"Our company is always thirty days away from going out of business."

— Jensen Huang, repeated as a weekly liturgy inside Nvidia for over three decades.

On November 12, 2026, that thirty-day paranoia met its most unthinkable reality.

After 33 years, 12,000 consecutive days of founder leadership, and an engineering crusade that transformed a San Jose diner booth into the world’s first $4 trillion compute sovereign, Jensen Huang resigned as President, Chief Executive Officer, and Board Director of NVIDIA, effective immediately.

His departure was not an orderly retirement, nor was it a planned succession. It was the explosive climax of a seven-day geopolitical and corporate governance crisis intertwining international espionage, clandestine semiconductor supply chains, FBI counterintelligence, and an internal board investigation that unraveled the most powerful CEO in modern technology.

Within hours of the announcement, Wall Street entered a state of sheer multiple compression:

text script
┌────────────────────────────────────────────────────────────────────────┐
│                   THE WALL STREET VALUATION CLIFF                      │
├──────────────────────────┬──────────────────────┬──────────────────────┤
│ Institutional Desk       │ Previous Stance      │ Post-Resignation     │
├──────────────────────────┼──────────────────────┼──────────────────────┤
│ Bernstein (Stacy Rasgon) │ $400 PT (Outperform) │ $200 PT (Cut by 50%) │
│ Dan Ives (Yorkville)     │ $300 PT with Jensen  │ $225 PT (-$1T Cap)   │
│ Morgan Stanley           │ Top Pick Selection   │ Removed from Top Pick│
│ Global Brokerages        │ Unanimous Strong Buy │ 7 Downgrades in 24h  │
└──────────────────────────┴──────────────────────┴──────────────────────┘

Bernstein’s veteran chip analyst Stacy Rasgon—who famously coined the Wall Street gospel "Jensen is NVIDIA, NVIDIA is Jensen"—slashed his price target in half from $400 down to $200. Dan Ives established a split valuation: $300 a share with Jensen Huang; $225 without him, stripping nearly $1 trillion in prospective market capitalization.

Yet the financial carnage is only the perimeter. To understand why Jensen Huang fell, one must trace the shadow war for global silicon supremacy.


I. The Shadow War: Alice Huang, the MSS, and the FBI Probe

The crisis did not originate on Wall Street; it originated in the counterintelligence corridors of Washington and Southeast Asia.

text script
┌────────────────────────────────────────┐
                  │    U.S. ELECTRONIC SURVEILLANCE & CIA  │
                  └───────────────────┬────────────────────┘
                                      │ Classified Memo
                                      ▼
                  ┌────────────────────────────────────────┐
                  │     MARCO RUBIO / STATE DEPT BRIEF     │
                  └───────────────────┬────────────────────┘
                                      │ Target: Alice Huang
                                      ▼
┌──────────────────────────┐      Alleged Liaison     ┌──────────────────────────┐
│   JENSEN HUANG (NVIDIA)  │◄────────────────────────►│  ALICE HUANG (MEGASPEED) │
└─────────────┬────────────┘                          └─────────────┬────────────┘
              │                                                     │
              │ Subpoenas & Shareholder Suits                       │ Suspected Channel for
              ▼                                                     ▼ High-End H100s / H200s
┌──────────────────────────┐                          ┌──────────────────────────┐
│     FBI INVESTIGATION    │                          │ CHINA MSS & GREY CLOUDS  │
└──────────────────────────┘                          └──────────────────────────┘

Between June 2024 and May 2025, Huang was repeatedly photographed alongside Alice Huang (no relation)—a Chinese national and executive embedded within Singapore-based data center firms and Nvidia partner networks, most notably Megaspeed and Shanghai Century Litong.

In early 2026, U.S. Secretary of State Marco Rubio oversaw the drafting of a "highly classified" intelligence memorandum. Compiled from intercepted communications and allied surveillance, the memo alleged that China’s Ministry of State Security (MSS) had recruited Alice Huang to target Jensen Huang directly—seeking to extract confidential silicon roadmaps, gauge U.S. enforcement resolve, and subtly influence chip export policies.

