#黄仁勋辞职 #每周来晒 Old Huang hands over the reins: Will Nvidia still dare to bet on tomorrow?
People often say the world keeps turning no matter who leaves. It does keep turning, but how it turns matters. In politics, this saying usually holds true. Presidents and prime ministers come and go so often you can barely remember their names, and none seems much better than the others. But industry is different: the heads of outstanding companies need the nerve to stake their fortunes on the future. That's the "entrepreneurial spirit."
Jensen Huang is a tough character who has that spirit in his bones. That's why there's a saying online: once Huang hands over the reins, Nvidia will sooner or later become mediocre. It's not that he's without precedent or peer; it's just incredibly hard to find another person like him. At least within Nvidia, there are no signs of one yet.
Some people push back by citing Apple: Jobs has been gone for years, and Apple is still doing just fine, with its stock hitting record highs time and again.
By the time Jobs left, Apple was already a finely tuned money-printing machine that had been running for more than a decade. The iPhone had fixed the definition of a smartphone, the iOS ecosystem had become a closed loop, the App Store collected its 30% toll year after year, and Cook had whipped the supply chain into shape.
What he left Cook was a product kingdom that could run on autopilot—there was no need to bet on tomorrow every day; as long as the wheel wasn't turned the wrong way, momentum could keep it going for years. The Cook era was essentially about "protecting the legacy and selling phones well." Add a Dynamic Island or switch to a titanium frame now and then, and consumers would still line up to pay. When Cook hands over to Ternus, what he passes on is still a closed-loop ecosystem: consumers care about the brand and their habits, not who the CEO is.
Nvidia is nothing like that. What Huang leaves behind isn't a self-driving car, but a game of Texas Hold'em where every move requires a decision and every hand raises the stakes. GPU iterations have been compressed to around two years, rivals are closing in at every turn, TSMC capacity must be secured, HBM locked down, export controls navigated, and custom requirements from cloud providers met. B2B customers have no sentimentality: if performance slips a notch today, tomorrow's orders are gone. In this situation, the new CEO's job isn't to "hold the line" but to "bet"—on architecture, capacity commitments, and geopolitical maneuvering.
Huang spent 33 years betting on CUDA, betting on the AI computing-power wave, and betting on securing TSMC CoWoS capacity. If his successor is a career executive skilled at balancing the books, they'll most likely turn "daring to bet" into "playing it safe": the short-term financial reports may look fine, but in the long run Nvidia will retreat from rule-maker to rule-follower.
The succession challenges at Apple and Nvidia aren't even in the same league. One is about holding on to an empire; the other is about opening up new worlds—and people who open up new worlds can't be picked from HR's candidate pool.
Let's not be too categorical.
The power of momentum shouldn't be underestimated either.
The real variable is strategic boldness at the top. At many past life-or-death moments, Huang made the call himself: daring to pour huge sums into a new architecture, daring to lock in priority TSMC capacity in advance, and daring to work the political and industry circles to tightly bind GPUs to the AI wave.
One lucky hit is luck; repeatedly catching the wave is skill. That kind of decisiveness is hard to replicate inside the company. A conservative successor can keep the business going; the hard part is developing the guts to make big bets and redefine the field. Preserving what exists is easy; pioneering is hard.
How can we tell whether Nvidia is sliding toward mediocrity? Three visible signposts.
1. The boldness of capital expenditure.
Huang's boldest move was to sign long-term capacity commitments before the market caught on—CoWoS, HBM, advanced packaging, all bets placed two or three years ahead. If his successor starts "flexibly reallocating capital expenditure," postponing non-urgent capacity, and replacing major capacity commitments with on-demand purchases, translated, that means: they no longer dare to bet. A CEO focused on preserving the status quo will always first think about optimizing cash flow, not capturing the future.
