AI right now is not replacing the judgment calls or the business judgment, the experience that some of us, the advisors have along the way. However, I do have a question about the tax practice about family offices in general. Kanwar is a family office advisor, how should families ... I'm sorry. Kanwar, I want to know, do you feel that AI can eventually replace investment advisors in the family office practice or any practice?

Kanwar Singh: Eventually is a big word.

Is all changing. Let's just talk about the foreseeable future the next two to four years. I see it as very unlikely and for a lot of reasons. I don't think the judgment that's embedded in the large language models and the biases that are baked into the training that they pursue gives enough comfort to families with nine, 10 figure wealth and more, or even eight figure wealth. It's the kind of thing where what we're seeing is clients are expecting and demanding that their advisors across the board, whether it's on the CPA side, the legal side, or on the investment side, are leveraging technology and AI to make them better, but they still want a human. And I think that the larger sort of maybe sociological part of this is I suspect that the reason we're seeing more and more desire for analog experiences, whether it's an in- person concert or an in- person talk, or even I see with my kids, they want vinyl record players and vintage typewriters.

I think there is a desire for the analog to give some sense of substance and to sort of revolve back to something that feels more human. And the idea of having someone in whom you trust has always been the foundation of each of our three practices. And I think the biggest issue is right now there still isn't that level of trust in AI, understandably so, nor is there the level of judgment that's informed by experience. And so I don't think you're going to see

Our respective worlds be replaced by AI, notwithstanding the dystopic prognostications of the elimination of all white collar labor. I don't think that that's really realistic, at least for the foreseeable future, but I do believe that if you're not on board or if you're not getting up to speed, you're going to get left behind because each of the three of us on this call have become power users, know how to leverage it to effectively make ourselves superhuman and thereby deliver better results for our clients. And so for anyone out there who's on the advisory side, I would strongly suggest that you get up the curve and not resist it and not fall back to that place where you thought, "Well, I tried it last year and it kind of sucked." I mean, this is not then and I promise you this by the end of the year, it'll be markedly different.

I'll give you one more anecdote. There's a family member in my family who is a pretty senior software engineer for a meaningful company. You can probably extrapolate to the handful of companies it could be. His boss, which is the head of technology, told him that by the end of the year, neither of them will be doing any more coding. I do believe that there are certain sectors where it is relevant. You've seen a 20% decrease in employment in the tech sector among sort of software engineers.That's because that's where they started. If you can teach AI how to code, you can teach it to do a lot more things, including learning itself. I think that that's where maybe some of the fear is going.

Dean Peterson: Yeah. Ted, what do you think for the legal profession? Is AI going to replace the M&A lawyer?

Ted Rosen: Over my dead body, maybe. No, absolutely not. And John said it and Kamaru said it, AI doesn't have EQ, right? So it can't negotiate a contract. I can put two different purchase agreements in the system. I can see the changes that are made. I could see potential suggested paragraphs, but when it comes to actually negotiating those changes, that's not going to be done by a computer. That has to be done by the lawyers who are experienced in business and corporate matters who know what's fair, that understand leverage and each transaction is bespoke. So it's a tool. We use the tool to make us better at what we're doing. So at my partner retreat a couple of weeks ago, they had a guy named Zach Cass and he was one of the early employees at OpenAI and he wrote a book called The New Renaissance with AI and Renaissance in both letters.

And I suggest anyone who has fears or wants to learn more about how AI is going to impact society, you should read the book because Zachcast is an interesting guy because he has a background, a degree in history and in computer science. So he traces all technological changes throughout history, the printing press, electricity, cars, and there's always a gap between adoption and founding of a technology and the cost. None of those are around anymore. Everyone has access to AI. The process of AI is going to change education. It's going to change medicine. It's going to change schooling. It's going to change everything. But as Camwar said, and what Zach has also reiterated, because Camwar is a humanistic human being, that's one of the things I love about him. We all need to balance being on a screen using AI but not losing what makes us human and that's human interaction.

And Zach has says, "My biggest concern is I want my children to go play in the park." And I think it's the same for us. There are nights where I'm on ChatGPT at two o'clock in the morning asking stupid questions. That's probably not a good thing. But at the end of the day, I think

Especially with these panelists, we all have what's hard to teach and that is it's an inherent desire to be great at what we do, to care about our clients and to have pride in what we do and it's very hard to replace that. And I want John and Kamar on my team. Chat's Claude is nice and there's great apps, but I want to be able to ask them when I need their expertise and vice versa. And that's why that teamwork that won't get changed by AI.

