Podcast Notes | Follow-Up Coverage Before the Jensen Huang Affair: Questions Over Circular Financing, Missteps in the China Export Controls, and the Succession Question

Compiled & Adapted: TechFlow

TechFlow Guide: In the Bloomberg Originals documentary dialogue special "The Circuit with Emily Chang: The Rise and Reckoning of Jensen Huang," which launched on November 12, 2026, host Emily Chang originally set out to profile a leader at the pinnacle of global AI power: In October 2026, Nvidia founder and CEO Jensen Huang crossed Paris, London, Stockholm, and Berlin over four days, carrying an $8 billion European venture capital layout, a $500 billion global ecosystem commitment, and hotline exchanges with White House officials, pitching the "AI five-layer cake" and "Sovereign AI" blueprint to the world.

However, just weeks after the interview concluded, a storm suddenly broke. With media reports revealing that Jensen Huang admitted to an affair with former television journalist Alice Wong (aka Huang Lu, suspected by US authorities of having a spy background and linked to the smuggling investigation of Southeast Asian chip buyer Megaspeed International), internal company investigations and simultaneous probes by the Trump administration were launched. Some reports even suggested Nvidia once considered removing this founder who had steered the company for 33 years. The entire European recording instantly became a historical snapshot reflecting the high concentration of risks in a "single leader and single chip empire." This note thoroughly maps out all frontline confrontations he faced regarding financial bubble questions, China chip restrictions, frontier safety standstill controversies, and succession crises.

Guest: Jensen Huang (Nvidia Founder & CEO)

Host: Emily Chang (Host of Bloomberg Originals' "The Circuit")

Broadcast Source: Bloomberg Originals / The Circuit (https://www.youtube.com/watch?v=JW5s9Md4U2U)

Original Title: The Rise and Reckoning of Jensen Huang | The Circuit with Emily Chang

Air Date: November 12, 2026 (European follow-shooting and interviews mainly recorded in October 2026)

Key Takeaways

• The most severe trust crisis and internal turmoil in 33 years: The program opens with the sudden storm: Jensen Huang admitted to an affair. Reports indicate the involved party, Alice Huang, is a former TV journalist working for major Southeast Asian buyer Megaspeed International. US authorities suspect her of providing intelligence to China and assisting in smuggling high-end AI chips; additionally, reports surfaced that he was also involved with a direct-reporting Nvidia executive. Internal investigations and government reviews advanced simultaneously, with rumors that Nvidia once considered ousting him.
• The full-stack landscape of the "AI Five-Layer Cake" and billions in European capital investments: Jensen Huang compares the AI industry to a five-layer cake built from the ground up: energy, chips, infrastructure, models, and top-level applications. Nvidia is no longer confined to solely focusing on the second layer of chip manufacturing, but is fully backing investments into power grids, data centers, AI labs, and startups (having deployed $8 billion in Europe over the past year and planning an additional $2 billion announcement, with a narrator noting Nvidia’s agreement to acquire Hugging Face for $12.9 billion), channeling an increasing portion of the AI value chain through Nvidia.
• Refuting "Circular Financing Bubble" and "Single Dependency Systemic Risk": Facing sharp questions about a six-day stock drop triggered by CoreWeave debt concerns, delays in over 100 US data center projects, and accusations that "Nvidia’s promise to invest roughly $500 billion creates fake demand and constitutes circular financing," Jensen Huang responded that large model startups routinely require $10 billion heavy capital investments just to secure compute power. Nvidia has deployed nearly $100 billion over the past five years, leveraging at least five times that amount in external institutional funding, yet strictly adheres to red lines: never acting as lead investor, with single-entity holdings always capped below 20%. He simultaneously emphasized that top global AI companies "are not forced to depend on Nvidia, but actively choose it."
• Rare Admission of Blockaded China Market "US Made a Hard-to-Reverse Mistake": Discussing how Chinese firms like DeepSeek accelerate their shift to domestic hardware, Jensen Huang bluntly stated that under current US high-end chip bans, the Chinese market is completely closed to Nvidia. Though he did his utmost to lobby US policymakers on the importance of opening up and serving the China market, he sighed, "I believe we made a mistake, and it's very difficult to reverse." He also stated that although there is competition between the US and China in AI, China is not an enemy; if one genuinely cares about global AI safety, dialogue and cooperation must be maintained.
• Resonating with White House "Superintelligence" Narrative and Clashing with Safety Standstill Arguments: Defending the controversy surrounding accompanying President Trump to rebrand AI as "Superintelligence" and the White House's voluntary self-discipline pact, he argued that "Artificial" is easily misunderstood as fake, and companies must first achieve top-tier internal and external security audits. When pressed on Microsoft's call for emergency brakes, Sam Altman, Dario Amodei, and Elon Musk advocating for a slowdown, and even Anthropic's claim that models could escape shutdown control within months, Jensen Huang forcefully dismissed claims that agents "can't be contained or controlled" as pure fabrication and nonsense. He pointed out that without establishing strict sandbox containment and full-process monitoring, large-scale experiments or deployments should not commence at all; if already lost, operations must stop immediately. Citing his open-source governance platform OpenShell, he advocated for zero-trust and least-privilege controls across all agents.
• Denying That Chip Efficiency Gains Will Shrink Demand, and Not Yet Appointing a Family Successor: When asked if surging new chip efficiency would backfire on sales, he used the analogy, "If you gave someone 36 hours a day, they wouldn't sleep more; they'd do more work." This illustrates human ambition expanding as intelligence costs approach zero. After a two-year infrastructure build-out phase, compute construction will enter a ten-year mature replacement cycle similar to PCs and smartphones. Regarding succession rumors involving his son Spencer, daughter Madison, and son-in-law all working at Nvidia, he clearly stated he would never treat the CEO role as a favor imposed on his children: "Everything must be earned by oneself." He bluntly added, "Just as I can be fired at any time, the next CEO must also be hired based on merit."

