Uber: the business thesis played out, the stock round-tripped to $75, and the margin of safety mostly closed. Is it a buy here or not?

Uber is the cleanest example I've found of the question value investors actually wrestle with: what do you do when a genuinely great business is trading at a merely *fair* price?

Quick context for the setup. Bill Ackman's Pershing Square began building an Uber position in **Q1 2025** — the quarter it first appears in their 13F — accumulating while the stock traded in roughly the **low-$60s to mid-$70s**, then going public on Feb 7, 2025 with a **\~$2.3B stake (30.3M shares)** that instantly became his largest position. He called Uber *"one of the best managed and highest quality businesses in the world,"* trading at *"a massive discount to its intrinsic value."* (13Fs don't disclose cost basis, so any exact average is an estimate — but that entry window sits right in the value band this framework flags.) And he's stayed: as of the Q1 2026 13F, Uber is still a **top-3 position at \~30M shares**. Since then the business has delivered, the stock ran to \~$100 last October, and it has round-tripped back to \~$75.

So I ran **today's** Uber through a strict, Buffett/Graham-style framework — four lenses: **Meaning, Moat, Management, Margin of Safety** — then leaned on the AI behind it for the parts a scorecard can't settle. Full teardown below.

**The moat: genuinely wide, and strengthening**
Uber's moat is a network effect compounded by efficient scale — more riders pull in more drivers, which cuts wait times, which pulls in more riders. The two-sided Mobility + Delivery structure means Uber Eats' courier density partly subsidizes customer acquisition for the whole platform.

The financials confirm the network is now *converting to cash*, which is the real test of whether a platform moat is durable or illusory. Annualized growth rates across the windows:

**Revenue:** \~25–31%/yr across the 5/7/10-year windows — remarkably steady
**Operating cash & FCF:** accelerating hard — Op. Cash +146% and FCF +193% in the latest 3-year window, the crossover from scale-*investment* to scale-*harvesting*

**10-year compounded growth: \~50%/yr vs. \~13% for the S&P**
Every growth axis scores at or near the top — Revenue, Equity, Operating Cash, and FCF all max out at 100. The one that stands out the other way is **earnings, scoring 75**, dragged by 1-year earnings growth of just \~4% — worth flagging honestly, because it's the single metric that lags all the others and bears watching (cost pressure, or one-time items?). Still, four of five moat axes pinned at 100 and the fifth at 75 is broad-based strength, not one fragile pillar.

**Honest soft edges:** Lyft (US), DoorDash/Instacart (delivery), and Grab/DiDi (international) all compete directly, and driver loyalty is thin because most drivers multi-app. The moat is real but not impenetrable. The single thing I'd watch is **take rate** over the next few quarters — if it holds or expands while gross bookings grow, pricing power is intact; if it compresses, the moat is being tested where the historical data can't yet show it.

**Management: the honest part (and where a framework earns its keep)**
This is the screen that separates a disciplined process from cheerleading.
The turnaround under Dara Khosrowshahi is **real and visible in the numbers**:
Return on equity: **−24% (10-yr avg) → +37% (last year)**

Net-debt leverage fell from **2.30x → 0.71x**, and net debt / FCF is **0.47x**(leverage risk basically off the table)

Buybacks have *started* — notable for a company that spent most of its life diluting shareholders
But here's what the framework refuses to wave away:

**ROIC is only \~11%** — improving, but it hasn't crossed the \~15% line that signals consistent wealth compounding. On a 10-year average, returns on capital are still dragged down by a decade of capital destruction.

The management quality score lands at **"good, not great"** — competent and clearly improving, but **stock-based comp dilution is still meaningful**, and insider ownership is modest.

When I pushed the AI on whether the turnaround makes this a "buy the management" story, it didn't cheerlead — it flagged that Uber is earning above its cost of capital *only recently*, so the jury is still out on whether that discipline is structural or cyclical. That's the read I trust: *the balance sheet is fixed; this is not yet a proven compounder.* Whether the new buybacks are sustained and accretive — versus a one-off — is the real test of whether management has shifted its shareholder orientation. The framework scores the decade, not the highlight reel.