2. Are there still "projects no one else understands" in the architecture roadmap?
Huang spent more than a decade nurturing CUDA despite shareholders' criticism. If the R&D roadmap becomes increasingly "explainable" and increasingly tailored to current customer tastes, that should actually raise concerns—no one is betting on obscure fields that nobody understands anymore. When a genuine frontier breakthrough is approved, someone will inevitably say, "This maniac is burning money again."
3. How it weathers the next darkest hour.
When the crypto mining boom faded, revenue plunged, and the data center business had yet to take off. Each time, Huang proactively pivoted instead of waiting for the market to provide an answer. When the next downturn comes, will the successor raise the stakes against the tide or passively retrench? Will they nurture long-term projects or cut "non-core" areas to protect profits? That reaction will determine whether Nvidia is driving manually or sliding downhill.
These three signposts boil down to one question: is it still "betting on tomorrow," or has it switched to "calculating what's right in front of it"?
As long as it is willing to pour money into Capex, nurture obscure fields, and bet against the tide during downturns, Nvidia is not yet mediocre. Conversely, the three signposts don't carry equal weight: Capex is the baseline, obscure fields are the backbone, and betting against the tide is the ultimate test.
If the first two are gone, it means management no longer has any offensive line in motion, and the computing-power machine is beginning to degenerate from "defining the field" to "following the field."
So the conclusion isn't "it will collapse as soon as he leaves," but that there are two possible paths: on the momentum of CUDA, Nvidia's core business can remain stable for many years; but continuing to stand at the forefront of the technology wave and consistently surprise the market will be a lot harder.
In short, a gradual slide into mediocrity is a possibility that cannot be ignored.
What Jobs left Apple was a self-sustaining product kingdom. What Huang leaves Nvidia is a high-stakes game still running at full speed, with decisions required at every turn. No matter how deep the moat, it cannot withstand a long-term absence at the top of a leader willing to bet. Nvidia's most valuable asset isn't its chips; it's the person willing to stake everything on the unknown. That kind of person doesn't show up in financial reports or KPIs, but without him, Nvidia's stock-price ceiling will be a lot lower.
Change the dealer and the game goes on. But with the top poker player gone, who will still dare to go all in?$NVDA
People often say the world keeps turning no matter who leaves. It does keep turning, but how it turns matters. In politics, this saying usually holds true. Presidents and prime ministers come and go so often you can barely remember their names, and none seems much better than the others. But industry is different: the heads of outstanding companies need the nerve to stake their fortunes on the future. That's the "entrepreneurial spirit."
Jensen Huang is a tough character who has that spirit in his bones. That's why there's a saying online: once Huang hands over the reins, Nvidia will sooner or later become mediocre. It's not that he's without precedent or peer; it's just incredibly hard to find another person like him. At least within Nvidia, there are no signs of one yet.
Some people push back by citing Apple: Jobs has been gone for years, and Apple is still doing just fine, with its stock hitting record highs time and again.
By the time Jobs left, Apple was already a finely tuned money-printing machine that had been running for more than a decade. The iPhone had fixed the definition of a smartphone, the iOS ecosystem had become a closed loop, the App Store collected its 30% toll year after year, and Cook had whipped the supply chain into shape.
What he left Cook was a product kingdom that could run on autopilot—there was no need to bet on tomorrow every day; as long as the wheel wasn't turned the wrong way, momentum could keep it going for years. The Cook era was essentially about "protecting the legacy and selling phones well." Add a Dynamic Island or switch to a titanium frame now and then, and consumers would still line up to pay. When Cook hands over to Ternus, what he passes on is still a closed-loop ecosystem: consumers care about the brand and their habits, not who the CEO is.
Nvidia is nothing like that. What Huang leaves behind isn't a self-driving car, but a game of Texas Hold'em where every move requires a decision and every hand raises the stakes. GPU iterations have been compressed to around two years, rivals are closing in at every turn, TSMC capacity must be secured, HBM locked down, export controls navigated, and custom requirements from cloud providers met. B2B customers have no sentimentality: if performance slips a notch today, tomorrow's orders are gone. In this situation, the new CEO's job isn't to "hold the line" but to "bet"—on architecture, capacity commitments, and geopolitical maneuvering.
