0:00
You're watching Christmas TV.
Search this video
0:04
Uh, so, uh, >> yeah, let's get into this.
0:05
Let's get into this piece.
0:06
Uh, uh, Jeremy Stern profiled Josh Kushner in Colossus. It released today. It's not in print yet.
0:15
>> It's going to be in print, right? >> Get ready.
0:17
>> Um, called The New World.
0:22
>> Where do you want to start?
0:23
>> Where should we start?
0:23
I mean, there's there's so much in here.
0:25
>> I I like this bit about Spotify.
0:25
So, uh, in 2012, uh, Thrive invested $6 million in a growth round of the Stockholm-based Spotify out of a $150 million fund, an allocation Kushner had regarded as a favor until learning nearly a decade later that Spotify CEO Daniel Ek was in need of exactly 6 million to close the round.
0:47
So, he's just like he's just like, "Oh, thank you so much for like making room for me. Like, this is amazing.
0:53
being like, "I'm so glad I could like, you know, get my 6 million in."
0:57
And and Daniel X over there being like, "The crowd would have completely flopped if you hadn't come in." So, thank you so much.
1:03
But, uh, there's more to this anecdote.
1:05
Because the reason EC knew he liked Kushner was that a few years earlier when Spotify was only available in Europe, E was notified by an executive that someone in the US had manager managed to register a fake UK address in order to download the app via the UK app store.
1:22
They discovered it was Kushner sitting in the library of HBS. >> Wow.
1:28
>> It's a uh >> I pulled out some highlights from the piece.
1:31
It's way too long for us to try to to get through entirely.
1:33
Um, but the uh some background uh on November 17th, 2023, Kushner was a 38-year-old spouse of a supermodel, brother and in-law of American political royalty and founder and CEO of what would what had very suddenly become one of the most coveted venture capital firms in the world.
1:51
firms in the world. He was also the grandson of survivors of the Novo Gruda ghetto massacres, indignant uh in indigent refugees who over the course of the Cold War built a New Jersey real estate principality that their son Josh's father expanded into a
2:06
multi-state empire before his conviction on felony charges and sent sentencing to federal prison and before the White House activities of his older brother Jared put Josh in the crosshairs of a torid political convulsion of which he wanted no part. So, um, So, um, >> kid spirits. >> Yeah.
2:24
I mean, ultimately it's it's it's this uh Josh has put on an absolute master class of how to be an icon while being relatively in the shadows, right?
2:38
He just sort of like pops his head up in these key moments.
2:43
Um, and uh, yeah, I think when I think about what Thrive what Josh and the and the Thrive team has built, it's actually it's I mean it's just like it's almost it's almost unbelievable >> like the story the story end to end.
2:59
Um, by the fall of of 2023, Thrive Capital, the New York-based investment firm Kushner started 13 years earlier, had become an overnight sensation.
3:09
In 2010, Thrive's first fund was 5 million and included companies like Kickstarter and Group Me.
3:15
By 2023, its eighth fund was 3.
3:18
3 billion, including it maniacally concentrated $2 billion investment in Stripe at a $50 billion valuation.
3:24
He's like, here's twothirds of my fund.
3:29
>> This is the Fifth Avenue strategy. Buy Fifth Avenue.
3:31
He talked about this on Invest Like the Best >> and a $150 million check into OpenAI at a $29 billion valuation.
3:36
The companies are now valued at 107 and 500 billion respectively.
3:41
Along the way, Thriv's bet on Instagram, Spotify, Warby Parker, Skims, GitHub, Slack, Robin Hood, and other companies had become uh conspicuous for being preient, aesthetic, and exquisitly timed among its most vindicated admirers, and for being absurdly priced, momentum chasing, and too highly concentrated in dysfunctional businesses with unproven returns among increasingly sheepish critics.
4:06
So, um, >> I like that they got Johnny IV to give a quote for this piece.