**Margin of safety: this is the close call**
Three price anchors, increasingly conservative:
**Intrinsic value:** \~$120/share
**Earnings-payback price:** \~$74.67
**Strict 50%-discount buy price:** \~$60

At \~$75, the stock sits **right on top of** the earnings-payback anchor — barely **$0.30 above it** — and roughly 25% above the strict $60 buy price. So it has cleared the most constructive of the three zones, but not the two stricter ones. The verdict lands at **"Fair" margin of safety**: real, but thinner than the textbook 50% discount.

One thing I appreciate: the framework doesn't extrapolate peak growth. Uber compounded \~50%/yr over the past decade, but the model **caps the forward growth assumption at 15%**— so the \~$120 intrinsic estimate is built on a conservative rate, not a heroic one. That built-in conservatism is *why* the value anchors sit where they do.

This is exactly where reasonable value investors split, and I don't think there's one right answer:
**Some are fine buying here** — it's essentially *at* a legitimate value anchor (under 1% above), the business quality is high, and holding out for a perfect entry can mean never owning a compounder at all.

**Some want all three anchors to line up**before committing, and would rather pass than bet on a stock that may never trade back to the $60s.
**Many split the difference and scale in via tranches** — a starter position near this level, adding on weakness toward the stricter prices — so you're neither chasing nor sitting out.

What the framework won't do is pretend a fair price is a screaming bargain. It's saying: good business, roughly fair value, thin margin — size and pace your entry accordingly.

**The wildcard nobody's framework can price: autonomous vehicles**
A scorecard is useless here — so this is where I stopped reading numbers and started arguing with the AI behind the tool, which reasons over the actual figures. Instead of a verdict, I had it steelman both sides.

**The bull read:** the Waymo/WeRide partnerships position Uber as the *demand-aggregation layer*rather than the vehicle operator — potentially eliminating its single biggest cost (drivers) while the network effect stays entirely intact.

**The bear read:** the mirror image — Waymo or Tesla bypass Uber entirely and build direct-to-consumer robotaxi networks, attacking the demand side of the moat in exactly the high-density markets that drive the economics.

Then the genuinely useful part. I asked it what single number would tell me which way this is actually breaking — and instead of hedging, it pointed at the **take rate** over the next four quarters. If it holds or expands while gross bookings grow, pricing power is intact; if it compresses under driver-incentive pressure or discounting, the moat is being tested in the one place the historical data can't yet show.

That's the judgment call no model prices. Having something grounded in the real numbers argue both sides and then hand me the one metric that settles it is a lot more useful than a rating — but this part's on you to underwrite, not the spreadsheet.

**The takeaway**
A framework that only flags what's *good* is useless. The useful signal here is the distinction it forces: Uber is clearly a good *business* (wide moat, real turnaround, cash finally compounding), and at \~$75 it's a *fair-to-okay price* — right at one value anchor, short of the stricter two.

Whether that's a buy is a conviction-and-sizing question, not a yes/no: full position, starter tranche, or wait for a deeper discount. And the biggest swing factor — autonomous vehicles — is something you have to underwrite yourself, because no model prices it.

*(Standard caveat: I'm sharing a way of reasoning about it, not a track record or a recommendation. Do your own work.)*

**Disclosure:** the tool I used is one I built myself — an app called Moatly. It runs any stock through this exact 4M framework in about 30 seconds, but the part I actually lean on is the AI mentor on top of the numbers: you can interrogate it, throw rebuttals at it, and test scenarios, all grounded in the calculated data. A study partner that argues back, not a stock-tip bot. I'm not here to pitch it — the analysis stands on its own and I mostly want the debate. I've kept the link out of the post; I'll drop it in the comments for anyone who wants it.

**So where do you land at \~$75?** Buy near the value zone now, start a tranche and add on weakness, or hold out for the stricter $60s? And how are you underwriting the AV question — free call option, or the thing that eventually commoditizes the platform? (Name a ticker and a thesis, and I'll post what the tool's AI fires back at it.)