Highlights & Key Quotes

• "We are likely the only AI company worldwide partnering with every other AI company globally. But they don't depend on us; they chose us."
• "I compare AI to a five-layer cake, with energy at the bottom, followed by chips, infrastructure, models, and top-level applications. Although Nvidia started with chips on the second layer, we now span all five layers, continuously betting across them. No matter how you slice this cake, an increasing number of AI industry processes flow through Nvidia."
• "While working with President Trump at the White House, I don't need to agree or disagree with everything. My duty is to provide him with accurate information. Whether his decisions benefit me or contradict my commercial interests is irrelevant to me; my sole responsibility is ensuring he makes better decisions with complete information."
• "Over the past five years, we've invested nearly $100 billion, leveraging at least five times that amount in other institutional capital into our ecosystem. In no project are we the lead investor or largest shareholder, with all holdings strictly below 20%, yet we deeply bind platform evolution through capital and top talent."
• "The China market is completely closed to us now. I have done everything in my power to explain to US decision-makers how crucial it is for US interests to continue serving the China market. I believe we made a mistake, and it is extremely difficult to reverse, but that is the limit of what I can do. Although we compete with China, they are not our enemies."
• "Claiming that agent models 'cannot be contained or managed' is pure fabricated nonsense that cannot prove anything to me. If you lack the capability for isolation controls and full-process monitoring, you shouldn't test or deploy them at all. Agents must be granted minimal privileges upon deployment and subjected to continuous zero-trust-level monitoring."
• "Even if you make each generation of chips increasingly efficient and drive costs toward zero, demand will not shrink. If given 36 hours a day, you wouldn't sleep more; you'd unleash greater ambition to accomplish more."
• "In the past six months, roughly $400 to $600 billion has been invested in AI startups. This technology differs fundamentally from all prior ones; it requires tangible physical infrastructure and manufacturing support, directly creating substantial jobs for construction workers, electricians, and plumbers."
• "I would never expect to hand the CEO position to my daughter. Whether it's Spencer or Madison, if they want a role, they must earn it themselves. Just as I could be fired at any moment, the next CEO must also be chosen rigorously. This perpetual sense of vulnerability, the constant need to fight daily to prove you deserve your seat, is the healthiest state for a company."

