[h] home [b] blog [n] notebook

what is Nvidia worth without Jensen?

the interesting financial question about Jensen Huang is what Nvidia would be worth if he left. we have a stock market, an options market, analysts with spreadsheets, and an enormous amount of commentary. getting a separately traded market answer to that specific question is still surprisingly difficult. i think that says something about what our markets are missing.

Nvidia’s board has announced an independent investigation following reports that Jensen was targeted by a foreign intelligence effort. its statement says it has no reason to believe national security or company secrets were compromised, and that it retains confidence in his leadership. an investigation is not a finding. departure is a hypothetical here, not an announced outcome. the financial question is worth asking without pretending otherwise.

start with two investors. one thinks Jensen is almost certain to stay, but that his departure would be disastrous. the other thinks departure is plausible, but that Nvidia would do fine without him. both can arrive at the same valuation. they agree about the stock price while disagreeing about the thing everyone is discussing. watching the stock cannot tell you which investor you are trading against.

this is because a share bundles possible futures together. leadership, demand, margins, competition, export policy, interest rates: all of it goes into one price. that is appropriate for an asset you own across those futures. it becomes a problem when you want to extract a particular belief from the bundle. a price can aggregate information without revealing how the information fits together.

the conditional market version is straightforward to describe. what is Nvidia’s stock worth at the end of September 2027, conditional on Jensen leaving the CEO role by June 30? alongside it, ask the same question conditional on him remaining. someone who understands Nvidia’s organizational durability could trade the departure scenario without first needing a strong view about whether the board will actually replace him.

suppose those markets imply $140 if he leaves and $180 if he stays. these are invented numbers, not price targets. the useful thing is that they separate two disagreements: how likely departure is, and what the business looks like in that world. a market asking only whether Jensen leaves gives you the first. the stock bundles both together with everything else investors care about.

this makes “Jensen is irreplaceable” a more demanding claim. irreplaceable over what horizon? does the stock lose 5 percent on announcement, or does Nvidia permanently lose its ability to make good bets? Nvidia without Jensen would still have engineers, software, products, and customer relationships. a large departure discount needs an account of what those assets lose without him. perhaps it is coordination, customer confidence, or judgment about which projects deserve another year.

the opposing view deserves a market too. perhaps investors give the most visible person too much credit for an institution’s capabilities. perhaps a successor would preserve the important things. it is easy to say either version in a research note. i want the person convinced that Nvidia survives its founder to be able to trade that view directly against someone convinced that it does not.

“just buy puts” is a reasonable response, but an ordinary put pays according to the stock price regardless of what caused the decline. you are buying protection against weaker demand, a market selloff, and other things alongside leadership risk. that may be exactly the protection you want. it is still different from exposure conditioned on a particular event. conditional instruments add specificity; they do not make options obsolete.

there is a portfolio reason to care. an investor might be comfortable with Nvidia’s commercial prospects but unwilling to carry a particular leadership scenario. today, reducing that concern often means reducing the whole position. a conditional hedge could let them retain more of the exposure they actually want. it would also make research more useful: expertise about succession could earn a return without requiring equal expertise about the next earnings report. that is a modest claim compared with automating corporate governance, but it is a much more immediate use case for real investors.

the construction matters. a claim paying Nvidia’s future price if Jensen leaves, and zero otherwise, does not by itself quote the conditional valuation. its price also reflects the likelihood of departure. you need appropriate paired claims or conditional settlement rules to separate those quantities. even then, collateral, risk premia, and liquidity affect prices. calling something a conditional market does not make every displayed number a pure forecast.

there is a more important limitation. the difference between the two branches is not automatically Jensen’s causal contribution. a planned succession after reassuring news is different from an abrupt exit following damaging findings. traders in the departure branch will price the circumstances they expect to accompany departure. the condition selects a whole set of worlds, including the other things likely to be going wrong in them.

this is the causal-correlational gap i have written about before. observing that Jensen leaves is different from intervening to remove him while holding everything else fixed. a shareholder may care about the entire departure scenario, so the conditional forecast is useful anyway. a board asking whether it should replace him faces a different problem. it cannot read the spread between branches as the value created by keeping him.

resolution is also less boring than it looks. does becoming executive chairman count as leaving? does temporary leave count? what if resignation is announced before the deadline but takes effect afterward? measuring the first close after departure asks a different question from measuring the stock three months later. these details determine which belief the instrument expresses. they need answers before trading starts, when nobody knows which wording will benefit them.

and only one branch happens. if Jensen stays, we never observe the stock price in the alternative world where he left. settlement rules can specify what happens to inactive claims; they cannot manufacture the missing counterfactual. this puts limits on how confidently we can evaluate the forecasts afterward. conditional markets organize uncertainty. they do not remove the inconvenient parts of it.

liquidity is the other constraint. a precise question with three disengaged traders is not much of an information product. splitting markets by event, deadline, and successor can fragment participation faster than it adds useful detail. market makers need a reason to quote, informed traders need a reason to participate, and everyone needs confidence in resolution. the mechanism has to work with actual capital, not just look elegant in a diagram.

despite all that, this seems worth building. analysts already construct these scenarios privately. investors already distinguish the probability of a shock from its consequences. the missing step is giving those judgments counterparties and public prices. the question is familiar; the tradeable instrument is what needs work.

Nvidia makes the point concrete. someone can be bullish on the company and worried about Jensen leaving. someone else can think he will leave and still be bullish. ordinary shares compress those positions into the same conversation. conditional markets would let us see where the disagreement actually sits. “Jensen is irreplaceable” would remain an opinion, but we could finally ask what price its believers are willing to pay.