What Do You Have to Believe (WDYHTB)
Reverse-engineer any bullish narrative into the specific, checkable assumptions that must be true for it to pay off.
- Difficulty
- Easy
- Time to result
- ~weeks to results
- Steps
- 4
- Confidence
- —
Whenever a company or market presents an exciting narrative — a huge TAM, a revolutionary technology, a bold valuation — this framework asks the analyst to work backward from the story to the concrete conditions that must hold true. Instead of accepting a top-down market-sizing slide (like 'we'll capture 1% of a $100T market') at face value, you isolate the real, falsifiable claims underneath it — such as 'real user value is being created by generative AI applications' — and then look for hard evidence (revenue, usage, retention) supporting or undermining those claims. This turns vague optimism or fear into a short, testable list that can be revisited as new data arrives, which is exactly how the hosts evaluate whether Nvidia's trillion-dollar data-center thesis is credible versus merely a hopeful narrative.
Origin
Extracted from Acquired
How to run it
- 1
State the Narrative
Write down the bullish (or bearish) story exactly as presented, including any market-sizing math.
- 2
Extract the Load-Bearing Assumptions
Identify which specific claims must be true for the narrative's conclusion to hold — not general vibes, but concrete facts.
Pro tip Ask 'what happens to this thesis if this one assumption is false?' to find which assumptions are truly load-bearing.
- 3
Find Real Evidence
Look for hard data — revenue, usage numbers, run-rates — that support or contradict each assumption.
Watch out Beware of evidence that is itself circular, like rising stock price used to justify the narrative that drove the stock price.
- 4
Weigh Probability and Monitor
Assign a rough confidence level to each assumption and set a plan to revisit as new quarterly data or product evidence emerges.
In the wild
Nvidia claimed a $1T total addressable market by assuming they'd capture 1% of $100T of global industries like autonomous vehicles, robotics, and the Omniverse.
→ Years later most of those specific markets (autonomous vehicles, robotics) hadn't materialized as projected, showing the original narrative's assumptions were largely wrong even though the company still hit a trillion-dollar valuation via a completely different path (generative AI data centers).
Common mistakes
Accepting top-down TAM math
Blindly trusting a 'we'll capture X% of a huge market' slide without checking whether the underlying use cases are actually materializing.
Treating narrative and evidence as the same thing
Mistaking excitement or spend commitments for proof that end-user value is actually being created.
Is it for you?
Best for
Evaluating high-growth or hype-driven markets, big valuation jumps, or bold new technology narratives before committing capital.
Not ideal for
Simple, low-uncertainty decisions where the causal chain from investment to outcome is already well understood.
From the transcript
“whenever someone paints a picture you say okay what do I have to believe the thing you have to believe is there is real user…”
“there was some stuff on the slide that was fairly speculative you know like autonomous vehicles and the Omniverse and I think robotics were a…”
From the episode
Nvidia Part III: The Dawn of the AI Era (2022-2023)