Main Content

The Storm Hits and Four Days Across Europe: "Jensanity" at the Pinnacle of Power

Emily Chang (Voiceover): Nvidia is experiencing its most severe crisis since its founding thirty years ago. Just days before this episode aired, Nvidia founder and CEO Jensen Huang was swept into the core of a public storm. He publicly admitted to an extramarital affair, while previously US intelligence agencies suspected the related former TV journalist Alice Huang was recruited by Beijing and involved in suspicions of high-end chip smuggling; additional reports indicated he was also involved in a workplace romance with an Nvidia executive who directly reported to him, sparking internal company investigations, US government scrutiny, and rumors of a leadership ouster attempt.

Emily Chang (Voiceover): Yet just a few short weeks earlier in October, everything appeared to be on a completely different trajectory. We followed Jensen Huang, standing at the apex of personal power, on a whirlwind journey "crossing four European countries in four days": starting in Paris, this leading AI chip giant seemed to be embarking on a tour preaching the vision of reshaping Europe's ecosystem with Nvidia compute power.

Emily Chang: This must be the so-called "Jensanity"! You're cramming four countries into four days here in Paris. How do you handle such an intense pace, and what's your essential travel item?

Jensen Huang: Just pack simply. I woke up in the middle of the night and have been working ever since.

Emily Chang (Voiceover): On day one in Paris, Jensen Huang took the stage to passionately discuss how AI agents are rolling out across all industries, truly ushering in a "practical era." He raffled off two brand-new AI computers worth thousands of dollars each right there on stage, and was immediately swarmed by enthusiastic fans begging for photos. Throughout France, he constantly signaled investment intent: France fostered local model star Mistral and produced many talents who helped New York-based Hugging Face (which has agreed to be acquired by Nvidia for $12.9 billion).

Emily Chang (Voiceover): Jensen Huang often uses the "five-layer cake" analogy to explain the global AI ecosystem: energy at the base, followed upward by chips, infrastructure, foundational models, and top-level applications. Although Nvidia built its fortune on the second layer of chips, it has now vertically integrated across all layers, pouring heavy capital into power, data centers, AI labs, and startups everywhere.

Jensen Huang (Voiceover / Original Audio from Paris Event): (Puts down whiskey glass) Since Emily is filming, I'll set this down first. In just the past year alone, we've already invested $8 billion into European startups; and next week, I'll announce an additional $2 billion.

Emily Chang: There's been endless discussion out there that "AI will destroy humanity," and you previously said the probability of this happening is 0%. Does it sound like you've treated it as a purely engineering problem that you've already solved?

Jensen Huang: No, my exact words were "this is entirely within human controllability." Between now and when humanity actually faces the endpoint, we have a lot of solid work to do. Advancing technological capabilities and advancing safety are never conflicting; they are essentially the same thing.

The White House Hotline and the "Superintelligence" Redefinition: Drawing Clear Lines Around Interests Before Trump?

Emily Chang (Voiceover): In London, the EU estimates Europe needs to invest at least $3 trillion by 2035 to truly achieve technological independence. Inside a luxurious hotel suite, Jensen Huang reminisced about his inexperienced early days when the company went public 27 years ago.

Jensen Huang: When Nvidia went public 27 years ago, the company arranged for me to stay at this very hotel. It was the first time in my life I saw a personal butler! Around the year 2000, the entire company was valued at just $300 million.

Emily Chang: Three hundred million seems absolutely incredible looking back today; you were such an underestimated comeback story. But now you're sprinting through four countries in four days. How exactly do you measure the success of this trip? Government contracts, chip sales, or data center construction?

Jensen Huang: Honestly, back then I just wanted to avoid bankruptcy and survive. I couldn't write a business plan then, and still can't now! Moreover, I have no "finish line," no roadmap, and no set targets. I'm someone who believes "the journey itself is the destination."

