$125.53
24h
-14.01%
21 Nov, 02:05
Leadership transition risk threatens NVIDIA's supply chain moat
Bearish0.00%
AI Summary
The author argues that NVIDIA's competitive moat relies heavily on Jensen Huang's personal supply chain relationships and capital investments, warning that the board's recent leadership transition and tightened investment policies have already triggered a 13% stock decline and could cause further ecosystem fragmentation if Jensen does not return.
NVIDIA has never been an "ordinary chip design company." Its barrier to entry lies not in blueprints, but in two networks woven exclusively by Jensen Huang: one is a thirty-three-year-deep industrial chain rooted in Taiwan, and the other is an ecosystem design that binds customers and suppliers into a community of shared destiny. Once these two networks lose their weaver, what remains is merely a pile of advanced GPU designs that cannot be brought to market.
The First Network: The "proactive coordination" of the Taiwanese industrial chain relies on Old Huang's personal favors and credit, not contracts.
TSMC handles over 70% of NVIDIA's high-end GPU capacity, exclusively supporting advanced processes like 4nm and 3nm, while CoWoS advanced packaging also hinges on TSMC. Foxconn, Quanta Computer, and Wistron are responsible for server assembly, while SK Hynix and Samsung supply HBM. The Asian supply chain accounts for 90% of NVIDIA's production costs. This "proactive coordination" did not come out of thin air. Jensen Huang has known TSMC founder Morris Chang for over twenty years; in 2013, Chang invited him twice to take over TSMC. Every year, Huang hosts a "Trillion-Dollar Banquet" in Taiwan, where leaders from TSMC, Foxconn, Quanta, and Wistron sit at the same table to discuss capacity, technology roadmaps, and investment schedules for the coming years. When suppliers decide to build a wafer fab or expand a packaging line, involving tens of billions of dollars, they often make these decisions before orders are fully confirmed. They dare to bet early because they trust the order credit accumulated by Jensen Huang over thirty-three years, rather than a new CEO's claim that "our supply chain planning has always been excellent." Some Taiwanese supplier executives have publicly stated after Huang's departure that they are willing to meet NVIDIA's capacity needs, but they are "worried that his successor may not be as proactive about reserving capacity as Jensen Huang."
The Second Network: Using capital to "front the money" for the entire industrial chain, locking the ecosystem onto NVIDIA's computing foundation.
This is Jensen Huang's true capability, exceeding that of a mere "chip CEO." Under his leadership, NVIDIA has invested more than $100 billion in equity and financial support for AI startups, suppliers, and chip clients, playing the role of an "unofficial banker" in the AI boom. This is not simple financial investment. Investments in OpenAI and Anthropic essentially lock in their next-generation models' priority adoption of NVIDIA's Rubin architecture, while using feedback from the model side to fuel hardware iterations. For "new clouds" like CoreWeave, NVIDIA provides purchase orders and credit endorsements, helping them borrow money to buy cards and build data centers, which in turn buy NVIDIA chips. It even provided lease and power expense guarantees of up to $105 billion for OpenAI's data center campus, acting as its own client's "landlord." This cycle of "investment—procurement—re-investment," criticized by outsiders as "circular financing," has practically paved the way for the entire industrial chain—from model companies to cloud service providers to server assembly plants—for NVIDIA's next-generation chips. White-collar professional managers calculate quarterly gross margins and return on investment; Jensen Huang calculates whether the entire industrial chain will still be on NVIDIA's track ten years from now.
What the board has done since taking over proves they do not understand how much these two networks are worth.
After Jensen Huang's departure, interim CEO Crese's first move was to advance annual salary increases and equity awards from March to December, attempting to stabilize employees with a "change in timing" tactic. Meanwhile, her external investment stance immediately shifted from Huang's "I invest in you, I give you money" to a "disciplined" and "rigorous" screening process, sharply raising financing thresholds for new cloud clients. Employees are not fools. Since Huang's resignation, NVIDIA's stock price has fallen 13%, wiping out approximately $500 billion in market value. With employee compensation deeply tied to stocks, asset shrinkage is a painful reality. A senior employee wrote to the board warning that without considering reinstating Jensen Huang, NVIDIA would lose "irreplaceable talent." Board Chairman Jones admitted that "finding a clone of Jensen Huang is unrealistic," yet continued to scout for external successors through headhunters. They preferred to place a professional manager with no supply chain credit or ecosystem design experience in that position rather than give Old Huang an advisory role to facilitate a power transition. This is not governance; it is dismantling the engine to replace it with a more "compliant" dashboard.
Over his thirty-three years, Jensen Huang repeatedly said the same thing: "The company is only thirty days away from bankruptcy." To the board, this phrase might just sound like emotional rhetoric from a founder. But those who have truly executed chip落地 know it is not rhetoric. Without TSMC proactively expanding CoWoS production lines, without Foxconn and Quanta preparing assembly capacity in advance, and without using equity investments to front the computing demand for clients, even the most advanced chip designs from NVIDIA would merely be an order on TSMC's schedule that could be delayed. Huawei's chip design is not inferior; what it lacks is the depth of execution to turn designs into products—and NVIDIA's depth of execution was woven inch by inch by Jensen Huang using thirty-three years of personal favors, credit, and capital operations.
Any future public statements by the new CEO will see comments repeating the same question: Where is Jensen? The supply chain will not ask the board "what is your strategy"; they will only ask "is Old Huang still in charge." If the answer is not affirmative, proactive coordination will become contract-based compliance, exclusive production lines will become queues waiting for slots, and ecosystem binding will become each party seeking its own path. What NVIDIA needs is not "a NVIDIA without Jensen Huang"; it needs Jensen Huang back, even if only to return to the table where the network is woven.