4:10
Some people just have innately wonderful taste and intuition, uh, said Johnny IV, Apple's legendary former chief design officer, who's now working with OpenAI on a hardware device.
4:21
Josh has wonderful taste.
4:23
I think it bleeds into his product intuition, which is just fabulous. >> It's great.
4:27
>> It's great. Um another another highlight here uh in 2020 in 2010 moreover the unknown Kushner's brother unknown little firm was making a bunch of large but weird sounding claims for itself like that it was a stage geography and sector agnostic venture firm that would concentrate all its investments in a
4:46
very small number of companies that it was not only an investment firm but also itself a company that incubated its own companies as well as invested in others and that it didn't just invest and incubate but functioned as a service prov provider, product creator, and embedded operational commando unit for founders. By 2023, every self-respecting
5:01
By 2023, every self-respecting investor on Sand Hill Road was also saying such things about themselves, even as they wondered how a New York firm made up of a handful of kids in their 20s and 30s, many of them with zero experience in venture capital.
5:13
And from the technology Bermuda Triangle of New Jersey had become some of the most desired investors in tech and the ones most closely associated with the otherwise distinctly West Coast boom in AI.
5:27
Um uh of course uh Josh had a very untraditional pathway to venture capital.
5:34
He w to Harvard for undergrad while uh at the time uh Mark Zuckerberg had just dropped out of Facebook.
5:40
Later he uh got a job uh at Goldman Sachs buying distressed debt and then went back to Harvard for his MBA.
5:48
Uh, and uh, of course I'm joking a little bit, but it but it's it's um, the what I what I appreciate about >> Yeah.
6:00
>> What I appreciate about Josh is uh, in many ways he he was on uh, he was on the perfect trajectory to be a venture capitalist.
6:10
And yet he has succeeded beyond what he has succeeded in a way that so many other thousands of people that had the same kind of pathway into the industry by being in the right circles like being early to a number of these these different trends.
6:28
Like he's succeeded on a hund time on a a scale 100 times greater than um many of the other people that were again had the same kind of like pathway in the industry.
6:40
Yeah, my read on it is that uh a lot of the VCs that started in the same like vintage that did not become massive institutions uh maybe just got caught in the trap of diversification.
6:54
Like I keep going back to that story of Josh saying he wants to buy Fifth Avenue, buy the best asset in the class, be concentrated, build a big allocation.
7:04
And that takes guts that I think a lot of VCs kind of fell in the trap of like I need to get a bunch of logos from I need from a whole bunch of companies or it's particularly cool like it's actually higher status to be oh first
7:20
check in this company or see even if you have a tiny allocation that gets completely diluted down like you're not actually making as much as many dollars like if you put two billion in stripe at at at 50 and it goes up to a hundred you made$2 billion dollars right? Versus you
7:32
made$2 billion dollars right? Versus you put 200k into some company gets diluted down you make 50 million like you made way less dollars on dollars return um but it's like somehow higher status in venture to be like yeah I was like the first check in and I think uh Josh's been early in a lot of companies but I think that he understands the importance
7:55
of like portfolio concentration actually getting all the dollars in the companies that matter and that's like And I think early on I think my my sense is that there was some like hunting for logos early on, but they were when the fund sizes were very small >> and that led to him having the the track record and ability to put $2 billion into Stripe out of a three $3.3 billion 3 billion fund.
8:20
>> And also like we're just like like the Thrive era, the era that Thrive grew in is an era where venture changed pretty dramatically.
8:29
I mean, we were talking about the Intel IPO, raise $6 million.
8:32
Like, what is your role as a venture capitalist?
8:34
Like, deploy $100,000 and then wait for the IPO like a year later.
8:39
And it was like that in the dotcom boom and Google IPOed pretty early.
8:43
Even Facebook IPOed at what, 50 billion, and it was like the biggest IPO of all time.
8:48
Uh, it was in like the >> Yeah, there were there were investments in that era that I'm sure the partners were underwriting.
8:53
It is I think there's 70% chance this company IPOs in the next two years.
