Transcript
Don't do that, kitty. Excellent. Hello, Matt. we're here for ya. ready to record. We got a fun episode about AI. So why don't we why don't we do that? three, two, one. Welcome to Shelter Primacy from Free Float Media, a podcast about activist investing, securities law, and all the ways the financial and legal worlds intersect and collide in real life. And Lipton and I are here. Anne Lipton is a law professor at the University of Colorado who teaches and researches securities and business law. She holds up the legal end of the podcast. And that's Mike Levin, an activist investor who lives and works in Chicago. He covers the financial side of our podcast. So you're doing well today, Ann. I know you have a bunch of meetings today, right? I've got a couple meetings trying to get a settlement put together for an activist situation that we're trying to close up so we can work on that a little bit, and then I'm gonna go see a grandkid a little later today. So it should be fun. Anyway, I know. So you know, Ann, we haven't talked much about AI, artificial intelligence on our podcast. I'm okay. Yes, sadly, and you? that's nice. Very cool. I'm sorry, I believe that is super intelligence. I I yeah yeah. Super. I was just gonna say we shouldn't be adopting that. I'm not calling it the Gulf of America either. Okay. Well, let's let's let's adopt it. Again, we have all sorts of naming conventions we worry about, like big A and little A. Activism, financial, whatever. I think we can continue to call it AI here. Great. Okay, that's but that's been what it's been it's it people researchers and scholars in the subject have called it that for decades. And I don't see any reason for us to diverge from that. I am gonna continue to call it artificial intelligence. So let's do that. AI. obviously it occupies like huge amounts of like the public discourse these days. That's all we talk about, hear about, read about. but we haven't really covered it. Not that there's nothing to say. There are some things on a podcast about securities law and activist investing we can say about AI. we just haven't kind of gotten our ideas together yet in a coherent fashion. Until now, I think we have some things we could say about it. Mm. so yeah, so why don't we plunge into that a little bit? So So I have a question for you as a baseline, before we talk about that, w are we like will AI lead us to extinction or paradise? Sure. Yeah. I I I have no idea. No. you don't have a view about whether this is gonna be the end of mankind as we know it, or whether it's gonna lead to I don't think it's gonna be the end of mankind as we know it. I I I am very concerned about it being essentially forced down everyone's throats when they don't want it and leading to something of a collective action problem where students feel like they have to use it to keep up with their peers, but they also know that it'll interfere with their learning and that's not great. Okay. Right. Okay. So there's there's a more immediate impact that you see like in the classroom. Yes, exactly. I think that AI evangelists enjoy talking about the doomsday scenarios and not the everyday problems. interesting. Yeah, so I'm I'm agnostic here. I don't think it's I have two observations before we get into some of the specifics. First, I don't think it's gonna lead to either. It's not gonna lead to human extinction. And it's not gonna lead to like a workers' paradise where all we do is read and write poetry. Well or well maybe A AI will f A AI will free us up to write poetry or something like that. I think it's gonna be fine, okay. Well no, because the po the and the AI writes the poetry. We're supposed to be just crank doing the blue collar work to maintain the machines. No, no, that's the whole thing. AI writes the poetry and we are confined to to fixing the machines. It's gonna be some somewhere in between. And it's gonna it's gonna take a lot longer than people think. That human behavior and human habits are very slow to change. And, you know, people will kind of grudgingly kinda go along and eventually, you know, customer service stuff, which is already starting to see changes there, will, you know, give up to AI applications and so forth. But it's gonna be decades before we really see a lot of significant impact in our day to day life. Yeah. I that's my guess. so with that backdrop, there's sort of some diverse AI or like AI adjacent subjects we thought we'd talk about. we'll talk about the coming anthropic IPO because there's some interesting investing and also corporate governance implications. I I don't think we can avoid talking about NVIDIA as Mm. Yeah, it's rather unique. No, as much as I've tried. a supplier to the the picks and shovels of the AI world. And then there's some more direct implications of like AI and activist investing or corporate governance and so forth. So maybe we'll try to tackle those three subjects. All right. So but before that, Ann Ha ha. Mm-hmm. Mailbag, yes, you everybody can email us if you like with questions, comments, as you wish. we are at shareholder primacy one word at freefloat one word dot L C. All right. beginning with anthropic. and its IPO, which is supposed to get priced tomorrow and will be floated I think on Thursday is the current schedule. Right. So they've been out in the past well, I the S one's been out for like four three, four weeks. Yes. Yeah, that's that's the goal. Yeah. For yeah, for a couple of weeks now. Yeah. Yeah, and then and it got leaked much to your chagrin and frustration. Much to my chagrin, it didn't get leaked. I mean leaked suggests it was involuntary, and I don't believe it was. I think this is just an SEC that doesn't care about these things anymore. I don't know how I teach it next semester. Right, but the SEC SE does didn't give a shit about about those kind of disclosures or enforcing that little bit of law. so it's coming. And we had a couple takes on this. it's just your usual IPO, right? Just you there haven't been that many this year, but it's just your usual IPO. Yeah, right, you know, a valuation like is comfortably over one trillion dollars. There's only been one of those before. We talked about SpaceX. Well now it's this is becoming the usual IPO, apparently. Yes. And they're interesting, they're raising close to a hundred billion dollars, which is actually more than SpaceX raised in its IPO and you know, magnitudes greater than much any other IPO has raised. and it also has a I shall we say, I think this is charitable, one of the most complicated corporate governance