- Buysiders
- Posts
- The AI IPO Race
Together with
Good morning! Thanks for your email, this summer I am working as hard as my European colleagues and will be OOO until September 9th. For urgent matters, don’t reach out.
Cheers,
Buysiders.
But, for those who want to get ahead…
You don't need to work at Blackstone to think like they do. Learn exactly how their analysts are trained with Wharton Online’s 8-week PE Certificate Program.
First time reading? Sign up here.
Have feedback? Respond here.
THE AI IPO RACE
Anthropic Claims its Next Victim
After dismantling the entire SaaS industry in a matter of hours, Anthropic just claimed its next victim - OpenAI.
This shouldn’t come as a total surprise. I assume most of you reading use Claude at work - don’t lie - and not ChatGPT (and certainly not copilot). It appears that investors are finally seeing that too. Whether or not they are too late remains to be seen.
Since people have cared about the AI race, OpenAI has been the clear leader in terms of valuation. Until last week when Anthropic announced a $65 billion Series H at a $965 billion valuation, led by Altimeter, Dragoneer, Greenoaks, and Sequoia. The company also released Claude Opus 4.8 the same day.
For the first time, Anthropic is worth more than OpenAI. OpenAI’s last confirmed valuation was $852 billion, set when its $122 billion round closed in March. The gap is $113 billion. Both numbers are probably stale already but that’s life in the private markets.
Twelve months ago, OpenAI was at roughly $10 billion in annual revenue and Anthropic was at $1 billion. The idea that Anthropic would catch OpenAI seemed like something you told your IC to justify your outrageous mark to market.
OpenAI ended 2025 at $13 billion in actual annual revenue, reached a $24 billion run rate by late March 2026, and is now generating approximately $2 billion per month. That is one of the fastest revenue ramps in the history of enterprise software. It is also, it turns out, not fast enough.
Anthropic ended 2025 at $9 billion in annualized run rate. By February it was at $14 billion. March, $19 billion. April, $30 billion. This week: $47 billion. The company went from $10 billion in annual revenue to a $47 billion run rate in approximately twelve months. Dario Amodei told investors the growth felt “too hard to handle.” Normal numbers would return soon, he predicted. They have not. Dario is doing fine.
The driver is Claude Code, the AI coding assistant that has quietly become the product every engineering team’s budget is being routed toward. One firm reportedly spent over $500 million on Claude in a single month. We don’t know which firm. We do know their CFO had questions.
Revenue is an exciting number, but let’s not kid ourselves that either of these companies actually earn a profit. Burn rate is the number that really matters. After all, how long can $100 billion last? Longer than you with your Hinge date, hopefully.
OpenAI is projected to lose $17 billion in 2026. Its burn rate stays at 57% of revenue through 2027. The company has committed approximately $1 trillion to infrastructure over the next several years, because Sam Altman’s view is that the risk of underinvesting in compute is greater than the risk of overinvesting. He may be right, and he is certainly spending as though he is.
Anthropic’s burn rate should fall to one-third of revenue in 2026 and 9% by 2027. The company projects positive cash flow by 2028. OpenAI projects positive cash flow by 2029 at the earliest, by which point Anthropic should have been profitable for a year and will have sent Altman a very polite note about it.
This purely comes down to business strategy. Anthropic is an enterprise company with consumer products. OpenAI is a consumer company building enterprise products. 80% of Anthropic’s revenue comes from businesses. OpenAI has 900 million weekly active users, roughly 5% of whom pay anything at all. Serving 895 million free users is very expensive. It is also, arguably, an extraordinary distribution advantage. The two companies have made genuinely different bets and both are, for now, winning in their respective fields.
Anthropic is also deliberately not building image generation, video, or the other compute-hungry modalities that eat margin without clear enterprise ROI. This is either disciplined strategy or the AI equivalent of not getting a gym membership so you don’t have to pay for one. The revenue trajectory suggests the former.
It’s not all sunshine and computer generated rainbows for Anthropic though. Anthropic is fighting a Pentagon designation as a supply chain risk, the result of CEO Dario Amodei refusing to grant the military unrestricted access to its models. Defense Secretary Pete Hegseth made the designation in February after a public clash with Amodei. The dispute is working through two federal courts.
This is not ideal timing for a company trying to IPO.
OpenAI does not have this problem, partly because its relationship with the government is warmer and partly because nobody has tried to designate it a security risk yet.
Both are targeting a listing this year. Anthropic confidentially submitted its draft S-1 to the SEC yesterday, while OpenAI filed its own confidentially after beating Elon in court. The question now: who hits the market first?
The company that files first sets the narrative. The company that files second gets compared to it for the rest of the roadshow. Both could be doing this in the same quarter as SpaceX.