8:58
this is why we're investing and then they became compounders and over time it looked like >> but they distributed and so a lot of the funds became raas but uh Thrive has been able to say yes we're we're investing in a company at 50 billion because we expect it to grow and grow and grow uh and actually deliver like a venture style return or like a significant return.
9:19
I appreciated this exchange between uh uh part of the article, the beginning goes into Kushner visiting um Rick Rubin in Malibu and uh Josh was telling Rick, "My deepest insecurities that I have these intuitions about things that I cannot explain to anyone," Kushner told Ruben as they sat in his garden overlooking the ocean.
9:37
"Sometimes I see or experience something and it makes sense to me.
9:40
I fall in love, but I cannot explain why.
9:42
Like when Thrive invested in Instagram or Spotify or OpenAI, I could not explain to anyone why the products made sense to me.
9:47
It is my job to learn as much as I can from my team and teach them as much as I can.
9:54
But often I have to push forward on my intuition alone.
9:56
Which is why my even deeper insecurity is what if I lose it?
10:00
Like what if I lose the capacity to feel or experience these things?
10:02
Uh he's basically the uh having this exchange because of course um Rick Rubin notoriously is just going off of raw intuition and uh and vibes.
10:14
Um, I think the other the other I mean it goes it goes into his entire um his family's crazy history dating back to >> Europe during World War II, >> getting out of Europe.
10:30
Um but the um I think uh can't be understated how formative it was I think for Josh to go through this this sort of uh you know right as he was uh the quote here by the time I was in high school my father had accomplished a tremendous amount.
10:46
He was deeply impactful in both the business and phil philanthropic worlds and then overnight our family were outcasts.
10:52
The world treat us treated us all one way for the beginning part of my childhood and then suddenly they treated us very differently.
10:59
That experience showed me how the world works and why you should not care too much about what people think.
11:04
think. Uh, of course, um, his father was embroiled, uh, in a, you know, wild crisis and, uh, ultimately went to prison, uh, for a couple years, but, um, I think, uh, Josh says elsewhere in the piece that he wouldn't, um, uh, I forget
11:21
the line exactly, but, something like wouldn't wish what he went through on his worst enemy, but at the same time wouldn't uh wouldn't uh uh is is sort of grateful for uh for what kind of turned him into a monster basically. Very kind monster. Very kind monster.
11:37
>> Well, if you want to try and reverse engineer Thrive's returns, dump all that data in Julius and chat with your data and get expert level insights, it's the AI data analyst that works for you.
11:47
Uh Tyler, what are you thinking about Kushner?
11:51
>> Uh I I think there's also underrated is he's he seems very um aura pill, right?
11:56
He has this kind of nonchalance about him.
11:58
Y there's a good quote I think that relates to this.
12:00
It says um >> uh it's about Andy Golden.
12:02
He's the um Princeton endowment head.
12:05
Uh he says Golden later recalled a happy hour for VCs in Cambridge in 2010 where he saw a 6'3 emo looking kid in a black cardigan standing apart from the group staring at the floor.
12:15
I mean that is just a pure aura. Yeah.
12:19
>> Uh apparently we're on French TV right now.
12:21
Thank you for the notification.
12:23
Send us the link if you can find it.
12:23
Uh would love to see that.
12:25
Um, yes, there seems to be somewhat of a of a cor correlation or inverse correlation between like uh just how much content and availability you like how how available you are in your aura.
12:38
Like Ilia like never does any press. >> Yeah. He's very mysterious. >> Very mysterious.
12:44
>> So because the mystery allows like people to like build this idea of you in >> Exactly.
12:49
You're like oh yeah like like even when Ilia posted yesterday like oh great the best day ever.
12:54
Everyone was like, "Clearly this is a like bubble is gone."
12:58
I mean, we're going to keep going. We're safe.
13:00
>> And and it's like that's not what it was about at all.
13:02
He was talking about geopolitics and the end of the Gaza war.
