Bill Ackman: Here's What the Market is MISSING
Bill Ackman (Pershing Square, ~**$25B AUM**) on All-In. The throughline — and the answer to the title — is a **mispricing call on the quality megacaps**: while short-term capital crowds into *'the new new thing… chips and semiconductors and energy,'* the durable compounders get left for dead. **Ackman owns Microsoft, Meta and Amazon and says they are 'undervalued' / 'really cheap,'** drawing the 2000 analogy where *'Berkshire Hathaway traded at the lowest valuation I think it ever traded at in its history'* as the dot-com crowd dumped *'old stuff.'* This is the **direct inverse of [Dan Loeb's 'SOX is up 40%' regime call](/issues/2026-05-31) and [Gerstner's semis/memory cohort](/issues/2026-05-31)** — same tape, opposite trade. He times it to a recent table-pound: *'stocks just got crazy cheap… of really high quality companies.'* On the **private-AI complex** (the cohort theme this week — [OpenAI's Sarah Friar](/issues/2026-06-07), [Laffont's $4T IPO wave](/issues/2026-06-07)), Ackman underwrites **SpaceX, Anthropic, OpenAI and Palantir 'as venture'** — *'people, opportunity, context, deal'* — noting they're *'not seed or series A… they're D or E.'* He's **personally in an xAI/SpaceX SPV** (via Ron Baron) and flags **SpaceX at a 'trillion / 750 billion'** valuation as plausibly *'the lowest cost of capital equity transaction in the history of the world.'* The bear note on OpenAI is precise and rhymes with the [ROI-reckoning thread](/issues/2026-05-31): the model is *'spending, making capital commitments… massively in excess of revenues,'* a *'degree of difficulty I would say is hard.'* On the **SaaS apocalypse** he bifurcates — *'I worry more about a salesforce'* and niche **$30k-a-seat** software than Microsoft at *'50 bucks a seat.'* On AI ROI he is blunt — every CEO's **number-one** worry, the McKinsey *'95% of enterprise initiatives fail'* stat goes unchallenged, and Pershing's own use is just *'legal… compliance, back office'*: *'we're still super, super early.'* The structural project is **Howard Hughes as 'Berkshire 2.0'** — a **$4B market cap** real-estate holding being recapitalised *'within the next week or so'* into an insurance compounder (float in treasuries, surplus in equities) with a stated **trillion-dollar** ambition.
Key points
- **The title's answer: the market is mispricing quality megacaps because capital is chasing 'the new new thing.'** *'What's interesting about markets is people always bring their eye to the new new thing. And the new new thing is sort of chips and semiconductors and energy. And that's where the shorter term capital is going. What tends to happen is really high quality things get left behind.'* He names the longs — **Microsoft, Meta, Amazon** — as *'old fashioned companies in… the OpenAI world'* and confirms *'yes'* they're undervalued. **This is the single load-bearing signal, and it is the direct contrarian of [Loeb's 'SOX +40%' semis-regime call](/issues/2026-05-31) and [Gerstner's memory cohort](/issues/2026-05-31): same megacap-vs-semis tape, opposite side.**
- **The 2000 analogy is the frame, not a bubble call.** *'I was there in 2000 in that sort of bubble… people got excited about Internet stocks and Berkshire Hathaway traded at the lowest valuation I think it ever traded at in its history… a similar thing is happening today in a sense to Amazon and Meta, Microsoft.'* He explicitly hedges — *'this is different, I'm not saying this is'* — so the trade is **rotation into left-behind compounders, not a short-the-bubble thesis.** Pair with [Laffont's '$4T AI IPO wave'](/issues/2026-06-07) and last issue's *'100x sales, I wouldn't buy a share'* — the bulls and Ackman agree the heat is in the new names, not the old ones.
- **The recent table-pound, dated and sourced to valuation.** *'Stocks just got crazy cheap, just incredibly cheap of really high quality companies.'* His mental model is reflexive: *'valuation is like a tether… when stocks get too cheap, the rubber band is actually pulling valuations up,'* and calling it publicly *'causes people to have kind of a psychological reset.'* **A live, repeatable Ackman pattern (he ties it back to his March-2020 CNBC moment)** — the signal isn't a number, it's that he chose to broadcast the long.
- **Disruption risk has structurally risen — the core of his underwriting work.** *'The most important and most challenging thing to do is determine what's the risk of disruption. What's the risk of two guys, two women from Stanford in a garage… That risk I think has gone up dramatically. This is the greatest era in history to build a business… unlimited access to compute… capital… talent, which means that the probability of your being disrupted has gone up enormously.'* **The same defensibility anxiety [the 20VC/Mercor 'application-layer has no defensibility' episode](/issues/2026-06-07) is built on** — Ackman's response is to pay up for non-disruptible quality, not to chase the disruptors.
- **He underwrites the private-AI complex 'as venture' — and the rounds are late, not early.** *'I think you underwrite a SpaceX the way you underwrite a venture capital investment'* on *'people, opportunity, context, deal.'* On Anthropic / OpenAI / Palantir: *'they're not seed or series A… they're D or E, but they're still like venture investments. These companies have proven they can generate a lot of revenues.'* **This is a buy-side framework for valuing exactly the cohort going public this week** ([Friar's OpenAI](/issues/2026-06-07), [Laffont's IPO wave](/issues/2026-06-07)) — treat them as concentrated venture bets on founders, not as DCFs.
