Newsletter · 2026-07-30 · Ashutosh Agarwal

Your Great Exit Is Somebody Else's Rounding Error - The VC Read - Week of July 30, 2026

Startups and venture newsletter for the week of July 23 to 30, 2026. New research pegging a billion-dollar exit at ten times harder than getting into Harvard set off a fund-math debate between small-fund discipline and Menlo's be-in-the-outliers-at-any-price approach to Anthropic, alongside the SpaceX post-IPO slide, the Stripe bid for PayPal, and a warning that the acqui-hire has stopped working as an exit.

The VC Read

Week of July 30, 2026: Your Great Exit Is Somebody Else's Rounding Error

New research says a billion-dollar exit is ten times harder than getting into Harvard. So why does almost every cap table get built as if it's the base case?

The Big Debate: who is the exit actually for?

For four weeks running, the loudest argument in venture podcasts has been some version of "is AI a bubble." This week something more useful surfaced, a fight about arithmetic that has nothing to do with whether Nvidia is overvalued, and everything to do with whether you and your investors are even playing the same game.

It started with a study.

Micah Rosenbloom, managing partner at the seed firm Founder Collective, had his team look back 25 years at every large venture exit they could find, and then asked a deliberately unglamorous question: what was the median outcome among the 500 biggest exits?

The answer, on Venture Unlocked, "Why $100M Exits Matter More Than Trillion-Dollar Dreams" (July 28):

"It's about $2.7 billion is the median value of an exited company, the top 500. And so you sort of say to yourself, if you're building a fund, if that's the median, like, you know, and you own 10% of that company, it's pretty tough to return even a billion-dollar fund. It's pretty meaningful for a $100 million fund."

Then he put a denominator under it. Those 500 exits came out of a pool of more than 100,000 companies started over the same period.

"The probability of starting a billion-dollar exited company over the last 25 years is 10 times harder than getting into Harvard. It's about 0.45%. Meaning of those 100,000 companies that have been started in the last 25 years, 400 and some odd have exited over a billion. And Harvard's undergrad is like 4% admissions rate."

He is careful not to sneer at ambition. His objection is to strategy:

"Look, I take nothing from the ambitious people who want to be the top 0.4%. I just don't want to build a strategy. It's hard enough to do what we're doing."

Rosenbloom also made the point that gets lost in the headline numbers: for companies outside the very top sliver, exit values have barely moved. "The average valuation exit for a non-1% company hasn't really changed too much. It's still sub 100", meaning under $100 million. "But we're now underwriting as if all these companies are going to be the top 1%."

The tension is structural, not moral

Here is the part every founder should sit with, because it explains a conversation you may one day have with your own board.

Take a $500 million exit. Rosenbloom walks the math both ways:

"So think about $100 million fund. If you're in a company that exits for $500 million, which is a great exit, by the way, in today's world... you own 10%. That's $50 million bucks of $100 million fund. You're returning half the fund. If you own 15%, let's call it $75 million... But the tension arises like when you're a big fund, a $500 million exit where you own 10% does nothing. It's 5% of your fund if you're a billion-dollar fund. And so those founders then get pushed for more growth."

Same company. Same outcome. Career-defining for one investor, a rounding error for the other. Nobody is behaving badly, the incentives simply point in opposite directions.

He then told a story that makes it concrete. A founder wanted to sell after his Series B, at $750 million, owning roughly 30% of the business:

"Great exit. Life-changing for this person who owned about 30%. But the VCs were trying to talk him out of it and saying, why are you selling? This could be much bigger."

And a second, quieter problem: ownership is not holding up the way the optimistic version of this story requires. Rosenbloom says average entry valuations at his firm have risen three to four times over the past decade or two, and yet:

"I think people are raising way more money, and they're giving out way more equity... some of them are even surprised at how little of a company they own by, like, Series D or E. And they're like, oh, my God. We only own, and it's like, I led the seed, and we own 2%."