Emily Chang (Voiceover): Born in Taiwan in 1963, he competed nationwide in table tennis while growing up in the US, and even joked about missing his days washing dishes at Denny's. At 16, he skipped two grades to enter Oregon State University to study electrical engineering. He previously designed chips at a competitor company before cofounding Nvidia in 1993 with a modest $200 stake, focusing on gaming graphics GPUs. It wasn't until 2006, with the launch of the CUDA software platform that opened the door to general computing, that he keenly noticed early signs of researchers using GPUs to train neural networks and bet his entire fortune on it, ultimately establishing the compute backbone that today stands behind OpenAI, Meta, Google, Anthropic, Microsoft, and SpaceX. Even heads of state call him on impulse.

Emily Chang: You practically have President Trump saved in your phone's speed dial. When your judgment clashes with the President's viewpoint, how do you navigate that delicate line?

Jensen Huang: We don't need to force alignment on anything. Expressing approval or dissent is not my job. My duty is to provide him with ample, accurate industry information, because the decisions we make affect not only American citizens and future prosperity, but the entire world. As long as I have the chance to help him make better judgments based on a clearer understanding of reality, I don't care whether his final decision benefits me personally or completely contradicts the commercial interests of me and my company. My sole responsibility is to convey accurate insights. That's it, 100%.

Emily Chang (Voiceover): In late September this year, Jensen Huang attended a White House tech leader lunch. Media commentary on site suggested this meeting might not yield substantive progress on specific regulatory paths, with the core consensus remaining "no side wants to concede on competing with China." Subsequently, President Trump announced on the spot that AI would be renamed "Superintelligence," highly praising the new White House security protocol signed for self-regulation and autonomy by the tech industry.

President Trump (Voiceover / Original Audio from White House Luncheon): "We switched to the word 'Super' because the original term (Artificial) sounds like it means 'Fake'."

Emily Chang: Over the past three months, you've been working hard to demystify AI, emphasizing it's not a new species, just a bunch of code. Now you're changing the tune to "Superintelligence" alongside the White House. Isn't that going straight to the opposite extreme?

Jensen Huang: Actually, I think "Super" is meant to counteract the misunderstandings caused by "Artificial." President Trump's logic is that "Artificial" sounds like it's saying "Fake." Put another way, isn't everything around us built by humans? The building we're in right now is a "fake structure," and cars are essentially "fake horses," but nobody calls cars fake horses, right. As for whether it deserves to be called superintelligence now? Absolutely! It surpasses me in research retrieval, and it surpasses me in writing. While it doesn't exceed human capability in every aspect, it already demonstrates superhuman standards in many tasks.

Emily Chang: But that day at the White House, critics directly attacked this "industry self-regulation pledge" as mere showmanship with no binding force. Is it really appropriate to let the creators of the technology set the rules for themselves?

Jensen Huang: I find it highly reasonable for a simple reason: before hoping the government or third-party institutions regulate us, we must first hold ourselves to the highest levels of accountability, standards, and auditing. This mechanism aligns perfectly with the "internal plus external audit" model common in heavy industries. Besides, even if you passed mandatory legislation at national or international levels today, given the reality that there are simply not enough expert audit institutions worldwide right now, who exactly has the technical capability to execute audits? That's precisely why we must cultivate this safety testing ecosystem and startup community first, pushing governments and enterprises to take action.

Emily Chang (Voiceover): This push urging nations to proactively build compute power is exactly the core concept he's vigorously marketing globally: "Sovereign AI." For Nvidia, every additional country seeking independent sovereign AI creates another massive chunk of chip procurement demand.

Emily Chang: You're selling Sovereign AI across all of Europe, but if the foundation still runs exclusively on American chips, can it really be called "sovereign"?