structures that would be thought of for any kind of company. And there's probably some good reasons for that. Yeah. Yep. Or they think there's good reasons for it. But Right. So all right, let's talk a little bit about valuation for a second. I again, there w which kind of affects, you know, some of the decision making here because I think there's this existential conflict in the leadership there between making money and saving or preserving humanity. And they can't avoid the money that's about to get made. you know, they were targeting a valuation of about well they say they think there's good reasons for it. Yeah. this week about one and a half trillion. So I said comfortably over one trillion. So I think one and a half, probably one four, one five. But interestingly, they had been talking about a range of anywhere between a little over one trillion. So after some of the initial NVIDIA news came out a couple weeks ago, you know, the things started to crash a little bit. especially for Anthropic, because Anthropic, of course, is what we would call a pure, pure play Mm. Yeah. Right. AI company. You know, NVIDIA sells chips to all sorts of people. Okay, Google has an advertising business. You know, Meta has social okay. All anthropic does, all its companion open AI does is AI applications. And they you know. so anyway, so the valuation was gonna go anywhere from it went as low as one point one at one point when they first had a floating the idea it was gonna be close to two trillion. Mm-hmm. Yeah. Yeah, and it's not just NVIDIA, obviously, it's also the election and the backlash against data centers and artificial intelligence and all that. yes. Right, exactly. Right, right, right, right, right. So so looking at about one point five trillion, they're looking to float six percent of the company. That's where we get to the hundred billion dollars. again, the only comparable IPO that we have in like history is the SpaceX one from two months ago for three months ago, which, you know, they raised about eighty six billion dollars, but at a somewhat higher valuation. Because again, SpaceX is not pure. SpaceX has space and telecom. Right, exactly. Well, I mean Anthropic has an actual business as we'll see in a minute. Anthropic's run rate and again, th the the metrics trying to measure these businesses is a little weird. It's basics, yeah. Mm-hmm. Right, it's got like actual businesses. Yeah, all right. Okay. You normally you'd look at the you know, the last fiscal year, you look at fiscal quarters. The way they're measuring things is like on trailing twelve month metrics. And so their last twelve months, as of I'm gonna guess the most recent quarter, I think was the number I saw, was eighty billion dollars of revenue. Okay, this is a company that had zero revenue like three years ago. Okay. So they have, you know, grown into this insane, you know, you're comfortably within the upper echelons of the Fortune 500 just on a revenue basis. Okay. But again, they're not, you know, it's it's really unusual to look at just a last 12 month. I mean you look at the last 12 month number for various kind of analytics, but in terms of reporting, you know, you're not gonna look at last 12 months. Anyway, so last 12 months they evidently have eighty billion dollars worth of revenue and the valuation is somewhere about sixteen or twenty times that And again, that's not even a real conventional. Usually you're like, you know, earnings, but they're not making money. They're losing money. For now. I mean, at this kind of run rate, they should be making it pretty soon. But they're not going to you know, you can't base this valuation based on profit. It's all speculative. but importantly, that valuation, sixteen to twenty X, is very comparable to some other like really fast growing tech companies, you know. Yeah, yeah, yeah. Broadcom, AMD. You know, these are companies that are also supplying the Okay. No. couple other interesting tidbits about it is that they are dedicating twenty percent of the IPO to retail shareholders. I've already I don't know if you've seen it, but I've gotten a couple of things from my brokers saying here's how to grab shares in the anthropic Well, I my my broker is very, very aware that I would not buy. well, okay. So maybe so I and I'm not I don't wanna I'm not gonna buy it anyway, it's way too speculative here. you know, we'll talk about NVIDIA in little b in a little while, but I think you know the the the I don't I'm not making a call here, but I think the likelihood of some sort of a crash in the sector in the short term is n is comfortably over zero. you know, either because the speculative fear fervor is going to abate Ha ha ha. Or because there's some stuff we're gonna learn or it's gonna start growing us slowly or whatever. But you know, the idea that this is coming out at a one point five trillion dollar valuation and could easily sink to merely one trillion and lose a third of its value, is enough to give me pause about, you know, try and again, I'm not getting an allocation, the IPO anyway. So anyway. and then the other tidbit is that it is six times oversubscribed. So I don't know, I again I don't Fair enough. I'm not I've never been a banker, so I don't work on IPOs. But you know, they spent the last two weeks kind of getting bids from bankers, investment banks saying, Okay, I want this much, this and this they've gotten bids six times the shares that are available. Yeah. Yeah, i I I don't know I don't know that that by itself is particularly impressive for an IPO. I mean, you're supposed to be oversubscribed and by a significant amount. si six times is a lot. Which is a it's a testament to the fervor around this kind of pure play kind of thing. And again, that's also suggests to me that maybe it's a little a little there's maybe a little exuberance here. I know if it's irrational, but there's a lot of exuberance. So anyway, so that's some of what's going on and that's kind of the backdrop for how they've kind of structured the company. and and the leadership and the corporate governance and so forth. and I thought that would be kind of we we wanna talk about that. So this is no mere dual class share company now, is it? I don't think. There are no, no. It it is a very Baroque and odd corporate governance structure. f f yeah, but actually we can start by saying legally it's not even a normal corporation. It is a it's a public benefit corporation. With five five classes of shares. okay. W go ahead. so it's w it's one of those small number of Delaware entities. Yeah, that's a public