9 min read5 viewsinvesting

Gooaye EP714: Whether a Role Can Be Handed Over Depends on How Many Things Are Tied Into It

A server factory corridor late at night, an empty chair in the foreground with a spec drawing still spread open on the desk, racks receding in rows into the depth of the hall, a single lamp still lit at the far end

Notes on the 14 November 2026 episode of Gooaye (EP714), where host Hsieh Meng-kung works through the resignation of Nvidia founder Jensen Huang. The market read it as bad news clearing out and bounced the stock, which he takes as proof that pricing runs ahead of headlines. His spoken figures that day: Nvidia's multiple already under 10x, Huang's stake a little over 3%, which he would have to sell into if he starts something new. He expects the long-term story to be rewritten without this one person, and says he has trimmed his own chip-related exposure. Every number here is his estimate or market chatter on the day, with no filings behind it. Educational listening notes only — no investment advice, no price targets, and the judgements here may be wrong.

  • Gooaye
  • podcast-notes
  • Nvidia
  • supply chain
  • succession
Contents
  1. Why did the stock rise first when the news broke?
  2. Without him, does this company really change?
  3. Will Taiwan’s supply chain be written down along with him?
  4. If a supply-chain management master takes over, who loses most?
  5. So how should I relook at the holdings that “charge with the specs”?
  6. Sources worth checking
  7. One Thing to Take With You

A server factory corridor late at night, an empty chair in the foreground with a spec drawing still spread open on the desk, racks receding in rows into the depth of the hall, a single lamp still lit at the far end

夫運籌策帷帳之中,吾不如子房;鎮國家,撫百姓,給餽饟,不絕糧道,吾不如蕭何;連百萬之軍,戰必勝,攻必取,吾不如韓信。此三者,皆人傑也,吾能用之,此吾所以取天下也。

In planning strategy inside the tent, I am no match for Zifang; in steadying the state, caring for the people and keeping the grain road open, I am no match for Xiao He; in leading an army of a million where victory is certain, I am no match for Han Xin. These three are men of rare gifts, and I was able to use them — that is how I took the empire.

—— Sima Qian, Records of the Grand Historian, “Annals of Emperor Gaozu” (Han dynasty; translation mine)

In EP714 of Gooaye, published 14 November 2026, host Hsieh Meng-kung spends half the episode on the resignation of Nvidia founder Jensen Huang. The market read the news as bad news clearing out and bounced the stock, which he takes as evidence that pricing runs ahead of headlines, so drawing a straight line from a story to a short-term move gets you hurt. His spoken numbers that day were a multiple already under 10x and a founder stake a little over 3%, and his judgement was that the long-term story gets rewritten without this one person — which is why he has trimmed his own chip-related positions. All of it is his estimate on the day, with no filings behind it.

Why did the stock rise first when the news broke?

Short answer: part of this had already been priced before the headline landed.

A comparison of two price paths: the upper one slides lower all the way before the vertical line marking the news release, then ticks up only slightly after it; the lower one only starts falling at the release point. The difference in shape is whether the decline happens before the news or after it.

He says the commentary filling his social feeds was all guessing in one direction, and then the market handed everyone a surprise in the other. His conclusion is that pricing happens early, and that the “this happened today, therefore the market will do that” calls have been wrong with remarkable consistency, from the first hours of the story right through to now.

That stretch caught a habit of my own. When a big story breaks, my first move is to work out which way it pushes the price, when the first move should be to ask whether someone has already run that calculation. He leaves one long-term use for the simple read, though: go back to the plainest layer of judgement and it gives you direction for the road ahead, while the short-term reaction is not something to turn into a conclusion.

Without him, does this company really change?

Short answer: he thinks it changes a lot, because the role is several things tied into one.

He describes Huang’s function as holding a whip behind the supply chain, with the pace of spec increases forcing everyone to keep up: cooling moved from air to liquid quickly, and then on to full liquid cooling in the next generation. Vendors who kept up earned excess returns. Competitors followed the specs he opened — AMD chasing behind, Google watching what he pushed.