If Anthropic really is better than OpenAI financially, then going second could make answering roadshow questions easier and going first would make OpenAI’s roadshow brutal, so either way it may be a net win for Anthropic.
Q4 2026 is going to be a good quarter to be an ECM MD. A less good quarter to be an ECM analyst.
PRESENTED BY WALL STREET PREP
The Best Investment You'll Make in Your Finance Career Takes 8 Weeks
The Wharton Online + Wall Street Prep Private Equity Certificate Program gives you the same hands-on training used inside Blackstone, Thoma Bravo, and Carlyle.
Earn your certificate in 8 weeks, online, without putting your career on hold.
Live office hours with Wharton Business School faculty and guest speakers who have actually closed the deals you're reading about
A 5,000+ network of finance professionals, plus lifetime access to events and meetups
A Wharton Online certificate (proof that you've done more than skim the KKR annual report and call it research)
The June 8 cohort is filling up. Use code SQUEEZ to save $300 and secure your spot.
INTERNATIONAL DEAL OF THE MONTH
DoorDash Goes International
Uber wants to buy Delivery Hero. It just doesn’t want to say so in writing.
Here is what has actually happened this week. Uber now holds a 36.83% economic stake in the Berlin-based food delivery company. Its voting rights are capped at 24.99%. German law requires any acquirer that crosses 30% of voting rights to make a mandatory cash offer to all shareholders.
Uber has deliberately structured its position to sit just below that threshold, owning more than a third of the company economically while keeping enough of the stake in instruments and derivatives to avoid triggering the mandatory bid.
This is not an accident. This is a very well-compensated German law firm earning its fee.
The timeline is worth understanding. A few weeks ago, Uber held roughly 7% of Delivery Hero. It quietly built to ~19.5% through open market purchases, becoming the company’s largest shareholder. The following weekend, Bloomberg reported Uber was exploring a full takeover.
Uber’s own shares fell 1.6% on that news, because markets were not thrilled about Dara Khosrowshahi paying $11.6 billion for a money-losing German food delivery company. With SpaceX getting ready to IPO at almost $2 trillion, money losing is irrelevant at this point.
Delivery Hero then confirmed it had received a formal offer from Uber at €33 per share, a price with literally zero premium to the prevailing share price. Uber offered to buy a company for exactly what it was already worth, apparently on the theory that the owners might not notice, but of course, they did.
Several institutional investors made clear they wouldn’t engage below €40 per share. Delivery Hero’s stock promptly traded up to a 52-week high of €39.79. The market had opinions.
Then on May 27, Uber bought Aspex Management’s stake at just under €40 per share, paying 21% more than its own public offer, in the secondary market, on the same day it was officially bidding €33. Uber’s public offer: €33. Uber’s actual transaction price: €40.
This is unusual. Most acquirers at least wait a few days before paying their own target price while publicly offering less.
The internal ceiling Uber has reportedly shared with advisors is €38. It has already paid €40 once. Nobody on Delivery Hero’s shareholder register is confused about where this settles.
Delivery Hero operates in 70+ markets (Middle East, Asia, Europe) under brands including Talabat, foodpanda, and Hunger Station. In most of those markets, Uber Eats either doesn’t exist or is losing. DoorDash has been expanding internationally and is circling the same assets.
If Uber doesn’t buy Delivery Hero, DoorDash might. And if DoorDash buys Delivery Hero, Uber’s international food delivery business gets substantially more complicated.
Prosus, a Dutch tech investment firm, has been ordered by the EU to reduce its Delivery Hero stake below 10%, which means a large block of shares is coming to market. Uber can buy that block and strengthen its position. So can DoorDash. Whoever moves faster controls the process. But it’s also European summer, so fast means by September.
Delivery Hero CEO Niklas Östberg is stepping down in Q1 2027 following pressure from major shareholders for a strategic review. A company in strategic review with a 36.83% anchor shareholder and a departing CEO has, functionally, already started a sale process and identified its own synergies.
The question is what the price is. The current answer from Delivery Hero shareholders: at least €40.
NEWS ROUNDUP
Top Reads
NEWS ROUNDUP
Which Bulge Bracket has the Best Mental Health?
Now you can find out. We just rolled out mental & physical health league tables on Wall Street 360, the first time the Street has ranked firms head-to-head on culture, not just comp.
Filter by:
Bulge Brackets
Elite Boutiques
Middle Market
Private Equity
Country, city, or industry
The results are eye-opening. Some firms with the loudest reputations rank middle of the pack. Others you wouldn't expect are quietly outperforming. And a few elite boutiques are putting some bulge brackets to shame.
What'd you think of today's newsletter? |