Actionable insights — Portfolio Update: UBER

1. Buy a great business the market is pricing for an existential threat

1. Start from quality: a business "one of the world's great" — large, durable free cash flow (~$10B) and fast earnings growth (+44% YoY). Don't apply this to a marginal name; the contrarian edge only pays when the underlying business is excellent.
2. Confirm the price embeds an existential fear, not a normal stumble — UBER at ~3.4x sales / 18x FCF "as though it faced existential threats." A great business priced for survival risk is the setup.
3. Size conviction by crowd disagreement, Druckenmiller-style: "I like it when I have extreme conviction in a thesis and no one else believes it — it gives me even more conviction." Lonely + right is the goal; popular + right pays little.

• Excellent businesses trading at a survival-risk multiple while one scary narrative dominates the tape; near-unanimous bearishness (or, as a sanity check, near-unanimous analyst Buys against a falling price — a tell the fear is narrative, not numbers).

2. Invert the consensus threat — prove the feared force is actually a tailwind

1. Name the one narrative doing the damage and find the unit-economics fact that flips it. AVs are "framed as an existential threat; the operational data proves the opposite" — an idle self-driving car earns nothing, so utilization is the economics, and Uber's demand network lifts a Waymo car's utilization 30%.
2. Carry the inversion into the P&L, not just the narrative: removing the ~70%-of-fare driver cost lets take rate climb from 28% toward 80% — ~$20B incremental revenue at 20% AV penetration on $193B bookings. The feared force makes the business more profitable.
3. Pressure-test the other bear narratives the same way (Delivery Hero "dumb money" → a Gulf super-app; AI-spend cuts read as distress → discipline at $9.8B FCF). A thesis survives when each headline fear has a quantified rebuttal.

• A consensus threat where the disrupting technology actually raises utilization, lowers a variable cost you keep, or needs your distribution to work — that's the inversion signal.

3. When you can't pick the winner, own the toll booth on all of them

1. Identify a fast-changing arena with many credible contenders and no clear winner (the AV makers: Waymo, WeRide, Pony.ai, Baidu, Wayve, Waabi…).
2. Find the choke-point every contender needs — here, demand aggregation and the white-label plumbing (fleet management, training data, remote ops, AV insurance via Uber Autonomous Solutions).
3. Spread bets across all of them (14+ partners, $10B committed) so the return is "we don't care who wins" — a cut of every autonomous trip rather than a bet on one platform.

• Platform fights where a neutral aggregator or infrastructure layer captures economics from every player; a proprietary, compounding asset (Uber's training data from 40M daily trips) that deepens the moat the more it's used.

4. Separate "the metrics improved" from "the price fell" — and trust the metrics

1. After entry, re-underwrite on fundamentals, not the quote: list every operating metric that existed at recommendation and check the direction. "Every financial metric that existed when we made the recommendation has improved."
2. Watch the second-derivative tells of a strengthening model — earnings scaling at >2x revenue growth, a subscription tier (Uber One, 46M, half of bookings) and high-margin ads (>$2B, +50%) compounding faster than the core.
3. If fundamentals improved while the price fell, the gap is opportunity, not warning — add or hold, don't capitulate to the tape.

• A widening divergence between improving fundamentals and a falling price; operating leverage (earnings growing materially faster than revenue) confirming the model, not just the story.

5. Value the pieces — get the optionality for free

1. Build a sum-of-the-parts / DCF that separates the proven core from the speculative upside: ex-AV, UBER's platform + advertising alone is worth ~$95–108 at conservative peer multiples.
2. Compare that floor to the current price (~$70): if the boring, provable pieces already exceed the quote, the contested upside (the AV "toll booth") is a free call option.
3. Sanity-check against independent marks (consensus target $104.45, Morningstar DCF $97, full SOTP $110–130) so the conviction isn't only your own model.

• Cases where a conservative valuation of the certain pieces already covers the price — that's the asymmetric setup, with the contested growth as the upside you didn't pay for.

6. Ring the register when a long-held name hits a multi-decade-high valuation — and raise cash on ominous tape

1. Revisit winners on valuation, not nostalgia: a name you bought cheap that's run ~50% and now trades at its highest P/S and P/E in 25 years has lost the reason you owned it.
2. Let the broader tape set the cash level: with the overall market flashing "ominous" signals, prefer raising cash and tracking fewer, higher-conviction names — sells, not just buys.
3. If you'd rather stay invested, rotate toward the cheaper, higher-conviction idea (UBER at 18.6x 2026E vs IBM's 22x) — but remember "cash is definitely not trash."

• A long-term holding at a multi-decade valuation extreme; market-breadth/risk signals that argue for carrying cash; the discipline to make a rare outright sell rather than a passive hold.