A four-step staircase rising to the right, labelled air cooling, liquid cooling, and full liquid cooling. The dot for the one who steps first stands on the top step, the two follower dots sit one and two steps behind, and an arrow points down from above back to the step the first mover has already taken.

Some people, he says, have the causation backwards, assuming others would have driven to this point anyway. His version is that he went first, and the rest followed once the road looked passable. Run it again with nobody out front and those companies might not have put the same money into opening those specs, and the pace would have slowed.

Then there are three things you cannot buy separately. One, command of large capital and real say over the supply chain, enough to secure the largest share of capacity and to get partners to the table. Two, the founder-mode willingness to open a path, always picking the technical route that looks riskiest, with a lot of people eating well once it is pushed through. Three, a flat organisation where dozens of managers report to him directly — and in past interviews he had said he was not running a succession plan. His Apple comparison: Tim Cook was raised inside Apple for years before taking over, and Nvidia has nobody who has sat that long. There is a woman running supply chain in Taiwan, he notes, but that is a specialised function — it covers components, not necessarily the dinner with assorted partners where something gets agreed.

On the left, three thin strands twist to the right into one thick rope, labelled capital and capacity clout, the will to open the road, and a flat organisation; on the right is the successor's hand, holding only one strand while the other two slip loose to the side.

He also flags a branch that touches the price directly: if Huang starts something new he needs cash, so some of that stake a little over 3% has to be sold, which presses on the stock near term. And people around him want to follow — he says acquaintances inside Nvidia have mentioned they would go if the boss would take them. His best case is that the man takes whatever sanction and penalty is due and comes back to lead the company. His worst case is that Huang starts something new and the company brings in a supply-chain management master instead.

Will Taiwan’s supply chain be written down along with him?

Short answer: he thinks crediting Taiwan’s visibility to one person overshoots.

The bigger hand is policy. The US wants supply chains out of China, so orders that would have landed there came out in bulk, and the fat orders Taiwan, Korea and Japan picked up came from that direction. China, he says, is capable of doing this board assembly work well, even better, grinding at it with fewer environmental constraints — memory is where you will see that next. His landing point is that isolation from China deepens after this round, so Taiwan’s position does not loosen because of one personnel story.

Two arrows push the same tray, one very wide labelled policy push and one very thin labelled one man's push; the tray reads orders Taiwan wins, and once the thin arrow is pulled back the tray barely moves.

The group to watch is a different one: the handful of names that charged in the front row every time he opened a new spec, which includes the most expensive stocks in the market. Those high multiples were buying a claim — whatever spec he opens, this plant can match it. That claim is what got shaken this week. Short term is fine, he says, because others will carry the roadmap forward now that an example exists. What gets rewritten is the multiple the market is willing to pay.

If a supply-chain management master takes over, who loses most?

Short answer: the ones out in front.

He uses the Apple experience: the early money in Apple’s chain was decent, and later the pricing was squeezed until some plants would rather not have the orders. They have looked at Apple’s servers in the chain — built with the finish of a consumer laptop, cabling laid out neatly, laser etching on the shell — but with spec increases that never charged the way Nvidia’s did, sometimes testing a laptop chip in there. Orderly and standardised, which is not good news for the supply circle.

He puts it bluntly: the market needs someone who sells a dream and charges at it. That kind of person gets beaten up on social media, and people call it hype at first, and it hardly matters if the roadmap arrives slower than promised, because what the market wants is a story and some room. His analogy is romance — a relationship with no visible future takes the motivation out of even talking.

What I came away with is that servers end up where PCs and phones ended up, standardised and fought over on price. The spec race is still running because someone out front is pushing it. The type of successor decides how much of that race is left.

A road splits at the point marked succession: the upper branch keeps climbing the spec staircase, while the lower branch turns onto a long horizontal runway labelled standardisation and price cutting.

So how should I relook at the holdings that “charge with the specs”?

