00:00Can I just clear this up here? Martin, my co-anchor and I, Matt, were having this debate as to
00:03whether SuperReturn is the biggest private equity conference.
00:05It is, right? Like, there's no other gathering of this size of people.
00:08I believe so. It's huge.
00:10Yeah, it's definitely. I mean, how's the mood been, by the way? Just in your meetings with people, with all
00:14your LPs, with everyone else.
00:15What sort of vibes are you getting at the moment?
00:17I would say the mood is definitely a little bit subdued. One of the topics is the lack of exits.
00:23That, in particular, in the mid-market is an issue. Frankly, a lot of those assets have only private equity
00:30as a buyer.
00:30And there's not enough dry powder out there. But for us, it's different.
00:35We focus on larger companies, high-quality companies. Those have strategic exit optionality.
00:41They have typically an IPO exit. We actually returned $14 billion since January 2025.
00:48Did a lot of IPOs. Just did the liftoff IPO last week, which has been a great success.
00:52So, our strategy, our exit engine is working.
00:55Can we talk about liftoff? Matt and I spoke to your CEO, who is very excited, obviously.
01:00It got a nice first-day trading pop. Where do you think this ranks in the success of Blackstone Exits?
01:05Well, it's certainly a very good one for us. And this one was personal for me.
01:09Thank you for having Jeremy on the show.
01:12Jeremy is one of the very best CEOs that I've ever worked with.
01:15And we tried to go February last year, and it was a brutal timing.
01:19It was the week of SaaSpocalypse.
01:21So, we had to pull it, but we backed the management team.
01:25The company kept on performing, and we went in June.
01:29Frankly, the banks advised us not to go, and we ignored that advice.
01:33It was the right call.
01:34We had one of the shortest, I think the shortest TMT Roadshow in six years.
01:38We were 11 times oversubscribed.
01:40We priced above the range.
01:42We made 10 times our money.
01:44So, a fantastic return for our investors.
01:47And it's our third IPO this year.
01:49And they come on the back of Medline, where we made over $4 billion for our investors in 2025 in
01:55September.
01:56Then Legions, which has traded up over, I think, 150%.
01:59We have seven more on file.
02:02So, the year of the IPO is definitely on.
02:04So, are you just done listening to your bankers, then, who are saying, don't IPO in the shadow of these
02:08bigger ones?
02:09Or do you think there is some nerves around SpaceX, OpenAI, and Anthropik as we go on?
02:14Well, I think if you look at something like a SpaceX coming at $75 billion, that certainly, I can see
02:20how that will feel daunting to people.
02:23But compare it to Blackstone Q1, we raised $70 billion.
02:27So, as long as the performance is there, I think there is demand for IPOs.
02:33And Liftoff shows that.
02:34And our oversubscribed, generally oversubscribed IPOs show that.
02:38It's always fun to look at what Blackstone businesses are coming to market,
02:41because you have a very digital age, Liftoff, an app company.
02:45And then there's always hopes that maybe we'll get Jersey Mike's coming public.
02:49Might that be one of the seven that we can see this year?
02:52We love Jersey Mike's.
02:53It's a fantastic company.
02:55I love eating the subs.
02:57So, hopefully, you know, it's interesting.
02:59Are you contractually obligated to eat the subs, Martin?
03:01I will eat the subs every day.
03:03Nothing is better.
03:04The Italian is my favorite.
03:06So, I mean, that's really interesting because, again, it's two different types.
03:09Do you think that that's the way this market is going to be bifurcated?
03:12Tech, really successful, maybe even AI-adjacent companies and real-world companies.
03:18And everything in between is kind of iced out.
03:20I would say there's truth to that.
03:22The demand for AI is very, very strong.
03:25And I think defensive assets get bid up.
03:28If you are in the middle, I think it depends on modes and it depends on growth.
03:32So, if you take Liftoff as an example, it's had 30% growth.
03:35And so, it's neither really an LLM nor a purely defensive play, but because the return
03:40potential is there, the investor demand is there.
03:43And that is a proof of generally how our IPOs have performed.
03:47So, just to be clear, you don't subscribe to this model that because SpaceX, OpenAI,
03:52and Anthropoc are all competing for the same institutional capital, that it dilutes the
03:56supply out there for you, or the demand, rather, that there's too much supply.
04:00So, that doesn't really concern you, this sort of dynamic that's emerging.
04:03Well, I won't comment on anybody else's IPO, but I will say as long as we have returns,
04:09investor demand, in my view, is there.
04:11One of the other interesting deals you did recently was this AI services company with
04:17Anthropoc.
04:17And then alongside that, you also committed $5 billion to an infrastructure JV with Google
04:23around TPUs.
04:24What's the TPU strategy specifically?
04:27Why enter this deal with them?
04:28Yeah.
04:28So, maybe let me start by explaining what a TPU is.
04:31The TPU is an AI chip.
04:34It's used to train and run AI.
04:37Gemini was trained on TPUs.
04:40Cloud Opus 4.7 was trained on TPUs.
04:44Sorry.
04:45It's a very strong chip.
04:46So, we committed $5 billion, 500 megawatts, to a partnership with Google to build a NeoCloud.
04:53And you might ask, what's the rationale?
04:55Why now?
04:56And that is increasing demand, in particular if the agentic economy takes off.
05:02And at the same time, the physical supply is constrained.
05:06It's hard to get a data center permitted.
05:09It's hard to get the power.
05:10It's hard to get the chips.
05:12So, as a result of that, the capacity is very valuable.
05:16And we are proud to be partnered with Google on this.
05:18Does that mean your portfolio companies get advanced access or they're prioritized for access for some of the infrastructure there?
05:25I would say this infrastructure is more typical for large AI companies, training models.
05:32But the anthropic partnership that you mentioned is one where we're excited about making great engineering talent available to our
05:41portfolio companies.
05:42We're close to the anthropic management team.
05:44The CFO, Krishna Rao, actually used to work in Blackstone Private Equity.
05:48The head of strategy was an analyst in our group.
05:51The head of capital markets came from Blackstone.
05:54So, we have deep trust.
05:56And what this anthropic partnership is about is making real engineering talent available to the companies that work with the
06:03anthropic partnership.
06:04So, we basically have Blackstone to thank for anthropic and thought is what I'm hearing from you.
06:08I think they are just fine without us.
06:10But we are proud to have partnerships.
06:12How are you approaching just the cost of this stuff?
06:15These stories keep cropping up of a consultant hears some story about a client that accidentally spent half a billion
06:21dollars in just the span of a month
06:22because they weren't tracking their token budget or Uber blowing through their entire year's budget in just a couple of
06:28months.
06:28As you're making companies more equipped with AI, how are you handling just the pure inference and compute costs that
06:35seem to be growing?
06:36Yeah, I would say those headlines about crazy spend, those are lack of controls.
06:42They're growing pains early in the AI revolution.
06:47I think it's more important to focus on the longer term.
06:51And longer term, what I think will happen is the demand will increase.
06:55It will likely exceed the supply.
06:57And when that happens, the value created will set the price.
07:02And there will be a paradigm shift.
07:03And that could very well lead to an increase in the token cost.
07:07We've actually seen that already.
07:09And it will particularly benefit the frontier models.
07:12That seems not great for margins if the whole project of this thing is also to make companies more efficient
07:17and decrease margins.
07:19I think the bidders on this will bid to make a return on any token spent or anybody who procures
07:27it.
07:27And for the TPU partnership that we mentioned, which will have compute capacity, that will actually be a positive.
Comments