benefit corporation, meaning that so a regular corporation, the board has a fiduciary responsibility and maximize shareholder value. They have a tremendous amount of discretion in deciding how to do that. But that is ultimately they should, in good faith, act to maximize shareholder value. That is not true in a public benefit corporation. in a public benefit corporation, you have to identify the benefit, some kind of public benefit that you are advancing. Yes. Right. That and this is why, like, it's not Anthropic Inc., it's Anthropic PBC, Public Benefit Corporation. you have to identify in your charter some kind of public benefit that you are advancing in addition to a profit-seeking motive. And the directors are under a legal obligation to balance both the benefit that you are advancing and the profit motive. So they aren't, they do they don't have a fiduciary duty to just maximize shareholder wealth. They have a fiduciary duty to balance the okay. benefit being pursued alongside shareholder wealth. And in this case, the benefit that's identified in the charter is to responsibly develop and maintain advanced AI for the long-term benefit of humanity. Now, yeah. Humanity. And right and their and their S one had a lot of magical words about that kind of stuff too. Yeah. I mean now, I mean, I'm you know I mean th the reality is that for I I wouldn't say everybody, but for many of us, academics, the just the idea of the public benefit corporation is at least as currently constituted, is is very weak. I mean, there's a fiduciary duty to balance the sh profit seeking against the benefit that you identify as a but even in a regular corporation, you have so much discretion Yes. To decide how to run the company. That I don't think that the public benefit corporation legal status gives you any different amount of discretion. I mean, in a regular corporation, directors have plenty of discretion to decide to advance humanity. And they all they have to, you know, it and and a court will never question it because there's always, you know, a potential argument that doing so will ultimately redound to stockholders' benefit. So I don't think it really makes a difference legally. You know, I'll leave it to philosophers. to decide whether it makes a difference psychologically in terms of how directors approach their duties. But they they don't have a duty to maximize shareholder wealth. They have a duty to advance this benefit and shareholder wealth. and the w yeah, and so there are a whole bunch of different share classes. So there's the yeah Ha ha ha. So they're and these are these are share classes within this P B C corporation or whatever. Okay. Right, exactly. So they're selling, you know, they're selling to the public A shares, class A shares, which have one vote per share. There exist N shares which have no votes, they're just economic interest for like prior investors in the company that are like part of their business, like Google or NVIDIA. And I don't know and if though I I I mean, I assume to some extent that's because they didn't want big slugs of voting power to be handed to others, and also because for all I know, and I don't know whether there's an antitrust issue if they gave them. voting shares. But so so the end shares have no votes, and those are held by businesses that are essentially intertwined with anthropic. Yeah, they they gave like they were like part of various private placements and early rounds of investment. Yeah, okay. Exactly. Yeah. And but yeah, but the main is the main issue are the control rates. So the the public investors have one vote per share. Then the board, at least currently, is supposed to have seven people, although at least when I last checked, they hadn't filled all the seats. But it's supposed to have seven people. Four of those are appointed by a long term benefit trust. So basically this is a non profit. It's not unlike I think how OpenAI did this. It's a nonprofit that is essentially devoted to making sure that Anthropic stays on mission, this trust. The trust has no economic interest in anthropic at all. It has trustees who are self-perpetuating, so they replace each other. and the trustees get to appoint have just without economic rights in the company, they just have the right to appoint four directors on a seven-member board. and then separately. The founders, there are seven founders and they all have an interest in an LLC so that they and that LLC then collectively acts on behalf of the founders. And the founders, the LLC has a single share called F stock, which for founder, which has the right to appoint the remaining three board directors. F her founder, right, okay. Yeah. And also has 50.1% of the voting power, no matter how many A shares issue, for everything else that shareholders vote on. So it has so basically this is a floating ability to vote share. It gives you three of the f seven directors, but also anything else that might be put to a shareholder vote, like I don't know, merger or charter amendments or things like that. Equity, yeah, yeah, things like that. Equity issuance, sir. Who knows what it right. Anything else, it will always have 50.1 per percent of the vote, no matter how many A shares issue. So the trust gets to a point four of the seven directors, but doesn't get to vote on any other shareholder matters. So essentially the board is controlled by the trust and the founders. And then the trust of the four directors that the trust appoints, one of them has to be approved by the existing nominating and governance committee of the board. Which is dominated by founder directors. So essentially one of the four directors that the trust appoints as a compromise between the founders and the trust. The trust appoint names somebody, and then the founders have to app or the the nominate sorry, the nominating governance committee, which is dominated by founders, has to approve the trust's choice. And if they disapprove the trust's choice. wow. Nominated committee, right. That's when the employees get a vote, and this is called class E share, with no economic rights, but just allows the employees to vote. And if the employees approve the trust's pick, then the emp then that then that person becomes the director. And if the employees do not approve the trust pick, then the trust has to identify someone else. So what it really means is that the trust gets three directors outright, the founders get three directors outright, and the seventh director is something of a tiebreaker director appointed by the trust. Right. founder through the