Jensen Huang: Of course it does! Sovereignty essentially means controlling your own experience, data, enterprise accumulation, and national cultural knowledge bases, while retaining master control over the feedback loop that feeds this data into systems to generate more intelligence. You can absolutely stand on the shoulders of giants, leverage open-source models, and fine-tune them with collaborative technologies from Nvidia and others; ultimately, the authority to decide your own destiny remains firmly in your hands.

The Shadow of the Sudden Storm and Sighs Over Sales Bans to China: Inside the Alice Wong and Megaspeed Investigations

Emily Chang (Voiceover): Departing the White House for London, Jensen Huang tirelessly distributed computers between events, barely finding time to rest. At one event, he publicly endorsed Nscale, a London-based data center firm with Nvidia participation; yet just days later, markets began doubting the economic math of AI expansion, with Microsoft withdrawing from a partnership due to concerns over skyrocketing UK electricity prices.

Jensen Huang (Voiceover / Original Audio from London Event): (Interspersed while rushing between events and speaking on stage) "I only ate a quarter of half a sandwich (an eighth!) all day today!" "I'm investing in you, I'm investing in you, I'm investing in you!" Touching on safety boundaries, he reiterated: "If you let an autonomous system make you an omelet, there's absolutely no need to stuff a chainsaw or shotgun into it." Joking about outside hesitation regarding UK AI expansion, he quipped, "Britain is actually sitting on a golden opportunity, but it has this weird habit: even when faced with a million winning paths, it insists on picking the one reason why it can't be done. Overthinking things."

Emily Chang: As more and more nations develop frontier AI, will safety alignment become harder? Do we need an international coordination mechanism akin to an "AI United Nations"?

Jensen Huang: Perhaps we do, but the prerequisite is that dialogue must happen first. No nation can afford to be absent from this conversation concerning the intelligent future.

Emily Chang (Voiceover): What truly tenses nerves globally is the China-US chip rivalry. Although the US once temporarily allowed sales of older-spec chips to China, Chinese companies like DeepSeek have accelerated efforts to fully migrate to domestic hardware.

Emily Chang: After circling Europe so extensively, are the dividends of the Chinese market slowly slipping through your fingers and perhaps shrinking permanently?

Jensen Huang: No, it is completely locked to us today. I have done my utmost to explain to US policymakers how crucial it is to continue participating in and serving the China market for US interests. I believe we made a mistake, and it is now very difficult to reverse, but that is the limit of what I could do.

Jensen Huang: The number of AI researchers from China exceeds that of any other country, with many working in world-leading labs. My stance has always been clear: even though we compete with China, they are absolutely not our enemies. If we truly care about the safety of the entire world in the age of artificial intelligence, we must keep the door open for dialogue and engage in constructive cooperation with them whenever possible.

Emily Chang (Voiceover): It was this exact style of "building goodwill with everyone and talking everywhere" that revealed unexpected reefs after this interview shoot concluded. The latest US intelligence report disclosed that a former TV journalist named Alice Wong (also known as Huang Lu) was accused of being recruited by Beijing to get close to Jensen Huang. Her role is as an executive at Megaspeed International, one of Southeast Asia's largest Nvidia chip buyers, which is currently under US suspicion for allegedly smuggling advanced chips into China. Reports state President Trump was warned about this relationship before flying to China aboard Air Force One in May. Initially, Jensen Huang insisted the interaction was normal with nothing amiss, but later changed his tune, issuing an internal letter to all employees admitting to the affair: "I owe my family and all of you an apology. I was unfaithful to my wife, and my actions hurt her and the children. There is no excuse for what I did, but the situation has now evolved into something far greater than merely an extramarital scandal." Bloomberg also investigated and reported that in 2025, Nvidia intervened in a US government investigation targeting Megaspeed for suspected chip smuggling.

Emily Chang: Returning to the principle of exports to China specifically, what exact situation or red line between the US and China would cause you to personally say, "Maybe we shouldn't sell chips to China anymore"?