The stakes were existential:

1. The H100 Sanctions Evasion Conduit: Federal investigators were probing whether relationships inside Nvidia’s partner networks facilitated the illicit rerouting of billions of dollars worth of restricted Hopper (H100/H200) clusters into mainland Chinese AI labs via shell data centers in Southeast Asia.

2. The Executive Code Breach: Concurrently, Nvidia’s board uncovered an internal relationship with a direct-report employee at Santa Clara headquarters, compounding the national security scrutiny with an unambiguous violation of corporate conflict-of-interest bylaws.

The dominoes fell with terrifying speed:

  • Sunday, November 8: Jensen Huang issued a public confession and apology admitting unfaithfulness in an extramarital relationship. Nvidia’s board convened emergency sessions to initiate an independent corporate investigation.
  • Monday, November 10: Donald Trump confirmed publicly that the FBI had launched a formal investigation into Huang’s ties with Alice Huang. Federal shareholder class-action lawsuits were filed alleging concealment of catastrophic national security risks.
  • Tuesday, November 11: Leaders of the House Foreign Affairs and China Select Committees demanded Huang appear for closed-door testimony under threat of congressional subpoena.
  • Wednesday, November 12: With institutional rebellion mounting and legal firewalls collapsing, Jensen Huang submitted his immediate resignation as CEO, President, and Board Director.

II. The 33-Year Arc: From Booth #4 to Planetary Compute

To understand the vacuum this creates, one must appreciate that Nvidia was never an ordinary corporation. It was a 33-year high-stakes gamble engineered around a single founder’s idiosyncratic willpower.

In 1993, Jensen Huang, Chris Malachowsky, and Curtis Priem met over pancakes at a Denny's in San Jose, sketching out a radical proposition: specialized acceleration hardware for 3D graphics. When their initial NV1 processor nearly killed the company in 1996, Huang fired 70% of the staff, staked everything on the RIVA 128, and barely made payroll.

Then came 2006. Long before "Artificial Intelligence" existed as an enterprise budget item, Huang mandated that every single silicon unit leaving Nvidia fabs must incorporate CUDA runtime engines.

For nearly a decade, Wall Street punished the stock. Investors regarded CUDA as a grotesque tax on consumer gaming margins—a vanity project subsidizing academic researchers with expensive silicon die space.

Huang refused to yield.

When University of Toronto researchers Alex Krizhevsky and Ilya Sutskever trained AlexNet in 2012 using two off-the-shelf Nvidia GeForce GTX 580 cards, computing crossed the Rubicon. Modern AI was born not in a DARPA supercomputing facility, but on the parallel architecture Huang had stubbornly forced into the global supply chain.


III. The Anatomy of the Moat: Hardware is Cheap, CUDA is Gravity

Nvidia was never just a chip fabricator. That persistent misconception is why AMD, Intel, and countless AI ASIC startups have spent years shipping mathematically faster raw FLOPS on paper while capturing almost zero data center profit.

Nvidia’s moat was a vertically integrated gravitational field:

text script
┌────────────────────────────────────────────────────────┐
      │           APPLICATION LAYER: LLMs, AGENTS, ROBOTICS    │
      ├────────────────────────────────────────────────────────┤
      │   PYTORCH / JAX / TENSORFLOW RUNTIME COMPILATION       │
      ├────────────────────────────────────────────────────────┤
      │   CUDA LIBRARIES: cuDNN, cuBLAS, TensorRT, Megatron    │
      ├────────────────────────────────────────────────────────┤
      │   COMMUNICATION PRIMITIVES: NCCL, NVLink 5, NVSwitch   │
      ├────────────────────────────────────────────────────────┤
      │   FABRIC & INTERCONNECT: Mellanox Quantum InfiniBand   │
      ├────────────────────────────────────────────────────────┤
      │   SILICON PACKAGING: TSMC CoWoS, Blackwell GB200       │
      └────────────────────────────────────────────────────────┘

1. The Software Lock-In (CUDA & NCCL): Every AI researcher for fifteen years learned to write matrix kernels in CUDA. Millions of lines of custom Triton, PyTorch bindings, and distributed tensor-parallel communications (NCCL) are hard-coded to Nvidia registers.