nominating governance committee approves it or the employees approve it. So, and the theory here is that through this structure, the founders, you know, they have their own vision of AI safety or whatever that they'll be able to enforce. But also the trust, since it doesn't have a financial interest in the company, they're supposed to also keep the company on mission because it's supposed to be a nonprofit that is just to make sure that things are developed safely. So the theory here is that the trust is enforcing this mission to not seek profit to the point where humanity would be destroyed. That said, I am skeptical of the effectiveness of this, which people keep asking, because I mean, among other things, like let's talk about who the trustees are. Like they're not AI experts, and the directors they put on the board are not. No, it's Ben Bernanke. Right. I thought they were. I thought the I thought the trust the trustees were It's very impressive people, but they're not like engineers. And the directors they've chosen are not engineers. So you have to ask yourself whether they have the ability to really oversee what the anthropic is doing. And this is something I'll just give a shout-out to Gad Weiss because he's a professor at San Diego and he wrote a whole paper on this. They don't have the expertise, and it's not even clear they have the information rights. I didn't look, yeah. Like I I don't know what kind of information rights they have to really view Anthropic because apparently, I mean, they'll have directors that they appoint the directors who are also not necessarily AI people. maybe, but w but but there there's some kind of agreement between the trust and anthropic that is called a resources and information agreement, but it is not as of this recording public. So Right. They can they can inform them about stuff. Information. Yeah, right. Yes. I don't know whether they have special information rights and they don't apparently for whatever reason have any particular AI expertise. So I am very skeptical that they can really are in a position to exercise control in a way that would, you know, main keep keep the mission. Yeah. Or helps them promote their mission for for safety and security and so forth. So go ahead. Right. And then it's well, th then just one thing happened, which is what happens when the founders are no more. Because one day the founders, I assume, will no longer be on this planet and or they can step down, and so forth. So essentially, I mean it gets kind of complicated, but what it really comes down to is when there are no more than two founders, when it's basically fewer than three, the whole structure changes. At that point, it just bec They will move on. They will right, yeah, right. Okay. The founder shares disappear. They once the the like they no longer have voting control. The employees no longer have tiebreaker role because there's no point because the founders don't have voting control anymore. At that point, there are only the share classes of the A's with voting power and the trust that can nominate four of the seven directors. And the board switches to a staggered board with the four directors appointed by the trust as part staggered over the over the terms. those okay, go ahead, yeah. Yeah. Right. Ha ha ha. So essentially it's somewhat controlled at that point, interestingly. The the LTBT will still be able to appoint four of seven directors, but the A shares will then have voting power on everything else that shareholders vote on, not just the remaining three directors, but also any other matter that comes to a shareholder vote. So that's what happens. Right. What? because the because the forty nine point nine percent will will will now be a hundred percent because the F s because the fifty point one at the F share zone will have disappeared. Okay. will be now be hundred percent because because is gone. Exactly. But that's happens when there are no more than two no more than two founders left. Okay, well, here's a couple of observations because those all those founders are relatively young people right now. It's gonna be a long time. They could step down, but they're not stepping down from this. Yeah. And they could also step down, but well, you know, people keep saying that, but you know, God laugh mortals plan and God laughs. Of course, but for the next small number of years for investment purposes, this is clearly a controlled company. Okay? And the two main shares that have an economic interest, which is the public investors and the strategic partners, have no say in what's going on, okay? Even though there's you know almost fifty percent, it's not fifty percent. So any kind of material vote outside the directors, the public shares that I would be Yeah. Yes. Yes. acquiring or any other investment we acquiring are you have no no say. No c it's controlled a company, okay? Right. No, yeah, yeah, yeah. There's y this is absolutely you have no voting rights. I mean, as a practical matter, there are no voting rights at this time. Right. So and I can't help but think that sure there's, you know, all sorts of concerns about whether this is going to lead to human extinction and so forth, and the tr trust is supposed to prevent that. But this is still a business with competitors and assets that they've acquired and investments they've made and cash flow that they need to monitor and taxes they need to pay and people they need to hire. Okay. And a CEO they need to supervise. Okay. And Yeah. Yeah. the idea that not just the public investors, but how about all these founders? I mean Google and Vidia and all those others are happy to, you know, cash in at this IPO because they're doing gonna do really well. Okay? But long term, this is this is this is crazy to try to think in terms of an investment case. This is this is not only is it opaque and ki I it's Not opaque. It's if you take a lot of time as you have to understand you can sort of figure out who's in charge. But once you find out, you know, the idea that, you know, they have to manage a business too to kind of make some money is seems to Yeah, well I mean that's why I'm I'm distrustful of the I mean, that's why I'm distrustful of this mission thing. I mean the fact is that they need a tremendous amount of capital. They're not gonna get it if they're not making money, and that is what's going to drive the decision making more so than what the L T B D says that doesn't, you know. Yeah. T B T. Right. All right. All right. So I'm not investing. Doesn't sound like you are either. But it'll be fun