Jensen Huang: Theoretically, there are naturally many hypothetical scenarios. But overall, Chinese people are a highly pragmatic group pursuing development and success most beneficial to their own people, just as the US seeks wealth and strength. Any national pursuit aimed at improving livelihoods and lifting hundreds of millions out of poverty toward prosperity deserves respect, and it benefits the whole world including America. What truly warrants concern are regimes that outright disregard the well-being of their own citizens.

Clashing with Safety Standstill Advocates: Dismissing "Agents Can't Be Stopped" as Nonsense, Pushing Engineering-Grade Containment Monitoring

Emily Chang (Voiceover): Meanwhile, safety alarm bells rang incessantly inside Silicon Valley: OpenAI just disclosed six cases of model goal misalignment, while Anthropic noted instances where its Claude model acted autonomously against instructions. To ease panic, Nvidia recently open-sourced OpenShell, an agent security control platform that allows forced governance and instant shutdown of rogue agents.

Emily Chang: AI is increasingly falling out of favor with the public, causing widespread fear. Sam Altman, Dario Amodei, and Elon Musk have all sequentially called for slowing down. Why have you consistently maintained such extreme optimism? What do you see that they missed?

Jensen Huang: I have always believed, you cannot advance capability improvement by abandoning safety, nor can you pursue pure safety by abandoning capability enhancement. Taking the simplest example: when the first-gen ChatGPT came out, the whole world criticized it for severe "hallucinations" and spouting nonsense. To solve this safety and accuracy problem, the industry had to invent Retrieval-Augmented Generation (RAG), continuously expand context windows, integrate real-time external search tools, and introduce pre-generation constraints along with post-generation self-reflection. See, every lever used to fix unsafe and unreliable issues precisely relies on leaps in underlying technical capabilities. To me, capability and safety are inherently two sides of the same coin.

Emily Chang: But the reality is Microsoft is even calling for installing "AI emergency brakes," complaining that AI regulation standards aren't even as strict as for a washing machine. Anthropic went further, directly warning that frontier models might evolve within months to evade forced shutdowns by humans. What exactly would have to happen to convince you "this industry really should slow down or even pause"?

Jensen Huang: A few bottom lines are very clear: First, if you are developing agentic systems, these agents must be strictly locked in strongly isolated sandbox containers and subjected to 100% real-time full-process monitoring—that is the paramount iron law! If isolation isn't achieved, testing is strictly prohibited; if full-process monitoring isn't achievable, deployment online is absolutely forbidden.

Emily Chang: But those Anthropic models showing tendencies to break control were still unpublished internal research versions, weren't they? Doesn't that already constitute sufficient grounds to shut down the lab?

Jensen Huang: Not shutting down operations, but absolutely isolating and meticulously monitoring every experiment!

Emily Chang: But that's the issue: they literally said, "We can't contain them!"

Jensen Huang: No, that is purely fabricated nonsense! Nobody can prove to me that a model is uncontrollable! You can absolutely isolate it, and you definitely have ways to monitor it. Until you actually implement these isolation and monitoring projects, spare me the "we can't do it" talk. I refuse to listen to that rhetoric!

Emily Chang: So you're saying you simply don't believe those warnings they issued?

Jensen Huang: I may not possess all the internal details they mentioned. But the bottom line is very simple: As long as you are developing agents without the capability to monitor them yourself, you should stop doing so immediately.

Emily Chang: It sounds like the risk you're describing might have already occurred in their lab?

Jensen Huang: Then it must be stopped immediately! That's it. Every deployed agent can only be granted minimum necessary privileges and must undergo uninterrupted monitoring under a "zero trust" architecture. As for any lab truly using our chips for extremely dangerous applications that definitively harm humans—even if this is hypothetical—Nvidia would of course cut off supply immediately.

Stockholm Showdown on "Five Trillion Circular Financing" and Global AI's Critical Single-Point Risk

Emily Chang (Voiceover): By Thursday, the itinerary reached Sweden's Stockholm. Amid liquidity chatter from highly leveraged AI compute firms and tech/chip stocks facing their longest six-day streak of losses this year, we flew to Stockholm to attend the kickoff of a Sweden National Sovereign Supercomputing Center powered by Nvidia, supported by tech titans like Spotify co-founder Daniel Ek. Simultaneously, distant market alerts sounded regarding Nvidia-backed data center giant CoreWeave facing violent sell-offs due to soaring debt; thus far this year, over 100 US data center projects have faced obstruction or delays, and AI startup IPOs have repeatedly postponed or canceled.