- **The OpenAI bear note is specific and rhymes with the ROI reckoning.** *'You get a company that's spending, making capital commitments, they're massively in excess of revenues. And how do you do that… it's degree of difficulty I would say is hard.'* He praises [Sarah Friar's](/issues/2026-06-07) capital-commitment discipline as the missing message — *'if I were OpenAI, I would be getting that message out.'* **Continues [the enterprise-token-spend reckoning from Issue 07](/issues/2026-05-31): the spend is real, the proof isn't — now applied to the labs' own capex, not just their customers'.**
- **The SaaS apocalypse is a bifurcation, not a wipeout.** *'I worry more about a salesforce… if you're a software company today you have to be as AI enabled as you can. There have been sort of monopolistic type profit taking off of customers… a niche software product charging 30,000 a year… those companies are really at risk. Microsoft, when the average customer is paying 50 bucks a seat… that platform is worth a lot more.'* **The tradeable cut: short high-ASP niche SaaS, own low-ASP platform incumbents** — a sharper version of the model-commoditisation worry running through the cohort.
- **AI ROI is real but unproven — every CEO's number-one worry, and Pershing's own use is trivial.** He endorses the [Issue-07 McKinsey stat](/issues/2026-05-31) unchallenged — *'95% of enterprise initiatives actually fail'* — and on adoption: *'every CEO in America today is like, how do I use AI… on the hierarchy of things they worry about, it's probably number one as both an opportunity and a threat.'* Yet Pershing's own use is *'really on the legal side and… compliance, back office'* and *'we're still super, super early in terms of big companies using AI effectively.'* **A blunt demand-side reality check from someone sitting on Fortune-500 boards.**
- **Howard Hughes is 'Berkshire 2.0' — a real-estate-to-insurance recap with a trillion-dollar ambition, recapitalised this week.** *'The company's got like a 4 billion market cap and the goal is to build it into a trillion dollar thing over time.'* Mechanics are pure Buffett: *'100% of the insurance float… in short term treasury… 100% of the surplus… in common stocks,'* funded by redirecting Howard Hughes' real-estate cash flow into insurance *'within the next week or so.'* **Directly relevant to a CRE-investor reader: Ackman is explicitly repricing a land/development balance sheet (high assumed cost of capital) into an insurance compounder (low cost of capital) — the same reflexive cost-of-capital lever he credits Elon with, where SpaceX at a 'trillion / 750 billion' would be 'the lowest cost of capital equity transaction in the history of the world.'**
Notable quotes
What's interesting about markets is people always bring their eye to the new new thing. And the new new thing is sort of chips and semiconductors and energy. And that's where, you know, the shorter term capital is going. What tends to happen is really high quality things get left behind.
I was there in 2000 in that sort of bubble. This is different. I'm not saying this is, but there's some analogies and the analogies are people got excited about Internet stocks and Berkshire Hathaway traded at the lowest valuation I think it ever traded at in its history. As people said, okay, that's all old stuff. I think a similar thing is happening today in a sense to Amazon and Meta, Microsoft.
Stocks just got crazy cheap, just incredibly cheap of really high quality companies, right?
I think you underwrite a SpaceX the way you underwrite a venture capital investment.
What's helpful is they're not seed or series A. Right. They're D or E, but they're still like venture investments. These companies have proven they can generate a lot of revenues.
You get a company that's spending, making capital commitments, they're massively in excess of, you know, revenues. And how do you do that and get, you know, it's degree of difficulty, I would say is hard.
I think there have been sort of monopolistic type profit taking off of customers. When someone had a kind of a niche software product, they're charging, you know, 30,000 a year or something like this. I think those companies are really at risk. You know, Microsoft, when the average customer is paying, I don't know, 50 bucks a seat or some small number, that platform is worth a lot more.
The company's got like a $4 billion market cap and the goal is to build it into a trillion dollar thing over time.
If SpaceX goes public at a trillion $750 billion, it'd probably be the lowest cost of capital, equity capital transaction in the history of the world.
Themes
- Megacap mispricing vs the new new thing
- Private-AI underwriting as venture
- AI capex vs ROI reckoning
- SaaS bifurcation under AI
- Insurance-float compounding (Berkshire 2.0)
Mentioned
People
Companies
Ideas
- the new new thing vs left-behind quality
- megacaps undervalued (MSFT/META/AMZN)
- 2000-bubble / Berkshire-cheapest analogy
- underwrite private AI as venture (people/opportunity/context/deal)
- D/E-round 'venture' valuation
- capital commitments massively in excess of revenues
- SaaS bifurcation (niche $30k-seat at risk vs $50/seat platforms)
- AI ROI super-early / 95% of initiatives fail
- Howard Hughes = Berkshire 2.0 insurance flywheel
- cost-of-capital reflexivity / army of believers
- valuation-as-rubber-band + public table-pound