to sort of see how the IPO does this week. All right, let's talk for another few minutes about the other AI poster child, NVIDIA. we could talk we could devote two or three shows to what's happened to NVIDIA in the past two, three weeks. and we probably should. there's CorpGov failures, the board of directors oversight of Ha ha ha. Yeah. Mm. Jensen Wong. There's a possible all sorts of delicious securities fraud theories. all sorts of good stuff. But let's focus on one little aspect that was caught my eye. I don't know if you saw it, in the news. There was all sorts of chatter about how NVIDIA did not have directors and officers liability insurance. And that got th that was Yeah. Yeah. Yes. The board was like being accused of negligence, of of of mismanaging this little part of its business. And once I kinda looked into it a little bit, it occurred to me right away, and I think maybe you too, that their decision, which we'll explore in a moment, was quite rational. Yeah, I mean I was very surprised, but I you're you're exactly I I became convinced that this is not an un an usual dis a rational decision. Right. S so s and and and and I don't know if you're if you're if you recall this, but I actually early in my career going back a I had a lot of experience with the subject of property casualty insurance and yeah, I was I was an actuary for a week and worked in in risk management and property casualty corporate stuff for first ten or more years of my career. y yeah you were you were in you were in insurance, right? Or something? wow. and then for various reasons we can talk about offline. I didn't that didn't interest me as much. So but anyway, this is something I've I've known a lot about. And also on my service on boards I've had to learn about this. So I'm I'm I'm relatively familiar with what's going on. So what NVIDIA did is they they did buy a little bit a little bit, a little slice of directors and officers liability insurance, but it covers only the directors directly. to use a phrase. and it doesn't provide any coverage for the company. And we'll explore that in a minute. And so and it makes good sense in that way. So let's go back to first principles for a minute and you can help me with this. Is that if the problem here is that if a director gets sued for something, okay, it could be some sort of direct, you know, thing where a director insulted a vendor and the vendor sues them in their capacity as a director for defamation. I don't know. Mm. Okay, or they made a decision that screwed over some vendor. Or more generally or more frequently, directors make a decision and then a derivative suit is a security is fraud, whatever. There's all sorts of things that directors can get sued for. Of course in the US everybody sues ev anybody for everything. So there's all sorts of possibilities. So okay, fine. and of course there's two potential costs. Or or even a securities fraud lawsuit. Right. Well that's 'cause we don't have regulators, but yeah. that a given director can bear. First is obviously damages. You know, maybe, you know, the share price behaved in a certain way that, you know, you gotta pay that. And you know, you don't want to pay that out of pocket. But there's also the legal fees and other expenses, but mostly legal fees. Because defending yourself here is expensive at you know, twenty five hundred dollars an hour for really good lawyers from you know, from Wachtel or from Sullivan. so what normally happens, your first recourse, if you're one of those directors that gets sued is the company on whose board you're sitting on, right? That's where all this corporate indemnification principle comes from. Is that I I there's with very few exceptions, corporations are allowed to cover to indemnify Yeah, but they can't they can indemnify good faith actions and in Delaware if it's adjudicated liability to the company itself, the company can't cover that. But yeah. Right. Right. Sure. Right. Right. So so but but there's and and you know, most company bylaws most company most state laws allow it. Most company bylaws then implement it. And frequently there's separate indemnification agreements that just, you know, add a layer of paperwork, a layer of that say, okay, we're going to cover so you you you get your legal bills and you hand them to the CFO and you say, Please pay. You get the damages judgment and they you're supposed to pay that and the company pays it. That's the but sometimes, as you started saying, companies can't or won't or don't pay it. sometimes th you know, things aren't in good faith. Okay. Sometimes there's an ad you know, there's a adjudicated liability, like you said, in a derivative action is in Delaware. Or maybe the company kind of disputes it. Maybe you're sideways with the company and they're saying we're not paying your legal bills, forget it. And you don't want to sue for that. Or maybe the company just doesn't have the money. Maybe the company's like, you know, n teetering toward bankruptcy. Okay. What does a poor director do then? And any ideas? They get insurance. They they buy ins they buy insurance. Okay. So so the DNO yeah, yeah, they pay the premium. Right. Okay. But so the insurance kind of backstops the indemnification. Okay. And the first part of this is they backstop the indemnification if the company doesn't pay. and that tends to be called side A. But the company pays for the insurance. coverage sort of as a it's a description of one of the endorsements or something that the policies provide for. and then so that's usually what you do. If the company isn't there or goes bankrupt or goes out of business, the side A coverage helps keep these directors whole. And it's usually the side A coverage that the directors are really interested in. And that's what NVIDIA has. Nvidia has side A coverage with ample limits in case NVIDIA can't pay or chooses not to pay because inevitably they're getting sued over their inability to supervise Jensen Wong at this point. Okay. But so what were the observers kind of worried about, I think? Well, there's two other parts of director's and officers' liability insurance. One of them is when the company actually pays the indemnification or indemnifies or pays the legal fees on behalf, they say, Okay, cool, yeah, you those are the bills are good. We understand we're gonna pay you. Then the insur then the company can go back to the insurance company and say, please reimburse us for this. Okay, that's sci that's called Right. So yeah, so insurance for the expenses the company puts out to defend and pay damages