If exits get bigger and your slice gets thinner, the two effects can cancel out. You need an even larger outcome just to stand still, which pushes everyone further up the risk curve.

His last point is about risk that has nothing to do with valuation. Big rounds carry an obligation:

"When you raise $50 million or $100 million or whatever the round is, you're expected to put that money to work very quickly to generate growth almost at all costs. And that is a very risky thing that creates even more misalignment between founders and people on their board."

Founder Collective has put its own money where its mouth is. Its fifth fund is still under $100 million, Rosenbloom believes they may be "the only branded Seed fund that has stayed sub $100 million." He notes many peers who started with $40 million funds now run $300 million to $500 million and are, in his view, "not really a Seed firm" anymore. He also mentioned that roughly 2,000 new US firms or funds were created between 2012 and 2023.

His own history informs the caution. He and co-founder Eric Paley sold their company for $95 million rather than raising a Series B:

"We were about to raise a Series B. And thankfully, we didn't. How would things have played out? I don't know. But I'm glad the financial crisis hit. Like a lot of things could have turned that into zero."

Now the other side, and it is genuinely strong

The counter-argument is not "small funds are dumb." It is that the distribution of outcomes has changed so violently that the old ownership discipline is now the expensive mistake.

Matt Murphy of Menlo Ventures made this case about as vividly as it can be made, on 20VC, "Leading Anthropic's First Ever Round | Will Open Source Threaten Anthropic's Business | Do Margins Matter in a World of AI | Why Triple, Triple, Double, Double is Not Good Enough Today | Why Series A is Hard Today with Matt Murphy @ Menlo" (July 27).

Menlo led Anthropic's first round. The check was "a little over $10" million, at a $4 billion valuation. Murphy admits the internal reaction was scepticism, and it was about exactly the arithmetic Rosenbloom describes:

"It's like a $4 billion valuation of venture fund. That's not what we should be doing. What are LPs gonna say?"

Harry Stebbings then role-played the sceptical partner, and it is the cleanest statement of the traditional objection you will hear:

"$10 at $4, I would be sitting in your partnership going, well, let's just like outcome scenario plan this. If it's a $40 billion company or an $80 billion company, let's say you do $80, it's a 20x with dilution. Traditional says 50%, it's a 10x. We're going to turn the $10 into $100. Wow. Thanks for returning 12% of the fund, Matt."

Murphy's answer was essentially: that model is obsolete.

"I've been in the business for 25 years now, you were kind of saying like, hey, great outcomes are $300 million, $500 million, $1 billion. So you're like, hey, you have to own 20% to get $100 million or whatever... That's not how the game is being played anymore. It's like you have to be in the big outliers to drive great returns, and you're better off being in them at a very small percent than owning a large percent of a company that exits for $3 to $500. Those just aren't gonna move the needle."

Asked directly whether ownership matters less than it used to, he answered: "By far." And he was pointed about firms that hold the line:

"You kind of get into these situations where we have to own 15% or 20% ownership, or we don't do this and don't do that. And I think the new Menlo that I'm part of has shown extreme flexibility to just do what makes sense. Let's get in this great company because once you're in, hey, if it takes off, there's plenty of opportunity to put more capital in."

That is not theory. Menlo followed the first cheque with an SPV of more than $500 million, a special-purpose vehicle, meaning a one-off pool raised from its own LPs to invest in a single company outside the main fund. Stebbings did the back-of-envelope on what Menlo's Anthropic position is now worth and floated "around $10 billion in carry." Murphy corrected him: the position is worth north of that, and carry (the investor's share of profits) is a fraction of it. Stebbings settled on "$2 to $3 billion." For one position, from a $10 million cheque.