Huang spent 33 years betting on CUDA, betting on the AI computing-power wave, and betting on securing TSMC CoWoS capacity. If his successor is a career executive skilled at balancing the books, they'll most likely turn "daring to bet" into "playing it safe": the short-term financial reports may look fine, but in the long run Nvidia will retreat from rule-maker to rule-follower.
The succession challenges at Apple and Nvidia aren't even in the same league. One is about holding on to an empire; the other is about opening up new worlds—and people who open up new worlds can't be picked from HR's candidate pool.
Let's not be too categorical.
The power of momentum shouldn't be underestimated either.
The real variable is strategic boldness at the top. At many past life-or-death moments, Huang made the call himself: daring to pour huge sums into a new architecture, daring to lock in priority TSMC capacity in advance, and daring to work the political and industry circles to tightly bind GPUs to the AI wave.
One lucky hit is luck; repeatedly catching the wave is skill. That kind of decisiveness is hard to replicate inside the company. A conservative successor can keep the business going; the hard part is developing the guts to make big bets and redefine the field. Preserving what exists is easy; pioneering is hard.
How can we tell whether Nvidia is sliding toward mediocrity? Three visible signposts.
1. The boldness of capital expenditure.
Huang's boldest move was to sign long-term capacity commitments before the market caught on—CoWoS, HBM, advanced packaging, all bets placed two or three years ahead. If his successor starts "flexibly reallocating capital expenditure," postponing non-urgent capacity, and replacing major capacity commitments with on-demand purchases, translated, that means: they no longer dare to bet. A CEO focused on preserving the status quo will always first think about optimizing cash flow, not capturing the future.
2. Are there still "projects no one else understands" in the architecture roadmap?
Huang spent more than a decade nurturing CUDA despite shareholders' criticism. If the R&D roadmap becomes increasingly "explainable" and increasingly tailored to current customer tastes, that should actually raise concerns—no one is betting on obscure fields that nobody understands anymore. When a genuine frontier breakthrough is approved, someone will inevitably say, "This maniac is burning money again."
3. How it weathers the next darkest hour.
When the crypto mining boom faded, revenue plunged, and the data center business had yet to take off. Each time, Huang proactively pivoted instead of waiting for the market to provide an answer. When the next downturn comes, will the successor raise the stakes against the tide or passively retrench? Will they nurture long-term projects or cut "non-core" areas to protect profits? That reaction will determine whether Nvidia is driving manually or sliding downhill.
These three signposts boil down to one question: is it still "betting on tomorrow," or has it switched to "calculating what's right in front of it"?
As long as it is willing to pour money into Capex, nurture obscure fields, and bet against the tide during downturns, Nvidia is not yet mediocre. Conversely, the three signposts don't carry equal weight: Capex is the baseline, obscure fields are the backbone, and betting against the tide is the ultimate test.
If the first two are gone, it means management no longer has any offensive line in motion, and the computing-power machine is beginning to degenerate from "defining the field" to "following the field."
So the conclusion isn't "it will collapse as soon as he leaves," but that there are two possible paths: on the momentum of CUDA, Nvidia's core business can remain stable for many years; but continuing to stand at the forefront of the technology wave and consistently surprise the market will be a lot harder.
In short, a gradual slide into mediocrity is a possibility that cannot be ignored.
What Jobs left Apple was a self-sustaining product kingdom. What Huang leaves Nvidia is a high-stakes game still running at full speed, with decisions required at every turn. No matter how deep the moat, it cannot withstand a long-term absence at the top of a leader willing to bet. Nvidia's most valuable asset isn't its chips; it's the person willing to stake everything on the unknown. That kind of person doesn't show up in financial reports or KPIs, but without him, Nvidia's stock-price ceiling will be a lot lower.
Change the dealer and the game goes on. But with the top poker player gone, who will still dare to go all in?$NVDA