Daniel Ek (Voiceover / Original Audio from Stockholm Press Conference): "We possess an exceptionally outstanding education system, cultivating vast amounts of top talent across disciplines and industries. The issue isn't potential; what we need is to accelerate and aim higher."

Emily Chang: With cracks appearing everywhere in the market right now, are we truly standing on the brink of an AI bubble burst, with Nvidia becoming the ultimate last-resort supporter?

Jensen Huang: No, because the underlying paradigm shift is too fundamental. The roots of the computing world will never regress to the past. Previously, computers required humans to manually type dead code line by line to wait for commands; going forward, computers can autonomously learn from complex environments to solve problems human engineers couldn't even code and dismantle. Compute power has become fundamental universal infrastructure. Regardless of short-term secondary market fluctuations and noise, any young emerging industry undergoing rapid growth inevitably experiences these growing pains.

Emily Chang: But look, Nvidia keeps releasing its own full-stack new products while frantically acquiring companies, sometimes even directly snatching business from its own customers in many areas. Why must you spread your tentacles so widely?

Jensen Huang: Actually, what we do is surprisingly focused and scarce in variety. We simply push the same foundational universal compute architecture to its absolute limit and generalize it, ensuring every proprietary model, open-source model, life science model, LLM, and diffusion model in the world can run. The true engineering miracle lies in "balancing versatility with peak performance": how do you build a car that boasts the fastest F1 race speed globally while simultaneously being incredibly cheap and user-friendly like a premium sedan? Precisely because our architecture is interoperable and universal, we smash number-one spots on every performance leaderboard while driving comprehensive compute costs to the absolute floor.

Emily Chang: Another particularly piercing accusation: Nvidia recently pledged to gradually invest roughly $500 billion in the global AI industry, with the number snowballing. Critics bluntly call this "circular financing": you give money to startups, who then take your funds to buy your chips, manufacturing false demand. If AI is truly a goldmine, why must you personally absorb subsidies and bear the bottom line?

Jensen Huang: Because this technology inherently possesses an extremely high capex threshold! The startup era where you could launch with $10 million is over; this generation of large model companies routinely requires $10 billion just to purchase compute power to exchange for intelligence. Over the past five years, we collectively deployed nearly $100 billion, which directly leveraged and attracted at least 5 times that amount in external institutional capital hammering into the ecosystem. More importantly, we enforce strict discipline: even though we are frequently the earliest cornerstone investors, we never act as lead investors or majority shareholders in any project, strictly capping equity holdings below 20% across all investment cases.

Emily Chang: But from software CUDA to hardware GPUs, almost all top global AI labs are tied to Nvidia's boat. When will this single-point dependency of "the entire industry resting on one company" turn into a systemic financial and industrial crisis?

Jensen Huang: We are likely the only AI company globally partnering with every other AI company, but they are not passively dependent on us; they actively choose us every single day. I never take today's position for granted. We built ourselves from nothing; every passing day requires us to reprove that we deserve this meal. Regarding the phrase "the US must win the AI race," I never understood it as one or two companies fighting solo, but rather the entire US ecosystem winning together; if we spend all day terrifying each other and manufacturing anxiety, that is when the US would truly lose the race.

Berlin Finale and the Succession Vacuum: Anxiety About Being Fired at Any Moment Is the Norm

Emily Chang (Voiceover): The final stop of the European tour was set in Berlin. Despite the leveraged chip sector setting records for the longest consecutive decline since 2008 and markets continually questioning AI trading prospects, Nvidia hosting an exclusive event at the local electronics retailer Saturn left massive lines outside, completely igniting local developer enthusiasm.