Dean Peterson: Yeah. So there's real value there in communication, human interaction, human connection. That's something that's missing from the AI world that cannot be replaced. So gentlemen, I'm going to ask you a couple of questions and I'm going to start with John and I want all three of you to answer this question. I just want to hear what you have to say. How should I think about AI as an investment versus a tool? What do you think, John?

Jon Zefi: So it's a really great question, right? So let me put my Eisner a tax partner hat on. I clearly view large language models as a tool that will extend my ability to be a force multiplier from my clients.

There is no doubt in my mind that in the last six to 12 months I've become even more efficient and hopefully that efficiency has created greater opportunities for me to engineer increase value for our clients. So from that perspective, it's been a tremendous benefit to my clients in converting what I do on a day-to-day basis and having a multiplicative effect on positive outcomes for our clients on a daily basis. So that's the tool side of the equation. On the investment side of the equation, obviously this is idiosyncratic to my own personal views and it doesn't reflect the views of the firm, but you can't deny the fact that when over $800 billion of capital is deployed in any given year, that that is a tremendous driver of value and GDP development here in the US. And so it's incumbent upon all of us to pay very close attention to where capital flows are going and maximize the benefits from an investment perspective to each one of our portfolios, given our risk tolerance.

And that's a different answer for each person that's listening to this because they come at the problem and at the opportunity with different expectations and sensitivities. And so I think that covers my view on

Dean Peterson: It. Yeah, good. Thanks very much. Kanwar, same to you. How should I think about AI as an investment versus as a tool?

Kanwar Singh: I think that this time period, they say history doesn't repeat itself, but it rhymes. It looks like a lot of prior time periods where you saw these big transformations, certainly centered around technology, right? Clearly there will be significant value creation. Certain companies are going to explode. You're going to see industries follow on that were unexpected that will be developed. They'll also be really not evenly spread or distributed and it's not always obvious how to capture it on a sort of consistent basis, et cetera. And the issue is oftentimes the initial leadership doesn't represent the companies that will ultimately be the ones that fundamentally capture the value. So you've got to be careful in diversifying your exposure across those different areas, which is why we talked about earlier about maybe not just betting on the hyperscalers, but also the infrastructure play that seems a lot less sexy underneath.

When you think about the company specific risk, the execution risk, the valuation risk, you've got to balance all of that. I do happen to believe that this is changing a lot and it's happening at a pace that I think will be faster than society's ability to fully adapt. So there will be some gaps. And I think my concern underneath this is the gap between those who invest and own assets and those that are sort of functioning from a paycheck to paycheck basis. I think you're going to see a separation there. But bottom line, from an investment perspective, I would be wanting to spread my risk. I'd be wanting to look at different sectors within it and I'd want to make sure I understand that the companies that first build the technology aren't always the ones that win at the end of the day. So don't put all your chips on that one play initially.

Dean Peterson: Sure. Thanks. Ted, let's go to you. Bring us home on this question. How should I think about AI as an investment versus a tool?

Ted Rosen: I think it's both. So I think because in many ways AI is becoming an app, right? It's the same way our cell phones now are full of apps. AI is to some extent just another app. What I'm seeing, especially for investment banking, is when you look at these different AI tools, they're extremely complicated. So I think part of the investment is not just getting an app but getting consulting firms to help you implement it. I think that's one of the keys. So AI is a tool to make you more efficient, to make you smarter, but you have to invest money to learn how to get your team to utilize it and maximize it as we discussed and to make sure that it's done properly.

From an investment perspective, and I'll use John's caveat, this is my own personal opinion about that of my firm, but my daughter is in tech sales and AI and she was looking at Anthropic and OpenAI and Nvidia and it's going to be really, really interesting to see how they take their technology and create a business. And there's a big leap to go from having an AI tool to create a business, to create a SaaS business, to go into various segments. So it's going to be a really, really interesting ... There'll be IPOs in this space, the valuations will be crazy, but there's still an execution play and I think that's what Kenworth is talking about. It's going to be very, very interesting how this all plays out, but at the end of the day, I think you either adapt or die. I'll quote Zach Cass, one of the things he talked about was autonomous vehicles.

So if you talk to 10 people, seven people will say, "I think it's a terrible idea." But the other three people who've seen the data will realize it's going to happen. It's not when. It's not if, it's when. And I think all of this AI adoption and development, it's not if, it's just when. So we and our clients need to be ... If you want to win in the next five years, you better invest in these tools and grow with it.

Dean Peterson: Great. So guys, I really enjoyed these last couple of weeks speaking with you off camera, putting these scripts together. It's been really, really educational hopefully for all four of us. So I want to get to your final thoughts on these two episodes. From our first episode, from our second episode today, what are your final thoughts? What should the audience take away from these episodes, changes in M&A, AI, and investment? John, let's start with you.