This is the question that stayed with me. The confusion a lot of people have this week is simpler: the news is huge, so do I cut this position or hold it?

I split a stock’s multiple into two pieces. The first is the orders and capacity it has now, visible in the filings. The second is “it will win the next spec cycle too,” which rests on one person’s momentum. This episode moved the second piece, and the second piece takes a few quarters to show up in filings. So the work is to record them separately: how much of this company’s profit comes from orders already in hand, and how much from everyone believing it will catch the next round. The higher the second share, the heavier this week’s news lands on it.

Two bars of equal height are each split into two blocks, the lower one orders in hand and the upper one hopes for the next spec round. The left bar's upper block is thin and the right bar's is thick, and only the upper blocks are drawn as dashed outlines and shifted by an arrow pushing them.

Hsieh’s own move was trimming chip-related exposure — volatility and risk up, too many variables, wanting to wait for things to settle, with the money going to places that do not take this kind of hit in the front row. I file that as his position and his tempo, not as advice; he says he trades in this market and his worst fear is no volatility and no story, and my holding period is not his.

The test I have been using on myself: for this holding, if I delete the sentence “it will win the next spec cycle too,” am I still willing to own what remains? Where I can answer, this week’s news is noise to me. Where I cannot, my homework is the order structure and customer concentration, not guessing whether that one man comes back. Two of mine are still unanswered, so I did not add this week.

Sources worth checking

  • Gooaye EP714, 14 November 2026, the source for this piece
  • Nvidia’s filings and earnings calls on product and regional mix, useful against the spoken estimates on multiples and revenue structure
  • Nvidia’s proxy statement and insider ownership filings, to check the stake of a little over 3% mentioned in the episode
  • Public reporting and past interviews on Nvidia’s org structure and succession arrangements, the source of the flat-organisation and no-succession-plan points
  • US export control and supply-chain policy announcements, the background for that stretch of the episode
  • Reporting on margins and order terms along Apple’s supply chain over the years, against the claim that pricing gets squeezed once things standardise

One Thing to Take With You

Whatever a standard answer can replace was never your position.

There is a stretch early in the episode with nothing to do with stocks. He was going to meet a possible funding partner, a games developer, and he asked around first about the man’s tastes: Taiwanese food, a private room, which categories to avoid. Then he put together three wines — one floral, a 1998 Italian bottle with age for anyone who likes it heavy, and a 2016 blend to close. Comparing vintages and price tags is something anyone can do, he says, because money solves it; what he picked for was plots and stories. The man turned out to be easygoing, holding nothing up to the light, swirling nothing — the preparation was surplus. He laughed that it was what the drill sergeant taught: prepare extra, and if the person across the table is a monster you can still hold the line.

The part that follows is my favourite of the episode. AI pulls everyone up to an above-standard answer: ask it about wine, ask it to translate, the answer scores fine. But once you have read enough of those answers, a few minutes of conversation tells you who looked it up. What is left to tell people apart is where you hold a line, whether you deliver what you promised, and whether you have lived anything you can tell as a story. New tools help little with those.

A horizontal standard line carries a dense row of short bars all of the same height, with only a few clearly rising above the line; those few are labelled persistence, delivery, and experience.

Which is the same point as the first half. Huang’s role is bound out of several things nobody can buy separately, so it is hard to fill; a person’s taste is bound out of years of money wasted and roads taken wrong, so it is hard to copy.

One thing I have tried, and it works outside investing: pick something people ask you about that makes you reach for a search — coffee, car repair, getting a kid to sleep, the patch history of some game. For the next month, every time you look something up, add one line of your own version in the same notes file: what I tried, which attempt failed, which way I now prefer. Read the file at the end of the month, and the “I tried” lines are the part nobody can look up.

This article is an educational discussion of investment method. It is not advice to buy or sell any individual security, offers no target prices, and does not analyze any current holding. Investing carries risk; make your own decisions or consult a qualified professional.

Comments

Loading comments…

Sign in with Google before posting. Only your name and profile picture are shown.