Jensen Huang Built the Central Bank of AI. His Empty Life Palace Is Why

Put Musk and Huang side by side and the difference is not ability — it is which road each of them chose. Musk’s core move is "I will do it myself": his own rocket engines, his own reusability, even his own satellite network. He does not join a platform someone else built; he builds the platform, and you may come or not.

Huang took the other road. He does not train large models, does not run a cloud, does not build the data centres himself — other people do all of that. What he does is bring those people into his system: whoever builds the models, the clouds, the servers, every one of them ends up connecting through his layer, CUDA. He does not build the platform; he becomes the ground that everyone on it stands on.

In the language of martial-arts fiction: men like Musk and Soros are grandmasters who found their own school — the technique and the school are raised together, and the world either joins or goes around. Huang is closer to the well-connected diplomat: able to work with anyone, offending no one, until he is pushed up together onto the seat of the alliance leader.

But this leader leads differently: he does not rule the world of the craft by decree — he invites more and more people into his own system to ally with one another. Those who come in are not taking orders; they are plugging into each other — the model builders plug into his software layer, the server makers into his specification, the clouds into his chips. Every new member makes the network more useful to everyone already inside it, and lifts him with it. That is the difference from the grandmasters: the bigger a grandmaster’s school grows, the more the world splits into camps; the bigger his system grows, the more the world links into one piece, and everyone inside grows together.

Neither road is higher. What is worth seeing is something else: both of them chose the road that runs smoother for their own hand. Not the stronger road — the one that costs less for that hand to play. That is what "suited to your own gift" actually means: it is a match, not a ranking.

That hand of Huang’s has a name in Zi Wei Dou Shu: a Life Palace with no major star — conventionally called an "empty palace". Cast from publicly circulated birth data (Jensen Huang, 1963-02-17, not verified by the individual), the Baize engine produces precisely such a chart: no major star in the Life Palace.

"The central bank of AI" refers to CUDA — nearly the whole AI industry issues its compute through that layer. The defining feature of such a position: it does not stand in front; it is the layer everyone else runs on. The value of an empty palace grows in exactly these places.

An empty palace is often read as bad news: other people’s Life Palaces carry major stars, yours is vacant, as if a piece were missing. In youth, the verdict handed to this kind of chart most often is that plain one — can’t really tell what’s special about him; as if having no label equals having no value.

What it actually says is something else: this person’s centre of gravity is not on "who I am" but on "what I work with". An empty palace takes its definition from the environment — you read it by which stars are borrowed from the opposite palace. So early on, that vague verdict is simply the normal state: the specialty only shows up once the right situation arrives, and beforehand nobody can pin it down, including the person himself.

One line: a chart has no good or bad; what matters is how you play the hand. An empty palace is dealt a "partner-type" card — its ceiling depends on the game it enters.

Huang’s public route is the complete footnote to this structure.

He co-founded NVIDIA in 1993. The first product failed; the company came close to bankruptcy and made deep layoffs. In 1999 the GPU turned it around and it went public. The real turning point came in 2006: CUDA was released free of charge. For years afterwards it was financially one-way — analysts kept questioning it — until 2012, when deep learning exploded and the whole industry found that its models could only run on his cards. In 2022 the gaming business slumped, and the company moved its centre of gravity wholesale to data centres — the main engine of the AI era shifted gears in that year.

Put back against the chart, the correspondence is remarkably neat. His chart is a Direct Wealth structure with Eating God generating wealth — accumulation inside existing rules, skills turned into products, no gambling outside the rules. Weak day master with the resource element (Earth) as fuel — support systems are his keyword. The standing advice for a weak structure is "borrow leverage, work with others", and his publicly documented way of operating is exactly that: team- and ecosystem-driven, not lone-hero-driven.

The most interesting step is the resource element itself. In this system the resource is "what supports and qualifies you". CUDA is the support he added for the entire industry: free, open, and something everyone came to depend on. He first built the support system for the industry; the industry handed him the position back. An empty palace does not define itself — he traded the right of self-definition for being depended on.