for the to the extent they identify their own officers. Yes. Right. And that's called side B. All right. I don't know where the side word came from, but side. And then finally, again, in these litigation, you know, resourceful plaintiffs are gonna name everybody. They're gonna name all the directors, they're gonna name all the officers, and they're also gonna name the company. Right, okay. Right. So occasionally NVIDIA will be sued directly. next. Well, it de ba in a securities case if it's not a derivative action. So this is the distinction between securities case and derivative action. Not the directors themselves, but in Vedi will be sued for allegations that are similar to what are covered under DNO policies. And you know, they may have to lay out legal fees and they may have to pay damages. And there's insurance for that too. That's called side C. Okay. Now, what's and none of this is this is all you know, this isn't anything you find in securities finding. This is just kind of what you can learn from the news reports. Evidently, NVIDIA bought only side A to protect the directors if they can't. And that's something again that directors are just going to insist on. They're going to say, I have to have this. Okay. They did not buy B or C. And that's what people were kind of all worried about. And the problem is, is NVIDIA is a very wealthy company. Okay? There is absolutely no need for them to buy side B or B. Mm-hmm. Yeah. Or see any of the corporate coverage. They should be self insuring all this stuff. Okay. So it was, you know, the the the premium that they'd have to pay is is hardly worth it, you know, give giv given their financial resources. You know, this is, you know, they get trillions of dollars of market cap, billions of dollars of cash. They they could they could pay any amount of damages that they have to reimburse, that they have to pay themselves. So it it makes perfect sense. Mm, okay. I said. for them for them not to not to do this. So I was I Yeah, I I actually I heard that it's you know, the ri like they can make more money by investing in their business that what they would otherwise pay in the premium. So it you know Mm-hmm. Right. Yeah, the premium for this would be, you know, tens or, you know, multiple tens of millions of dollars. And they don't need to spend that. They could just, you know, given the risks, you know, they could they could bear this easily themselves. So Yeah. But but as long as we're talking about insurance and indemnification, I do want to throw one thing out there, which is, you know, the last couple days have been all about like other tech people agitating to reinstate wrong. Yeah. And you know, you can I don't know if that's gonna happen. I'm not gonna make a prediction. The board has said the board has said they won't, who knows? But Yes, go ahead. Yeah, yeah. man. Well, the company has disavowed any of this. Right. you know, they're clearly trying to replay the open AI thing where the board fired Sam Altman. Yeah. Or yeah, but here yeah, well he's he he apparently he's helping to organize it. There are employees who are rebelling just as they did in open AI. He's got all the tech people and VC people who are in his corner, which reads very much like I mean to some extent that they want to establish a precedent of impunity. But but the this is the difference. Yeah. Yeah, yeah. But here's the difference. Right. Or Huang is trying to do that, f assuming he's part of this, yeah. They got this petition, there's a big petition with a hundred thousand signatures. The difference is that obviously this is not just about affairs. There are very real concerns of potential like law breaking, evading export controls. Jeffrey Sonnenfeld had a thing that there might be accounting problems in terms of h misleading the public as to how much they were dependent on China. And I have absolutely no idea whether any of that is real or gonna turn into a something. But there's definitely right now on the table at least a question. Yep. Of whether Nvidia went more than simply, you know, finding loopholes and actually was evading export controls by sending to like sending chips to I don't know, straw companies whatever. Right. This Sing Singapore to go sent to China to use against us in our yeah. It if if this is something if this was something that Huang was doing and the board, having suspected that, hires him back, now the board is those individual directors are directly complicit in the law breaking. And that's gonna look like bad faith. It's going to look like intentional illegality. And that's where not only is there liability, but that's where indemnification and insurance may fall off the table. So Yeah. Right, yeah. No no no no DNO shurer that's on the side A of this is going to easily is gonna say, Okay, sure, directors, you took this guy back, even though you yeah, th th th th that coverage dispute will be epic. And they will not pay it. Go ahead. Yeah. Yeah. So that's exactly that I mean, and and and the company actually, if they're adjudicated liable to the company themselves for causing it to break the law, they don't get indemnification. And damnification is only available for good faith cover. I mean, like the level of like personal jeopardy they're putting their own finances, leaving aside whatever other penalties might exist. I mean, is quite real. So This is what like I'm again, I'm not making a prediction of what they're going to do. I'm simply saying this is what the difficulty is that is not that was not present in the open AI situation. Yes, correct. There's this th there there is a a much more intricate set of facts here that could lead to immense amounts of liability for for some of the parties, including, you know, NVIDIA itself. And again, there's you know, if if if some of this is right, part of the problem is NVIDIA is in the middle of they're called the central banker of AI. If they're in the middle of all this and their balance sheet takes an enormous hit and they can't make good on some of their obligations to fund data centers and fund businesses and so forth. they're gonna they're gonna this is this is one reason why I'm not investing in this sector. So anyway. All right. so that's NVIDIA, why don't we shall we take a quick break and then we can come back and talk a little more about AI like in activism? All right, let's do that. we'll we'll do that in a moment here at shareholder primacy. Right. Sure. Yep, okay. And Matt, you just gotta remember to put the transition in here and the music. So all right. And this one should actually be I've thought about it some. This is the one where I was gonna try to r I know. This is something where I've been trying to write about it and I'm like I'm not doing well, but maybe this will help me work out some of the thinking. All right. let's go. Three, two, one. Over to you, Anne. 