Murphy's account of how conviction built is worth reading, because it wasn't a spreadsheet. Anthropic was pre-revenue and pre-launch when Menlo first invested, but benchmarks showed it at or above ChatGPT's level having spent, in his words, "like, I don't know, a 50th of the capital", evidence of real technical efficiency. Then Amazon and Google arrived as investors, technical partners and distribution partners via Bedrock and Vertex. Then revenue started compounding, "you'd see them adding 10 this month, 8 the next." The trigger for the big SPV was an LP meeting in November where an Anthropic executive presented:

"He blew everyone away. Like after the meeting, our LPs were like, 'This is crazy. This company is amazing.'... We literally came out of that meeting and said, 'All right, we've got to do this. We've got to figure out a way to lead the round.' And 2 weeks later, we signed a term sheet."

Notably, Murphy does not want an enormous fund either, "we don't have a $10 or $20 billion fund, nor do we aspire to have that." His resolution is the SPV: keep the fund a sensible size, and go to LPs when a single company deserves more than the fund should hold.

The people writing the cheques to the cheque-writers agree

Scott Voss has been at HarbourVest for 27 years, investing in venture funds on behalf of institutions. On How I Invest with David Weisburd, "E408: HarbourVest: Why Venture Capital Is Chasing Trillion-Dollar Companies" (July 27), he framed the objective in a way that leaves no room for the median:

"What you want is economic interest in the short list of companies that matter. So if it's the 10% of companies that are going to return 90% of the market over the next decade, how do you kind of gain access to those companies?"

His conclusion is to stop caring about how you get in. HarbourVest will reach the same company through a fund commitment, a direct co-investment, or by buying someone else's stake on the secondary market. He used the security company Wiz as an illustration: the primary fund portfolio made roughly 100 times its money, the direct portfolio 3 times, the secondary portfolio 2 times. Same asset, three entry points, wildly different returns, and wildly different fees. Primary fund commitments are the most expensive access, starting at "2.5 and 25" (a 2.5% annual management fee and 25% of profits) "and it goes up from there." Weisburd noted he has heard 2.5 and 30. Co-investments are sometimes free of both fee and carry; buying a secondary stake usually carries no fee on the interest itself.

Voss also offered the most useful reframe of the fund-size argument all week, that it need not be a binary. He pointed to Michael Gilroy's $400 million Marathon Venture Partners as an example of a firm that stays boutique and leans on LPs when it needs scale:

"If you've got a $500 million fund or a $300 million fund, but you have sophisticated LPs around the table who have hunger for co-invest and are predictable in how they'll behave with that co-invest, it allows you to run your fund in more of a boutique way, like a $300 million fund. But when you need to scale up, you've got that optionality available to you through your limited partners."

He set a hard bar for the asset class overall. If venture returns a 15% median, "that's not a reason to invest in venture capital." Twenty percent, sustained for 10 to 20 years, is the game. And he was blunt about timing:

"Venture's on a good run right now and everybody wants in, but you needed to be in five or 10 years ago... You've convinced yourself you want to be in this asset class. That's great. But convince yourself you should be in it for the next 20 years and don't change your mind."

Where does he think this ends? A split market:

"It becomes an oligopoly where it's probably 10 players that dominate the scale venture part of the market. But then there's going to continue to be that traditional venture market that we all grew up knowing. It's the highly idiosyncratic, fragmented part of the market targeting the entrepreneur and the idea, trying to nurture that company from 0 to 1."

If he's right, both Rosenbloom and Murphy are correct, about different businesses that happen to share a name.

The strongest evidence against "always let it run"

Here is where the debate gets genuinely hard, and it came from the bull side, not the bear side.

Voss brought up Klarna as a caution against assuming that holding on is always right:

"Klarna, $15 billion value at IPO. I think it trades below $10 billion now. There was a private market financing in Klarna that valued at $50 billion. And there were investors that sold into that financing. So that was a 10X decision."

Read that again. The investors who sold into the $50 billion round, who took the "unambitious" option, made ten times what the holders did. The decision to sell was the high-conviction call.

Voss's own framing of how rare these moments are is worth keeping:

"I can go back at my 27 years at HarbourVest and I can honestly say I've probably made thousands of decisions, investment decisions along the way. But the ones that are like super consequential, I can count on two hands."