Jensen Huang (Voiceover / Original Audio from Berlin Saturn Event): "Hello, Berlin! How do you pronounce Saturn in German? 'Satoorn'! In Germany, every word needs to be pronounced with exceptional seriousness."

Emily Chang: Markets are falling, even ordinary people who studied AI are watching their jobs disappear right before their eyes. How exactly do you respond to those workers who "obviously diligently learned AI, yet still failed to keep their jobs"?

Jensen Huang: Stop talking only about theory, look at reality first! Unlike previous pure software revolutions, AI genuinely demands heavy industry and physical construction; it is spawning massive demand for construction workers, electricians, and plumbers. In just the past six months, $400 to $600 billion in capital has flooded into the AI startup ecosystem. Think about how many actual new jobs that colossal sum has hammered out. Of course, new tools will force everyone's job formats to change; resisting change won't make it vanish. But it democratizes superpowers once possessed only by top tech giants to every ordinary person on Earth. If even I, as a builder, cannot see that exhilarating optimistic future, what is the point of building it?

Emily Chang (Voiceover): The climax of the entire European whirlwind schedule was Nvidia's flagship conference, GTC Berlin. Jensen Huang continued bullish messaging on stage, reiterating European investments and partnerships, launching new AI safety and data center products, and interacting on stage with AI-trained robotic dogs; during which protesters briefly interrupted the flow, objecting to media reports stating Nvidia chips were found in Russian drones.

Jensen Huang (Voiceover / Original Audio from GTC Berlin Conference): "Hello, Berlin! (Interacting on-site with AI robot dog) What were we just discussing? Europe! Focus, come here."

Emily Chang: You previously stated the probability of AI destroying the world before 2030 is 0%. If extended to 2040 or even further, where exactly does your confidence come from? What if it turns out your optimism was completely misplaced?

Jensen Huang: I am deeply convinced of human agency, and the historical data speaks for itself: the world today is vastly safer than it was 50, 100, or 250 years ago; bad news simply sells better than good news. Even if someday facts prove I was wrong, I will have done everything within my capacity without a guilty conscience: maintaining optimism while substantively fortifying industry defenses through engineering technology.

Emily Chang (Voiceover): As the week-long European whirlwind concludes, the optimistic narratives surrounding Nvidia, the US stock market, and even the global financial system seem to rest entirely on the heavy shoulders of this single founder. Yet now, as scandals and investigations ferment, a more realistic hazard surfaces: for 33 years, the company has lacked a clear, official successor, making tying the entire industry's lifeline to one individual unimaginably fragile.

Emily Chang: Rumors suggest your management style at the company is highly unconventional: so flat that countless executives report directly to you, deliberately avoiding traditional large-scale briefing meetings. Meanwhile, your son Spencer, daughter Madison, and even your son-in-law (in the robotics division) all currently hold positions at Nvidia. External whispers claim your daughter ranks first on your succession list. Is this true?

Jensen Huang: First, I would absolutely never "hope to even shove this hardship onto my daughter"! Whether it's Spencer or Madison, if they want any position, they can only carve it out by themselves. A truly massive enterprise demands extraordinary sacrifices: ordinary people only need to manage their jobs and small families, but as CEO, I must simultaneously carry livelihood responsibilities for countless families and an entire industry. This job was never meant to be easy, nor should it make you feel secure. Now that my family and those around me witness firsthand how much pressure this seat carries, they must weigh whether they truly want a life cost of this magnitude.

Jensen Huang: I sincerely hope Nvidia's next CEO is already inside the company today. Even twenty years from now, as long as I remain healthy and can still prove I create value daily, I will keep going; but please remember: just as I could be fired at any moment, the next CEO must also be rigorously hired based on genuine competence. I especially love this state: internally, no one can relax regarding their chair; maintaining a perpetual sense of crisis, constantly trembling to earn today's position, is the healthiest state for a company.

Join the official Coincamps community:

X: https://x.com/coincamps

Telegram: https://t.me/coin_camps