13:05
But um >> so Rim K in the chat says, "We asked for a link."
13:09
And Rim uh Rim K says France 2, which apparently >> you just have to go >> just go to France.
13:14
France 2 is the channel name. Yeah. No, I know. >> Find it.
13:18
Is there a way that or something? >> French cable here.
13:21
>> I would love to see >> go sign up try to sign up for an account. >> Did they translate?
13:24
Did they put subtitles over us?
13:25
I'm so I'm so curious about this.
13:27
Uh anyway, while we dig >> this is very interview. >> Yeah. Yeah.
13:31
But I'm wondering because they could have put French subtitles over us or they could have dubbed us with 11 Labs or something.
13:38
>> Ask le chat where how to how to get access to French.
13:43
>> There there is uh more details on how the OpenAI deal came together in this article that are pretty interesting which you should go read.
13:48
Uh I feel like I'm allowed to trash investors because I was one for most of my career.
13:52
Sam Alman said, "Most investors don't work that hard.
13:56
They're usually not available for midnight at calls and won't drop everything to fly across the country on short notice to do you a small favor for you the next day."
14:03
Josh is consistently willing to do all those things.
14:05
He's incredibly hardworking for his companies.
14:07
He'll do whatever it takes, any amount of time.
14:09
Nothing is too big an ask.
14:10
Uh during that crazy week where I got fired and rehired, he just put his entire life on hold.
14:14
Uh he didn't leave his hotel room for 72 hours.
14:17
He just worked non-stop, very strategically, very effectively to get things back on the rails.
14:22
Thrive had first gotten involved in OpenAI in early 2022 when they met Altman to discuss a new round.
14:28
They had been given a preview of GPT3, the foundation model that preceded chat GPT, which Kushner reportedly became so obsessed with, he almost seemed haunted by it.
14:35
Now, that's that needs to be like corrected because GPT3 came out in 2020.
14:40
So, I think they're either talking about GPT 3.
14:42
5, Da Vinci 002, which was really popular, or it's just a preview of ChachiBeept because investors were getting previews of ChachiBT beforehand and it was like blowing everyone's mind, but it was a really complicated deal and a lot of people passed basically for the wrong reasons in my opinion.
14:57
So, uh, >> on on Kushner's work ethic, there was a he was, uh, we met up earlier this year >> for coffee >> and he was he was uh I I had like gone to sleep.
15:10
He was still on the East Coast. It was like 2:00 a. m.
15:12
He's texting like, "Yeah, do you want to like uh I'll see you at like 10 or something like that."
15:20
I'm like, "You're on the East Coast. >> It's 2 a. m. for you.
15:22
you're going to get to the west coast and be ready for and and so >> absolutely >> that is part of the part of the story obviously is just insane you know insane work ethic. >> Yeah.
15:34
>> Uh I see in the chat somebody said uh we're on France too and Taylor says dang they released a sequel to France.
15:42
[Music] >> Tyler Tyler is MVP of the chat this week. >> Yeah.
15:46
Sorry Taylor >> Taylor's been crushing it.
15:49
>> Tyler has uh been struggling to learn French.
15:52
I I thought you studied French. You took French.
15:55
>> I took I took one or I took two semesters of French. Okay. I found it. Um Yeah.
15:58
Let me I'll screen share. >> Okay. Yeah. Yeah. Yeah. Let's figure out. I I'll keep reading.
16:02
Uh so, uh OpenAI at the time was a caprofit subsidiary controlled by a nonprofit board with a mission of safe AGI development, taking legal precedence over profits.
16:13
uh Microsoft's $1 billion 2019 investment gave it a dominant position in OpenAI with complex revenue sharing agreements and preferential access to the company's technology that and the company was reportedly valued at 29 billion while doing a trivial 50 million in revenue.
16:26
What a ramp from 50 mil to what are they doing now like 10 billion 12 20 billion something like that.
16:32
Uh it was all very weird.
16:32
There was no reason uh there there was a reason no other investor submitted a term sheet for that round. Wow.