Against that, he told a story about an investor who kept rolling every dollar of Roblox gains, including crystallised carry, back into continuation vehicles rather than banking anything. Voss told him to take money off the table, that his wife would kill him. The reply:

"Look, Scott, this keeps us aligned. It keeps us hungry. I've seen too many venture investors who, when they had that big monetization event, they lose that edge that they had that allowed them to become great."

Two investors, opposite conclusions, both defensible. That is what an actual debate looks like.

The quiet resolution: founders hold more power than they think

The 20VC roundtable of July 23, "OpenAI and Anthropic Threatened by Kimi? | Should the US Ban Chinese Open-Source Models | Should Openrouter Sell & Value in the Routing Layer? | Stripe Buying Paypal: What You Need to Know", landed on the part of this that founders most often get wrong.

Rory O'Driscoll of Scale Venture Partners, on the fear that investors will force a sale:

"Founders agonize about this when they think about control and they think, oh my God, these people are going to make a sale. Even if, as the VCs, we have board control, 70% ownership, and a drag along for the founders... If the founders who are core to the business don't want to sell, it's not going to happen. So the first thing I tell the founders is, it's your decision, which I think is very empowering because it takes it away from people are going to make you. That's the beauty of being private, unlike being public, where you don't have that degree of freedom."

His script when an offer arrives is a small masterclass in how a good board member behaves:

"One, the window's open, it doesn't open often, we should take it seriously. Two, I'm going to support you in whatever you want to do. Three, if there's concerns that you have that you've been sitting on and not telling me, now would be a good time to share so we can make an informed decision... I don't think you should pressure people into selling."

And the question he actually puts to the founder: "Do you think you can be worth 3x this in three, four years?"

That is the whole debate compressed into one sentence, not "is a billion possible," but "is the extra risk and the extra years worth it to you."

Where I land

Rosenbloom has the data; Murphy has the receipts. But they are answering different questions, and the honest reading is that the mismatch is the story.

The Menlo strategy is unfalsifiable as a general rule. "Be in the outliers at any price" works beautifully if you correctly identify Anthropic in 2023 and can raise a $500 million SPV in two weeks on the strength of your LP relationships. Most firms cannot do either. For everyone else, "ownership doesn't matter" quietly becomes "we will overpay for things that are not Anthropic," and the fund math still has to close.

Meanwhile the founder in Rosenbloom's story, $750 million and 30% ownership and told not to sell, is not a hypothetical. There is a real, growing gap between what makes a family wealthy and what moves a multi-billion-dollar fund. As Rosenbloom put it: "Life-changing money is still life-changing money. You don't need trillions."

The swing factor, as ever, is dilution. If founders and seed funds keep giving away more equity per dollar raised, the bigger exits get consumed by the bigger cap tables, and the alignment gap widens for everyone. Watch ownership at Series D, not headline valuations.

And a footnote that supports the pessimists from an unexpected direction. On Goldman Sachs Exchanges, "Is the Surge in US IPOs a Warning Sign for Investors?" (July 28), Jay Ritter of the University of Florida explained why the number of companies going public has stayed low even as dollars hit records: modern tech markets reward scale so heavily that "getting big fast is more important than it used to be," and as a result "there have been exits increasingly for successful VC-backed companies that have been trade sales rather than the company remaining independent and going public."

Translation: the trade sale, the $100 million to $2 billion exit that big funds shrug at, is not a consolation prize. Structurally, it is becoming the main event.