Jon Zefi: I think the single largest takeaway that I would say is that individuals need to demonstrate some certain amount of interest and invest time understanding the tools that are available to them currently and seeing how they can best utilize those tools given whatever they do for a living to generate efficiencies and deliver value for clients. So that's part and parcel with what needs to be done. So a little bit of desire to kind of spend some time exploring that most of us don't do. We don't take an hour out of our day to kind of learn new things on a regular basis. Many of us don't. I think it's incumbent upon all of us to invest the time and do so. And then once you begin spending some time with the various tools, you'll see opportunities open up in your careers and providing services to friends and family.

Second thing I just would like to mention to everyone, and that's more directly relevant to what we do, Dean, you and I, is that tax architecture and structuring really generates alpha in these investment opportunities and it needs to be part of the early part of the conversation when we're looking at potential structuring and opportunities that are coming across our desks and be focused on the notion that it really just does generate alpha as compared to other investments that don't spend the time and effort setting up the appropriate tax structure. So personal view and a professional view on those two points.

Dean Peterson: Yeah, John, I totally agree. There's a couple of things on the tax front that I think everyone should be aware of. First of all, tariffs have changed the rules and they are not going away. You're going to have to plan for tariffs from now into the future. Pillar two and it's all the rulings around it are going to change the rules. So one of the things I want to note is that anybody who has pre- 2024 structuring, your structuring is now stale. You should revisit that with your tax advisor. So that's a plug for anybody interested in talking to us about tax advice. Conmor, let's go to you for your final thoughts. What are your takeaways or your final thoughts for these two episodes? Well,

Kanwar Singh: First I'll note that it's really enjoyable doing these kind of conversations with friends and people that I respect very highly. And so the three of you, this has been tremendous, really enjoyed it. I have three words that came to my mind, curiosity, humility, and collaboration. You cannot succeed and survive in this dynamic and in this market unless you are curious and you stay curious and open. If you have closed perspectives on whether it's the tax law or whether it's what's happening with the advancement of technology, I think you're going to get left behind. You have to stay curious and open. That then comes with a dose of humility, knowing that you might not be able to know everything and you probably can't just rely on the ChatGPT bot or the Claude code or whatever else to come up with the answers, which then leads to collaboration.

I think that humanity part of it, but also the knowledge part of it, bringing in people like John and Ted and yourself into a transaction, into a conversation, just even to scenario plan is always going to result in a better outcome. It doesn't mean that you don't use the tools to help inform some perspectives and maybe ask better questions, but ultimately we have to stay collaborative.

Dean Peterson: Yeah, absolutely. Thanks for that, John and Kanwar. And Ted, you're the only one wearing a jacket today, so I figured I'd save the best for last. So Ted, give us your final thoughts on these two episodes going forward. What should the audience take away from this?

Ted Rosen: It's a great question. I'm an M&A lawyer. My passion is helping entrepreneurs sell their companies. So I think the takeaway is that there's always uncertainty in the economy. There's always uncertainty politically. I've never had a client say, "Ted, I'm really sorry you made me a millionaire. I'm really sorry I don't have generational wealth." So I think you continue to move forward if you want to sell, start the process sooner, prepare, get your team together. And I think if you're a buyer, I think it's a great opportunity for buyers because a lot of my sellers are 50, 60, 70 years old, maybe wanting to retire, may not have adopted AI. And I think this is going to create lots of opportunities for PE firms that are much more sophisticated as buyers to use these tools to make these portfolio companies more successful. And I agree and this is something in the M&A world that is really important to me and the AI doesn't teach this, but it's being respectful.

It's being respectful of sellers, being respectful of buyers, being respectful of advisors, and working together collaboratively and using AI as a tool, but not as the solution.

Dean Peterson: Yeah, definitely.

Ted Rosen: And it's exciting. Don't be afraid. I think we're all saying the same thing. Don't be afraid. I think the world's going to be a better place at the end of this. Let's hope so, but let's do it in a way that is better for everyone.

Dean Peterson: Yeah. Thank you very much, gentlemen. And I'll leave my parting words for the audience and I'm going to hark back to something that I was mentioned earlier. It's adapt or die. And those are the three words I'm going to use. Conroy used three much more constructive words, but I'm saying adapt or die. So gentlemen, I want to thank you very much for these two episodes. Very informative, very interesting, and a really good time speaking with the three of you. And to our audience, thank you very much for joining us today. I'm going to toss it back to Savannah for some final comments on housekeeping.

^{Transcribed by Rev.com AI}