16:38
After several conversations with the investment team, however, Kushner marshaled his case.
16:44
Forget the nonprofit structure and Microsoft and all the red flags.
16:48
He argued if you can create this much enterprise value, everything else is solvable.
16:51
In any case, colleagues recall he kept repeating things like I saw the future and this is the one. Great call.
16:59
Great, great, great call. Extremely hard.
17:02
I know some other investors who literally passed on that round because they were like, yeah, like we we we talked to our lawyers and they were like, "This doesn't make any sense."
17:09
So like we couldn't understand like how we were getting it.
17:11
Like there were lots of smart people that passed.
17:14
>> How early did people feel like Chad GBT was a going to be a meaningful threat to Google?
17:23
that probably started in like March of 23, but uh I but like November, I mean people when chat when when GPT3 came out in 2020, people were starting to say like you could use this to you could ask it a question.
17:41
You had to you really had to prompt hack because you couldn't just ask it a question.
17:46
You had to like ask it a question for a list of bullet points and put a couple a couple. >> I'm just laughing.
17:52
Rim K is saying it's live now on news. >> It's live now. >> It's on France, too. It's on the news. >> It's on the news.
18:01
>> Just we're trying to find it. >> It's on the news.
18:03
Um and so, uh Thrive submitted a term sheet 130 million from its main fund at the $29 billion valuation.
18:11
Uh in November of 2022, OpenAI launched JBT.
18:13
Within two months, it became the fastest growing consumer app in history.
18:16
By summer of 2023, Thrive was working on a $90 billion round.
18:21
And in August, they agreed to anchor uh 400 million of a $500 million employee tender offer, critical leverage in the explo explor exploding war for AI talent against OpenAI's publicly traded competitors.
18:33
Um by November 17th, 2023, Thrive had committed around $700 million to the company.
18:40
So, uh, pretty pretty remarkable results and, uh, really just goes to show you that like there are certain pitches in venture like seeing chat GBT before it launches that you need to really, really, really, really, really swing hard at and and Josh swung at the right pitch at the right time, which is like fantastic and much harder than I think people think.
19:03
so easy now that everyone uses CHP all the time to to to just be like, "Oh, yeah.
19:08
I would have I would have made that deal, too." >> Yeah.
19:11
And ignoring the ignoring the FUD and just like doubling, tripling, quadrupling down every single possible >> Yeah.
19:18
I mean, when when that when that initial round was happening, I was I was also like I mean, I wasn't an investor, but I was I was I was thinking like it just makes a lot of sense.
19:27
just makes a lot of sense. I was looking at the team and I was like if you just make the bet on like you have the CTO of stripe Greg Brockman you have the former president of Y Combinator just the resume of the founding team at that time
19:40
everyone was midcve mid-curve response was the corporate structure weird >> like like this isn't a nor this isn't a C corp it's breaking the rules I don't fund nonc corps like I'm out >> but just going back to basics of being like the founding team who's running the company are they the best? You know, is
19:57
You know, is this an important category and do I have the best team?
20:01
It's like it was hard to make an argument against it at that time in my opinion.
20:05
Um anyway, should have invested, I guess, but you know, like doing content instead. It's more fun.
20:11
Um in other news with OpenAI, um uh there's they did an interview with the uh Broadcom team and Sam Alman shared a little bit more about the custom chips that they're working on.
20:24
Uh before we tell you about that, let me tell you about fall, the generative media platform for developers.
20:30
The world's best generative image, video, and audio models all in one place.
20:35
Develop and fine-tune models with serverless GPUs on demand clusters >> used by Adobe, Shopify, Canva, Corora, and many more.
20:42
>> So, uh Sam Alman says to zoom out a little bit, if you simplify what we do in this whole process, you know, melt sand, run energy through it, and get intelligence out the other end.
20:49
as we realized we were going to need the whole system together to support this, it's just gotten more and more complex.
20:54
So, it turns out Broadcom is also incredible at helping design systems.