Signals

• SpaceX became the cautionary tale about getting what you wished for. For years the complaint was that no liquidity ever arrives. It arrived, and it has been rough. On Best Stocks Now with Bill Gunderson, "Monday July 27, 2026", Gunderson noted the stock went public at $135.00 and was trading at $110.54, with market capitalisation down from nearly $3 trillion to $1.53 trillion. Four days earlier, on "Thursday July 23, 2026," he described it as down roughly 50% from its $225 high and in a "massive downtrend," calling the earlier valuation "the most ridiculous" he had seen in years. The Elon Musk Podcast, "Alphabet Funds Orbital AI With SpaceX Gains" (July 27), added the structural detail: an $86 billion raise at $135 a share, only 4-5% of the company floated, and Musk retaining more than 80% of voting control through super-voting shares. One clean winner: Alphabet's stake, originally about $1 billion, is now worth roughly $94 billion. For the "why," the sharpest version came slightly earlier, on Hidden Forces, "SpaceX, Europe's China Problem, and the Populist Backlash | Patrick Boyle" (July 20), where the pricing was picked apart as roughly 100 times sales for a company losing about $5 billion a quarter and growing 15% a year, against Google's 8 times sales and 200% growth at its own IPO, with the price set by Musk unilaterally rather than through the usual investor feedback process.
• The lockup is the next real test. On RiskReversal Pod, "Rick Heitzmann: The Gray-Area Deals Funding the AI Buildout" (July 23), host Dan Nathan flagged that SpaceX reports on August 4, at which point 20% of eligible shares become sellable. That, not the last three weeks of chart damage, is when we learn what pre-IPO holders actually do with their paper gains.
• Stripe going after PayPal is the year's most interesting piece of M&A arithmetic. On the same 20VC roundtable (July 23), Rory O'Driscoll's opening verdict was simply "price clears all." Both companies process roughly $1.8 to $1.9 trillion a year. Stripe is valued around $150 billion; the offer for PayPal is in the $50-billions, Motley Fool Hidden Gems Investing, "Paypal to Stripe: You're Going to Have to Do Better Than That" (July 28), put it near $60 a share, about $53 billion. The comparison is trickier than it looks because Stripe reports revenue net (roughly $6 billion) while PayPal reports gross (roughly $30 billion, about 1.7 times revenue). On a like-for-like basis, O'Driscoll reckons PayPal is about one and a half times Stripe's size, "still a company buying something one and a half times its size for what looks like a third less," with the private equity firm Advent sharing the load. He called it "a big ballsy play to double your market cap," compared it to the Dell buyout, and was candid about the culture risk: Stripe will "have to do a lot of hard-nosed stuff to turn PayPal around," and "you're probably going to be looking at that place and saying, we're going to get rid of a lot of people."
• The growth math is the catch. Jason Lemkin pushed on the obvious problem with buying something growing 7% when you grow 20-30%: "if I take 30% and seven, that's 37% and divide by two, I'm only growing like 18% now. I've fallen below the Mendoza line of 20% growth at scale", the rough threshold below which software companies lose their premium. O'Driscoll's rebuttal: "There's no such thing as a Mendoza line for growth at 5 billion and above because you can get out." Lemkin also explained why staying private made the structure necessary: you cannot issue $50 billion of stock, so you need debt and a partner. And on whether it closes, he was unequivocal: the board's rejection is the tell. "You reject it because no investment bank will tell you you're allowed to make your highest offer up front... It's a dance. They're going to accept it."