'cause I haven't Yeah. Welcome back to Shoulder Primacy. I'm Ann Lipton here with Mike Levin. And we are talking about AI, but we're moving on about AI generally. Yes? Do you have some thoughts about activism and AI? Yes, well yes. I I've been I had some ideas I wanted to kind of run past someone and and you're as good as any, about about AI in the activism world. and my overarching observation is that everybody needs to kind of calm down. Hammer someone. Ha ha ha. that there's that there's a lot of attention being given to it and there's a lot of worry. you know, kind of going back to is it gonna, you know, what's it gonna do to activist investing? What's it gonna do to investing and so on and so forth. and my hypothesis is that consistent with the idea that it's not gonna lead to ruin or paradise, and also it's gonna happen much more slowly than we think, that the idea of AI having applications or AI being a subject of activism or so forth is probably a little over baked right now. And and people just need to kind of relax and focus on other stuff. And harder that is to do in this environment where everybody it's all we talk about. you know, I don't think a whole lot about AI and, you know, portfolio companies and I don't think other people really should either. So s Mm-hmm. So I thought about four different kind of aspects of this. Okay. First, there are a number and I think a growing number of call it precatory proposals that ESG proponents have drafted and moved and proposed at various companies kind of about AI. so and it has to do more with kind of responsible application. So here here's one at Amazon was and these are all from the proxy preview report. So there's looks like there's ten or twelve of these. So Amazon had a proposal from last year, alignment of AI sales and development with responsible AI. Okay. NVIDIA, someone proposed one a rationale for engaging in military sales. And NVIDIA of course makes AI chips. Palenter, the defense contractor in the data company, had you know, human rights impact assessment of AI. Salesforce had an AI-related human rights risk. And so there's a whole, there's a number of these proposals that seem to focus on how companies kind of use or apply or whatever AI, and the proponents are advocating for, I don't know, humane or fair or somehow measured use, you know, report about how we're using, I don't know. And it str it strikes me that these are, you know, pretty esoteric. And you know, I'm sure there's probably a little bit of basis for one or another of these. But it didn't seem to me, you know, and again, setting aside whether any precatory proposals are gonna get voted on any time ever, in the next, you know, you year two, three or whatever. it it it didn't seem that these were as as nearly as hard hitting or as I don't know, relevant as as some of the others. That i it looked like there was some proponents that were looking to kind of jump on the AI bandwagon rather than to, you know, focus on, you know, like, you know, proposals at the newly public anthropic about how they're using it. So I I wasn't sure about, you know, precatory proposals related to AI and whether they were Mm. you know, need to be as much a p priority even as like some of the climate stuff that we that we've seen. so so that was the first one of these. The second was a cousin of that was y you know, th there there's there's sort of growing noise about whether AI should be the subject of like a financial activist engagement. Again, it's a cousin of this, but, you know, r related to how does a company govern its AI? You know, w is are they appropriately using AI in their business? Are they using it and I think these focus more on should a company be doing more with you know portfolio company, doing more with AI rather than with the shareholder proposals in the first example with which is companies are, you know, being a little too aggressive or being careless with AI. And as a subject of an activist project, of an activist engagement, financial activist engagement, strikes me as way premature. I mean, we might as well be asking, you know, what role if we ask what role does AI have in a company's business, we should be asking what role do computers have in their business? What role does A does the internet have in their business? Go ahead. Hmm. Right. Yeah. Yeah. S so yeah, I mean, usually I mean, well, we know the standard there's a fairly standard act financial activist playbook. I'm sure everybody has different strategies and there are different kinds of activists, but a lot of it is take on more debt, you know, sell some assets, return cash. Yeah, so I mean so tell me something. Like, I mean, you've said before that sometimes, especially if you're a new activist, you wanna like make a splash and and kind of get attention. Return some cash, restructure the balance sheet. Right. There's there's kind of fundamental business imperatives, yeah. So my question is, if is activism aimed at like you should use more AI, that's my pitch, is that like more of a public relations camp at other shareholders and the business model than an actual activism engagement? Yeah. agree completely. That that any activist that I know that starts, you know, critiquing use of AI at a company is is got their head up their tush. That's not it's not it's not that's not really that it's not a really viable but importantly it's not a viable path to like imp improving share price in the sh short to medium term term or even the long term. Ha ha ha. Ha ha! And any any company that, you know, if we're if I'm looking at like a software company, any company that I'm gonna have in the portfolio, like like I said, a software company, that I have to talk to them about AI, I mean, that's crazy. You know, they're obviously thinking about this. You know, and I c I may kinda quibble with them a little bit about, you know you know, operationally, how they're working, but again, I'm not gonna you th that that gets way into the weeds about how an activist focuses on a company's operations. so anyway, so that's another example of where I think, you know, focusing on AI and whether boards are you know, whether I'm advocating for a board to focus more on AI, there's plenty of boards, there's lots of NACD stuff about