The other side of the same structure is public too. In 2018 the crypto market cooled, GPU inventory piled up, and the company did not cut upstream orders in time — a write-down was recorded in the public filings. Insufficient vigilance about upstream ecosystem dependence is on the record.

In structural language: an empty palace takes its position from the environment — when the environment shifts gears, his adjustment is half a beat slower than others, because "working with" something first requires confirming what the new rules are. This is not a list of defects; it is the two faces of one structure. What let him put thirty years into one ecosystem, and what cost him an extra tuition bill when the ecosystem shifted, are the same thing.

So what this piece hands to the reader is not "how Jensen Huang succeeded", but something plainer: a chart has no good or bad. The cards are dealt and cannot be swapped — the only thing you can move is this: read your own hand clearly, then play it well. An empty palace has its way of playing; a Direct Wealth structure has another. Misread your own hand, and the harder you push, the more it costs.

This is the other face of what we wrote in Musk and Soros: on the chart, the same type — a path grows out of a chart; it cannot be borrowed. Musk’s road grew out of his chart; hand it to Huang and it does not run. Hand Huang’s road to Musk as it stands, and no identical NVIDIA grows out of it. This is not about who ranks higher, only this: another person’s answer answers another person’s question.

Huang’s hand and Musk’s hand are not the same, which is why their roads do not carry over to each other — and why a private, customised comparison is worth more than any number of celebrity cases. The most direct way to read your own hand is not to read more cases: it is to hold it up against somebody.

Step one: Benchmark Figures. That is what this feature inside the custom reports does. Pick a reference figure, and it maps out where you are the same type, where you are the opposite, what can be learned and what breaks when copied, then lands back on your own next step. It is a difference analysis, not a success formula (charged per run).

Step two: if the person you want to compare with is not on the list, add them to Relationship Readings. That is the only feature on Baize that reads two people: enter the other person’s birth data, and you see the structural differences between you — not a verdict on whether the relationship is good, but a look at which hand he holds and how the two hands interact (charged per run).

Step three, and only then, the questions. Once the comparison is done, whatever is left is concrete. Paste these straight in:

• Given my chart, where is my centre of gravity — and in what kind of situation does this hand show its value?
• The person I want to model myself on: where are we the same type, and what breaks when I copy it?
• Given my structure, is the game I am in now worth ten years of me?

All three ask the same thing: not who you resemble, but how this particular hand should be played. It answers from your chart, not from a generic FAQ.

First time on Baize and not sure which step comes first? The full order — chart, question, review, coming back — is in First Time on Baize: Where to Start.

Is an empty Life Palace bad luck?

No. It describes where the centre of gravity sits: not on self-definition, but on what you work with. It is a structural description, not a good-or-bad label — the same structure is platform-grade value in the right game and half a beat late in the wrong one.

Did you really cast his chart?

Cast from publicly circulated birth data (1963-02-17), not verified by the individual. Because the data is not his own, this piece does structural comparison only: no prediction of his future and no chart-level verdict on him.

Are people with empty Life Palaces spineless?

No. The shape of a backbone differs: some structures stand on self-definition; an empty palace stands on making the collaboration excellent. The first risks self-enclosure, the second drift with the environment — both need managing, and neither ranks above the other.

What is "the central bank of AI"?

CUDA is the software layer NVIDIA opened up in 2006; essentially all AI model training runs on top of it — like a clearing system every bank must connect to. That is a statement of public fact, not an investment judgement.

Can I just follow his road?

No — the cards differ. Huang is an empty Life Palace with resource as fuel; Musk is another structure (see "Musk and Soros: on the chart, the same type"). A path grows out of a chart; another person’s answer answers another person’s question. What you do is not copy, but read your own hand clearly.

What should I take from his example?

Learn the structure, not the outcome. A chart has no good or bad; the cards cannot be swapped; the only movable part is how you play. That gets answered with your own chart — which is why a private, customised comparison beats any number of celebrity cases.