• The IPO market is setting dollar records while barely functioning as a market. The numbers cited on RiskReversal (July 23): 86 IPOs priced in 2026, down 26% from last year, raising $142 billion, up about 130%, and roughly half of that total is SpaceX alone. Filings are up nearly 10% to 152. On Exchanges (July 28), Owen Lamont of Acadian Asset Management argued this doesn't qualify as a wave at all: "In the past three years, we have been in the opposite of an IPO wave. We have been in an IPO drought." His test is refreshingly simple, in 1999 and 2021 there were about five IPOs a week, so "if there's not an IPO happening every business day, you're probably not in a wave." One enormous listing doesn't count; he compared it to Saudi Aramco in 2019, "an idiosyncratic event, not really reflecting anything about world capital markets."
• The most useful bubble tell nobody is watching: first-day pops. Also on Exchanges, Lamont noted that IPOs typically rise 15-20% between the offer price and the first day's close, and that in genuine manias those pops become enormous. "We have not seen IPO pops of the magnitude of 2021 and 1999. So, that's a sign that we are not in a speculative euphoria." He was careful to say issuance is one of his "four horsemen of a market bubble" (firms sell equity when equity is dear) but that the signal marks the beginning of a bubble, not the end: "In Japan, it lasted for years. The IPO wave of the 1990s lasted for years." His caution for anyone buying this year's listings is the historical one: "there's absolutely no guarantee that the winners of AI are the firms that are going public this year. Because the firms that went public in 1998 were not the firms necessarily that benefited most from the internet... Netflix and Google, they went public much later."
• Can the market swallow the supply? Ritter thinks the worry is overdone, and the numbers are striking: US public companies have been paying out roughly $600 billion a year in dividends and buying back about $1 trillion of stock, so "$1.6 trillion of cash being paid out that needs to be recycled." His caveat matters, though: that arithmetic is weakening in 2026 "partly because some of the big tech companies that have been big repurchasers of shares have instead moved to be net equity issuers." Lamont's historical bound: in the 1990s, technology share counts grew 3-5% a year for several years, so the market can digest a lot before price gives way.
• The acqui-hire has quietly stopped working as an exit. This was the sharpest new argument of the week, from Rick Heitzmann of FirstMark on RiskReversal (July 23). His example: a group of people left ServiceNow to start their own AI lab, the way Workday's founders once left PeopleSoft. The obvious ending is that ServiceNow buys them back. The problem is who can afford it now that big platform software companies trade at "mid-single-digit revenue multiples." As he put it, "this isn't meta, which is producing a ton of cash as a trillion-dollar market cap that can acquihire someone for effectively a billion dollars. That's not on the menu for almost every company in the world." Then the punchline, which every seed investor should read twice:

"If you invest in a company and put $100 million in it, a billion-dollar valuation, and ServiceNow says, you know, I could really squeak and pay $1.1 billion for a bunch of guys with an idea but not any product, not any revenue. The guys that raised it at a billion-dollar valuation are super excited. ServiceNow is kind of holding their nose on price, and then the investors kind of get their money back. And you're like, okay, this isn't really working for everybody, so it's not going to be sustainable."

A $1.1 billion exit where the investors merely break even. That is what a headline number can hide.

• Selling the plumbing while the plumbing is scarce. Reports that OpenRouter, which routes requests across different AI models, is in talks with several acquirers drew a firm "sell" from Lemkin on 20VC (July 23): "I think it's a great time for OpenRouter to sell. I think them leaking the story, it was very savvy." His reasoning is that the market has figured out it needs this, but it remains "a niche product that more and more people are going to build variants of themselves," and platforms like Databricks will build their own. At a rumoured five or six billion, "I might check out." O'Driscoll's contribution was the buyer's logic, and it is a genuinely useful lens: "it may well be that the NPV of the company on a standalone basis is, you know, a couple of billion, not huge. But the value to it right now to a hyperscaler, if they could shift 10% market share in the enterprise to them over the next half a decade... could be interesting." Then, the line that tells you what a venture investor really thinks: "shit, I wish I was in that."
• Secondaries and continuation vehicles are becoming routine in venture, and are surfacing awkward questions about carry. On Swimming with Allocators, "AI, SaaS and the Next Private Markets Shakeout" (July 29), Nick Kasson, who heads the secondaries practice at law firm Sidley Austin, described three venture deals on his desk simultaneously: an LP trade, and two GP-led deals (one representing the incoming money, one the venture sponsor). The interesting friction is alignment. New investors coming into a continuation vehicle typically demand that the manager roll not just its existing investment but any crystallised carry into the new structure. Kasson is currently negotiating an exception so a sponsor can "clean up shop" and let senior partners approaching retirement take money off the table. He also noted this isn't automatic: "there's no default that entitles the carry holders in the existing fund to those economics as it relates to the CV unless they have that specifically negotiated." As funds stretch across ever more years, this generational-handover problem becomes standard.
• "SPV sandwiches" are the week's most under-discussed risk. On the same episode, an institutional allocator described the practice of stacking special-purpose vehicles on top of one another to sell access to hot private companies: "it's just layer upon layer upon layer. And you don't know, number one, what the underlying assets may actually be. But you certainly don't have great transparency into the layers of fees upon fees upon fees upon fees." The comparison he reached for was not flattering, "it reminds me of something that happened maybe in like 2007, 2008, when we just kept layering things and nobody knew what they had... And then you package it and sell it and repackage it and resell it." Sometimes what is inside isn't even what buyers assume: restricted stock, an option, or in the worst cases outright fraud. His flag for the retail end of this: someone posted on X boasting she had made more from a single Anthropic SPV than in her life to date. His practical test for anyone pitching you private access is delightfully low-tech: check whether they have any credible investors in their LinkedIn network at all.
• Capital concentration, in one statistic. On How I Invest (July 27), David Weisburd noted that Anthropic, OpenAI and SpaceX together raised something like 75% of all venture fundraising in the first quarter of 2026, and offered a striking reason they went public: they had to, because "most VC firms that would have invested were already too concentrated in those names." When the private market runs out of buyers, the public market stops being optional. For scale, he cited roughly $2 trillion in private markets against $127 trillion in public ones.
• Valuations at the top kept climbing anyway. The Elon Musk Podcast, "Anthropic overtakes OpenAI as most valuable startup" (July 23), reported Anthropic raising a Series H at a $1 trillion valuation, with OpenAI at $852 billion. Separately, All-In, "The Fight Over Open Source AI, Anthropic's $1.5B Payout, NYC Socialists: Evictions = Violence?" (July 24), covered Anthropic's $1.5 billion copyright settlement with authors, a reminder that legal exposure at these companies is now measured in billions too.
• On IPO timing, the insiders and the press disagree. Bloomberg Intelligence, "Moonshot Plans IPO in Six Months After China AI Breakthrough" (July 20, just outside our window), reported Moonshot AI targeting a Hong Kong listing within six months at a $30 billion valuation, Anthropic expected to move in September or October, and OpenAI treated as a 2027 story. Heitzmann, who has companies in registration right now, was more sceptical about the frontier labs: "I don't think that it's imminent, the OpenAI and Anthropic IPOs." His more encouraging observation is that the window isn't only open for AI. Bending Spoons, a Milan-based roll-up of consumer subscription apps, went public roughly eight to ten days after SpaceX, despite everyone assuming SpaceX would absorb all the oxygen, and "traded really well."
• A useful reframe on what an IPO even is. On Run the Numbers, "The Four Stages of World-Class FP&A with Datadog's AJ Ljubich" (July 23), Ljubich, who took both Datadog and UiPath public, recalled Datadog listing in 2019 at a sub-$10 billion valuation on what was then a record software multiple; it now trades above $80 billion. His framing is a healthy corrective to treating a listing as an ending: an IPO is a fundraising and liquidity event with branding benefits, not a finish line, and going public does not make a company fundamentally better the day after.
• And the week's oddest exit. The Colin and Samir Show, "9 Lessons From TBPN's $100M+ Exit" (July 28), covered OpenAI acquiring the tech livestream show TBPN for a figure described as over $100 million. Co-founder Jordi Hays's account of the leverage is the interesting part: the business was performing well enough that they were not looking for a buyer, which is precisely why the terms were good. Sell from strength, or don't sell.

Quote of the Week

Micah Rosenbloom, Founder Collective, on why his firm's fifth fund is still under $100 million, Venture Unlocked, "Why $100M Exits Matter More Than Trillion-Dollar Dreams" (July 28):

"You need optionality. You need flexibility. And a small fund allows for that. A big fund forces a big outcome. And it's much more binary."