20:58
So, we are working together on that entire package.
21:02
Uh, we are able to think from etching the transistors all the way up to the token that comes out when you ask Chachi PT a question and design the whole system.
21:10
all of the stuff about the chip, the way we designed the racks, the networking between them, how how the algorithms that we're using fit the inference chip itself all the way to the end product.
21:21
end product. And so it feels very much like like even if they're not you know fully like plateau pilled like there's a lot more research to be done there's a lot more improvements like there are at least pieces of the chat GPT architecture pieces of the system whether that's you know 40 level inference or reasoning tokens that
21:39
should effectively be baked down onto it ship and so that's what uh what openai is working on with Broadcom at this point um uh uh stratey Ben Thompson had a great uh deep dive on or on OpenAI and Broadcom and uh and sort of laid out the argument for working for them working together and baking stuff down onto a chip. Um
22:01
Um it's uh it's a fun read.
22:05
You should go check it out. >> Should we move on? >> Yep.
22:09
I was just going to try to figure out uh what >> Well, while you do that, let me tell you about Turbo Puffer. Search every bite.
22:16
Serverless vector and full text search built from first principles and object storage.
22:19
fast, 10x cheaper, and extremely scalable.
22:22
>> Uh, Broadcom was a $220 billion company uh the day that Chad GPT launched. It is now >> Wow. >> 1. 7 trillion. >> Almost a 10x. Wow. Nvidia too.
22:33
Nvidia was was much much smaller.
22:36
That might be a 10x as well. Um, yeah. W Yeah.
22:38
Wild wild growth across the entire category.
22:45
Basically, everything is is has 10xed in price or something like that roughly.
22:47
Uh >> yeah, it's interesting.
22:49
It'd be interesting to go back and do the math and the VCs.
22:52
Uh there was some reporting in the Financial Times >> couple days ago talking about who actually owns OpenAI and it would be probably uh interesting and painful to look back at the investors, how much they invested in OpenAI versus just like market buying NVIDIA and Broadcom and some of the other beneficiaries and would they have made more money?
23:12
>> Um there the the OpenAI >> didn't alimter do both.
23:15
Yeah, I'm sure I'm sure the the players like that did both.
23:19
But, uh, OpenAI's expected ownership after the company restructures, Microsoft is going to have about 30%, OpenAI employees are going to have close to 30%.
23:28
Open AAI, the nonprofit, is going to have uh 20 to 30%.
23:35
>> But that doesn't mean you shouldn't still donate.
23:36
If you're thinking about doing some charitable giving this holiday season, consider writing a donation to OpenAI, the nonprofit.
23:43
Prioritize tipping your cap table first. >> Yes.
23:47
>> And then after that, consider donating to OpenAI the nonprofit.
23:51
>> Uh but anyways, Microsoft at 30%, OpenAI employees around 30%, OpenAI nonprofit at 20 to 30%, Soft Bank about 10% and the remaining is is uh Thrive, Kla, MGX, and a number of of uh bedrocks in there.
24:06
Um but >> I wonder how accurate that reporting is.
24:10
It feels like directionally correct.
24:11
>> I think it's directionally correct.
24:11
I wouldn't There's some pretty big error bars on either side. >> Yeah.
24:16
>> Um but still, I mean, you you just look at some of these like some of these numbers that we pulled up.
24:19
It's like uh you know, what was the original OpenAI deal? Uh it was that thrived.
24:24
It was at a 27 billion uh 130 million at 29 billion. So 2900.
24:36
So that initial deal was uh half a percent of the company, right?
24:39
And that was like and that was like a whole round.
24:43
And so that was a very low dilution round even at that time.
24:46
>> Um yeah, what what's what's notable is like looking at this kind of rough cap table.
24:51
Uh it's just astonishing to end up in a situation where VCs have the lowest ownership out of the employees, a nonprofit, a strategic partner.
25:01
Soft Bank kind of makes sense. But uh but yeah. [Music]