how boards should be looking at AI or something, good for them. But I'm not gonna make that a subject of an activist situation. Okay. Third area has to do with how AI is gonna change like activism, more like in how We work. Okay? You know, just like how AI is changing how you do legal research a little bit, I suspect. fine, okay. Yeah, there's other people or about how you, you know, find you know, cases or whatever, that kind of research. It is changing, of course, how Mm. Well, it's not it's not changing how I do it, but it's changing how other people do it. activist work, you know, it's it's making our life easier. We're gathering facts better and analyzing companies more quickly, even helping to set engagement strategies. But that's no different than again, using spreadsheets or using Bloomberg or something. Okay. It's just it's another tech tool. It's a really good tech tool. It's working quickly, okay. But it's not it's not really a game changer. I mean it's game changer in a sense it's it's taking me a lot less time to do certain things now. But it's not changing how activism works. It's still, you know, a lot of the nuts and bolts of financial activism is interacting with boards and people and changing their decision. And I can't, you know, I'm not gonna have an AI, you know, bot do my activism. It's gonna always be kind of me and my team. So the last area I think is one where at least certain people should be a little worried. Okay, and that has to do with like stewardship. Okay. I do think that AI poses sort of Yeah. on the longer term, sort of an existential threat like to proxy advisors. Okay. To ISS and Glass Lewis and others that make a living by telling investors kind of h what to think and how to analyze and how to vote at companies. Okay. And here I think, you know, i i it's almost a perfect application of AI. Mm-hmm. Because it can read like, you know, decades of documents from a company and create whatever peer groups you need and look at and write policies for you that you need to review a little more carefully and and even start advising on votes. Okay. You know, th this whole the whole the whole process of delivering a t you know, how how a shareholder a passive shareholder should think about how to vote is is ripe for this kind of automation, I I suspect. Right. Yeah, I I mean I just think it's funny because I think JP Morgan was one of the ones that announced it was gonna start using AI for this. And but part of that, as I understand it, is JP Morgan and Jamie Damon being upset that sometimes proxy advisors recommend against his pay package. Except that I think yeah, except that I think that, you know, what we've seen so far, I think, about AI analysis of proxy voting is they actually vote against management more than the proxy advisors recommend. Yeah. Right. There's a bit of a grudge, yeah. Yes, I've seen that. Right. Yeah. The A the AI AI engines are a little more sympathetic to shareholder concerns than than the than the people at right. Yeah. I mean I just yeah, I I guess I just don't know enough. I I don't know if a enough about AI works. Well the my my worry about these kinds of things is how much AI becomes self-fulfilling. Like, you know, if it it creates the data for doing this and then uses that data as the as the input for learning more, like, you know, at what point is this becoming like just the AI chugging along with its own analyses that isn't adjust that, you know, is is just Yeah. Well, but that's a global question for, you know, any application of AI and anything. So but I but I think there's an interesting a useful analogy here because so much you know, everyone complains complains, some people like it, I think you do, about indexing and algorithmic investing. Where, you know, it becomes so cheap Isn't it just it? Yes. Yeah, that it it'll it'll make that n there's no more market efficiency because it's just indexes, yes. Right. Okay. Well, you know, there's a lot of investing, it's become just you know, you know, giving it over to algorithms and so forth. And I think the analog analogy here is that may we might be giving over governance or at least the voting and so forth of passive funds i becoming a lot more algorithmic. Okay. That there's th that that's a possible direction like stewardship could go. Yeah, it's this, yeah. So, you know, big stewardship groups and they're not that big. You know, we've talked about stewardship groups where they do a valuable job, but you know, they're not a they're not you know, BlackRock doesn't view their stewardship group as a moneymaker. it's a way to burnish their reputation with certain types of allocators and it's a compliance function and it's it helps with their you know, being responsible, you know, participants in capitalism. Mm. It's a compliance function. But again, it's not and so so if they could spend a lot less on stewardship and give it over to algorithms the way they have with a lot of their investment decisions, I'm I'm not sure they'd be really disappointed. And so so that that's that's kinda the direction, a potential direction that this could go is making stewardship algorithmic and automatic. just the same way that investing has become for a lot of these big funds. So anyway. So those are the four things. So like I said, I think everybody needs to kind of relax about AI. Let's focus on some other stuff. There's a lot of other imperatives going on. Unless you're an investor in NVIDIA or unless you're trying to get the anthropic IPO, then it's a little more immediate. except if you're a proxy advisor or work in a stewardship shop. Maybe you gotta worry a little more. So that was that was kind of what I was thinking ab about this. And hopefully you give us something more to talk about. All right. So that was that that was the AI show all right. I think we've I think we've covered a lot. Let's let's return to this at some time soon after after we figure out what happens in NVIDIA and after the anthropic IPO. If we're st I think we will be. Fair enough. Okay. Yes. If we're still alive. This is Shareholder Primacy, hosted by Ann Lipton and me, Mike Levin. I'm an independent activist investor and advisor to investors about their activist situations. Anne is the professor of law and Lawrence W. Demuth, Chair of Business Law at the University of Colorado Law School. You can find me, Mike, at the activistinvestor.com, and you can find Anne at law.colorado.edu. Shareholder primacy is produced and distributed by the good folks at FreeFloat Media. Thanks for listening. We'll talk again soon.