2. The Systems Architecture (Mellanox & NVLink): When Huang acquired Mellanox for $6.9 billion in 2019 outbidding Intel, analysts balked. Today, training trillion-parameter frontier models is bounded by network latency, not raw FLOPs. A cluster of 72 Blackwell chips behaves not as discrete cards, but as a single colossal, liquid-cooled super-GPU sharing unified memory at 1.8 TB/s.

Huang oversaw every layer of this stack personally—from the thermal tolerance of copper heat pipes to low-level firmware scheduling.


IV. The Succession Machine: Colette Kress & The Institutional Era

Nvidia’s Board executed an immediate corporate firewall:

  • Colette Kress (Executive VP & CFO since 2013) assumes the role of Interim CEO. Kress has spent 13 years managing Nvidia's capital through crypto mining booms, silicon gluts, and the generative AI hyperscale explosion. She is universally respected by institutional investors as an unflinching, data-driven operator.
  • Harvey C. Jones (Nvidia board member since 1993 and former CEO of Synopsys) was appointed as the company's first-ever independent Chair of the Board, permanently severing the imperial consolidation of CEO and Chair under Huang.

"Colette has helped lead NVIDIA through a period of extraordinary growth and has a deep understanding of our business, customers, operations, and financial position... The Board has full confidence in Colette and the leadership team, and we expect the company's operations, customer commitments, and product roadmap to continue without interruption."

— Harvey C. Jones, Chair of the Board

The immediate roadmap—the manufacturing ramp of Blackwell GB200 NVL72 and the tape-out of Rubin (scheduled for 2026/2027)—is locked into TSMC’s advanced packaging fabs. For the next 18 to 24 months, the machine runs on inertia.

The crisis begins on Month 25.


V. The Compute Vacuum: Will Hyperscalers Break Free?

For three years, Big Tech’s four biggest spenders—Microsoft, Meta, Google, and Amazon—have lived in uneasy vassalage to Santa Clara. Together, they represent over 40% of Nvidia’s revenue, handing tens of billions in pure operating profit directly into Nvidia’s 75% gross margins.

Every hyperscaler has spent billions preparing for this exact moment:

HyperscalerIn-House SiliconOpen Compiler VectorStrategic Objective
GoogleTPU v6 (Trillium)XLA / JAX ecosystemTotal internal independence from CUDA
Amazon (AWS)Trainium 2 & InferentiaAWS Neuron SDKDeliver 40% cheaper cloud training alternatives
MetaMTIA v2 (Artemis)PyTorch 2.0 / OpenAI TritonCommoditize underlying compute for open Llama models
MicrosoftMaia 100ONNX Runtime & DeepSpeedProtect Azure margins against custom silicon lockouts

With Jensen Huang commanding the room in his black leather jacket, hyperscaler CEOs had little choice but to pay whatever Nvidia demanded. He personally decided who received allocation of scarce H100s and who was forced to wait six months in the queue.

Without that founder-king leverage, can a committee-led Nvidia maintain the power required to demand 75% gross margins from the most formidable software monopolies in human history?


VI. The Architectural Verdict

Every great hardware empire faces this threshold:

  • When Steve Jobs passed away in 2011, observers anticipated Apple’s decline. Instead, Tim Cook institutionalized operations, preserved margins, and transformed Apple into a $3.5T services empire—even if revolutionary hardware leaps slowed.
  • When Bill Gates stepped down from Microsoft in 2000, the company drifted through a decade of strategic paralysis under Steve Ballmer before rediscovering its soul in the cloud under Satya Nadella.

Nvidia now enters its Tim Cook era. The supply chain execution will be rigorous. The balance sheet will be protected. But can an institutional committee summon the lunatic foresight required to stake the company's entire survival on an unproven computational paradigm fifteen years before it matures?


What are your thoughts?

Tags:AI ArchitectureNvidiaSemiconductorsGeopolitics