◢ ALL-IN ALPHA research only · never trades · 2026-09-10
BI

Bill Gurley

regular guest ×1.0 · 14 ideas engaged · 4 episodes · E44 → E275
ChamathJasonSacksFriedbergAaron LevieBen ShapiroBrad GerstnerCyan BanisterElon MuskGavin BakerJoe LonsdaleKeith RaboisRyan PetersenThomas LaffontTravis Kalanick
0.62
credit over 13 calls how this is scored
7/2/4
right / half / wrong
+28%
avg α on scored calls α
1
live positions
+0
lifetime conviction force · 0 ideas born
2.21
avg strength · 0 dissents

Does his own confidence mean anything?

the same scored calls, split by how hard he was pushing at the mention that set his final stance. If credit rises with strength, his table-pounding carries information and you should weight it. If it's flat or inverted, his confidence is noise — treat a ×3 like a ×1. Cells under 5 calls are provisional and get no weight in the replay.

he wascallscreditright/half/wrongavg αread
×3 table-pounding 5 0.4 2/0/3 -5% worse than coin flip
×2 stated with reasoning 7 0.71 4/2/1 +45% better than coin flip
×1 offhand aside 1 1.0 1/0/0 +83% provisional (n<5)

Which kind of claim to trust

by the tier of the mention that set his final stance — a dated prediction is a different animal from a passing lean, and they don't have to score alike

tiercallscreditright/half/wrongavg α
explicit_prediction 5 0.8 4/0/1 +45%
sentiment 8 0.5 3/2/3 +18%

What he's actually good at

his row of the scoreboard — these are the exact cells rescore.py uses to weight his mentions. w = clamp(2 × credit, 0.3, 1.5), applied only at n ≥ 5.

🌍 macro n=1
1.0 provisional 1/0/0
🪙 crypto n=1
1.0 provisional 1/0/0
📈 equities n=9
0.56 w=1.11 4/2/3
🤖 tech n=2
0.5 provisional 1/0/1

Live book — 1 position

his latest stance on every ACTIVE idea. Expression = what he is effectively long or short: supporting a bearish idea is a SHORT, opposing one is a LONG.

ideahis stanceexpressionconvictionflageval in
🤖 Anthropic IPO (~Oct 2026) is attractively priced at $1.5-2T CONTESTED support ×1 ▲ LONG GOOGL 53.9 EARLY 177d

Where his book points

net push per primary instrument across his live stances — conviction × his agreement × the idea's direction. This is his implied book, not a position he disclosed.

GOOGL
+53.9 long

Best and worst calls

TOP α — his stance was right and beat the index
+165% HIT 🤖 Generative AI is the next Silicon Valley bubble cycle
+85% HIT 🌍 Fed ending QE drains liquidity-dependent assets in 2022
+83% HIT 🪙 Crypto bubble bursts in 2022 — 90% of projects blow up
+70% HIT 📈 2021 is the peak of the risk-capital golden era
+54% HIT 📈 The negative-unit-economics growth playbook dies with free capital
WORST α — stated straight, because the ledger doesn't flatter anyone
-58% MISS 📈 Private venture marks are stale by design — nobody is paid to fix them
-32% MISS 📈 Down-round IPOs wipe out the late-stage preference stack
-27% PARTIAL 📈 Prop 22 IC+ model becomes the national gig-work architecture
-15% MISS 🤖 A decade of zero rates under-trained a generation of operators
-11% MISS 📈 Tech drawdown is in the eighth inning - the bottom is close

