Samsung-Perplexity and the Google Question
Welcome to Cautious Optimism, a newsletter on tech, business, and power.
Monday, and it’s time to rock. Our economic calendar for the week is here. Let’s get it! — Alex
📈 Trending Up: AI companies paying for IP … rural consumption in India … modest AI productivity gains? … brain-computer interaction … cruelty … neobank profits … Microsoft investment in Switzerland … Google’s AI products … Russian influence in Poland … Chinese tech? … data center deals …
📉 Trending Down: Concerns about data center overcapacity (again) … SoftBank’s ability to not do deals … trade worries … China’s real estate crisis, at least in tier-1 cities … airplanes as hallmarks of diplomatic rapprochement … politics-free science …
Chime + Circle = Happy Alex
As discussed earlier today on TWiST, we’re heading for a good run of IPOs in the next two weeks. Here’s the latest:
American neobank Chime’s latest IPO filing indicates it is targeting a $24 to $26 per-share IPO range, which would give it a midpoint, fully diluted market cap of $10.5 billion. That’s just 40% of its ZIRP-era private-market valuation, but I suspect that Chime wants to reprice higher at least once before it actually lists. So let’s hold our fire on 2021-2025 pricing comparisons.
For fun: At an $11 billion price tag, Chime is worth 5.3x its Q1 2025 revenues, annualized.
South American rival Nubank, in contrast, is worth 4.5x its Q1 2024 revenues, annualized. From that perspective, Chime is looking pretty pretty when it comes to its valuation.
American stablecoin giant Circle’s IPO has a new, higher price range and a new, greater expected share sale count. From $24 to $26 per share, Circle now expects its shares to sell for between $27 and $28 apiece.
Even more: Circle now expects to sell 12.8 million shares (primary) with existing shareholders selling 19.2 million. Those figures are up from 9.6 million and 14.4 million, respectively.
As RenCap notes, the higher expected share price and the larger share count mean that Circle could raise as much as 47% more than it had previously anticipated. That’s a lot.
And there’s Voyager Technologies, an acquisitive space tech company that is also part of the private-market space station Starlab project, is looking at a $1.6 billion IPO when it itself lists.
When do these companies finally start trading? Circle should price on Wednesday and trade on Thursday, while Voyager is expected to price on June 10th and trade on June 11th, while Chime is expected to price on June 11th and trade on June 12th.
Good. Even this run of public liquidity won’t do more than dent the DPI-drought currently plaguing venture-land, but it is as welcome all the same.
Samsung-Perplexity and the Google Question
Currently, Google pays Samsung handsomely to install Gemini on its hardware. The tie-up ensures that Google accretes new AI users. The way I see it, Google pays Samsung to help it counter OpenAI.
But while Samsung has been content to cash Mountain View’s checks, it could have a new deal in mind. Bloomberg reports that Samsung may preload Perplexity’s AI tools in the future, even integrating “the startup’s search features into the Samsung web browser.”
Why would Samsung potentially limit its Google income while betting on a startup? Because it might be a part owner in short order. The same article notes that the South Korean hardware giant could put up one the largest checks in the much-rumored $500 million Perplexity round that is coming together at a price of up to $14 billion.
As Samsung is worth a quarter trilly, give or take, it can afford the outlay. And probably a reduction in its Google income. For Perplexity, though? Such a deal would be more than a coup. It could become a distribution deal par excellence, a JATO pack of sorts for the search and productivity startup.
Recall that Google is currently in trouble for its search deals, transactions in which it purchases default status on mobile devices. Could Samsung be hedging its bets ahead of rulings there?
Perplexity crossed the $100 million ARR threshold earlier this year. Without the Samsung deal, that feels expensive. With? At a minimum, the valuation can be argued into submission. It could prove a deal.
Roll me up
Consider this a starter entry, but we need to talk about venture firms getting into the rollup game. Here’s a good definition of private equity rollups, to get us started:
A private equity roll-up is the process of acquiring and merging multiple smaller businesses in the same industry into one larger consolidated company.
What does that have to do with VC and startups? Aren’t upstart tech companies — the main fare of any venture portfolio — very different beasts than the sorts of small businesses that PE likes to snap up and merge like smushed Play-Doh?
Yes, of course they are. But venture capitalists are doing more than merely dipping single toes into the rollup game. And, thanks to recent interviews and a little digging through venture capital literature, I explain what’s going on — and why, in just a few simple bullets.
First, the underlying trends that brought us to today:
Software companies cannot extract the full value of their offerings: An argument I’ve made for some time is that software is too cheap. Whatever TWiST pays to have me on its Slack instance is a pittance compared to the value it brings. I could not do my job at the podcast without the comms tool, so, the theoretical full value of Slack in my role is a huge — 10x? 100x? — multiple of whatever we are paying for it.
The reason why software is so cheap — as Jason argued this morning — is that if a software company wanted to charge more for its service, switching isn’t hard, and there are open-source alternatives to consider. Not to mention just vibe-coding up something in-house. Essentially, the insane gross-margin profile of software — that it is all but zero-marginal-COGS — means that it will always sell at a fraction of its delivered worth.
Great for customers, irksome for founders and VCs.
Venture capitalists have too much money: Ironic, I know, given the current venture fundraising drought, but I don’t think it’s too controversial to say that the rise of mega-funds has changed how many VCs think about investing. For example, VCs going the RIA route, building perpetual funds, rolling over deals between vehicles — it’s all in there.
Second, the solutions to the problem:
Invest in startups rolling up and automating traditional companies: Some VCs are backing startups that want to buy real-world businesses — the ur example is a midwestern HVCA company owned by a family — apply AI-led automation, improve their margins, and then use the cash generated thereof to buy even more of the same flavor of company.
For VCs, this is fun, twice: It’s possible to extract a lot more value from installed software if you own the business that earns all the benefits, instead of merely the cut you get from helping a company through, say, per-seat pricing.
And since buying companies takes capital, it’s a great sink for AUM.
Roll up and automate traditional companies at the firm level: The other version of the VC-rollup game is the investors doing the buying, upgrading, levering, and expanding. I doubt we’ll see as much of this, as it’s a lot of work. And what VC wants to spend time in Duluth arguing with dry cleaning suppliers. Still, some will try this to pass the time while trying to bid on yet another OpenAI secondary SPV.
Too clever by half is a phrase we plebs use often. In poker it’s called Fancy Play Syndrome. Or as my Dad says, excessive cleverness usually fails. None of those chestnuts apply to finance, however, where complexity is the name of the game.
Perhaps the VCs will do well here. I mean, I hope so. But the strategy of better monetizing the upside of software won’t be easy — or quick. I wonder if we’ll see similar IRR from roll-up VCs as we have from say, upper-quartile Seed-stage firms.
But as mega-funds are playing a different game, one that absorbs larger checks for more modest returns, perhaps that doesn’t matter?
when you are happy, i am more likely to be happy