How he argues

STANCE MIX
support 14
TIER MIX
sentiment 9
explicit_prediction 5

Latest from him

support ×1 sentiment 🤖 Anthropic IPO (~Oct 2026) is attractively priced at $1.5-2T E275 · 2026-05-29 ▶ 49:10
Claude is really good at product. Like Claude for Excel is better than Copilot. Not by a little, by a lot. So anyone that's going to run against them, they are a worthy foe.
support ×3 sentiment 📈 Private capital concentrates in a few mega-GPs that take themselves public E141 · 2023-08-11 ▶ 25:19
I think a lot of people realize that if you can get 2.5% or 3% management fee investing $300 million at a pop, that's an easier lifestyle than actually taking board seats and doing work. And so I think a lot of money and activity got pulled into the late stage market. Nearly every firm started doing that.
support ×2 sentiment 🤖 Generative AI is the next Silicon Valley bubble cycle E141 · 2023-08-11 ▶ 24:00
the AI deals, one thing that might be interesting is if you pulled the AI deals out, I bet those numbers would be more akin to what they were two or three years ago
support ×3 sentiment 📈 Private venture marks are stale by design — nobody is paid to fix them E141 · 2023-08-11 ▶ 36:24
The other thing is many LPs are actually bonus on the paper mark. And this is something that a lot of people don't realize. And so they don't have an incentive to dial around to the GPs and say, get your marks right, because it's actually going to reflect poorly on them if they were to roll those up.
support ×3 explicit_prediction 📈 Down-round IPOs wipe out the late-stage preference stack E141 · 2023-08-11 ▶ 52:32
because of hyper competition and investing in 99, 2020, 2021, there was a term removed from most term sheets that gave investors the right to protect their LICPREF on an IPO. That's gone in most of these cases. So you could convert LICPREF under, which for a founder or an early stage angel investor would be a huge win.
support ×2 sentiment 📈 The IPO window reopens - but only at a real discount to 2021 marks E141 · 2023-08-11 ▶ 53:58
There's two things. I think that cleaning up complexity is a great reason for the public market. And Brad already said, we've seen a massive recovery in software stocks. Like, the marks are better than they were two years ago.
support ×3 sentiment 🤖 A decade of zero rates under-trained a generation of operators E141 · 2023-08-11 ▶ 1:20:10
they lived and grew up in a day and age where they were told growth at all costs. And it's super hard culturally to go from that type of execution to the principal type execution you guys have been promoting over the past several months. It's just hard. It's not impossible, but it's very, very hard.
support ×2 explicit_prediction 🌍 Fed ending QE drains liquidity-dependent assets in 2022 E81 · 2022-05-23 ▶ 4:08
we basically had a mini pullback in March of 2020, but then the Fed hit so hard that things just blasted off again. ... Now, you guys have talked about this, but that tool is not in toolbox anymore.
support ×3 explicit_prediction 📈 2021 is the peak of the risk-capital golden era E81 · 2022-05-23 ▶ 1:13
It's more like a sawtooth. It risks on is a very slow process and it's reflexive. So it grows and grows and grows and grows. And then risk off tends to be very abrupt. ... This cycle, risk on was from 2009 to five months ago. ... And risk off is five months. And the thing that's really tough about that is it requires mental adjustment very quickly
support ×2 explicit_prediction 📈 Late-stage private marks reset 30-60% — down-round IPOs become the norm E81 · 2022-05-23 ▶ 3:40
1 was very abrupt, and we didn't really start to see liquidity again until, with a few exceptions, Elon mentioned PayPal, but like 05, 6 ... it's usually a long window on the other side
support ×2 sentiment 📈 Tech drawdown is in the eighth inning - the bottom is close E81 · 2022-05-23 ▶ 49:36
I'm very excited about public stocks here, actually. ... Yeah. Like the valuations are getting super interesting.
support ×1 sentiment 🪙 Crypto bubble bursts in 2022 — 90% of projects blow up E81 · 2022-05-23 ▶ 12:54
If you deployed two-thirds of your fund into crypto assets with no board seat in the past 12 months, are you going to call Harvard and Penn and say, hey, I need some more right now? I don't think you're going to make that call.
support ×2 explicit_prediction 📈 The negative-unit-economics growth playbook dies with free capital E81 · 2022-05-23 ▶ 45:30
It all depends on whether you can reign it back in or not. ... I think if 50 entrepreneurs try that trick, 49 are going to auger in.
support ×2 sentiment 📈 Prop 22 IC+ model becomes the national gig-work architecture E44 · 2021-08-28 ▶ 1:06:01
There's one big issue that I don't think is talked about enough, which is, if you poll the drivers, they're not looking for any changes. They're really happy with the flexible work product. If you look at the voters of California, they stepped up and voted and made it very clear.