Produced by | Zhixie Island

In Seoul's vibrant Hongdae district, outside a bustling barbecue restaurant named 'Brother, It's Me,' a crowd exceeding a thousand gathered with bated breath.

Their anticipation wasn't for a K-pop idol but for a 63-year-old man clad in a sleek black leather jacket.

Inside, around a round table, four men sat in quiet anticipation. In Korean dining culture, where hierarchy reigns supreme, the eldest occupies the head seat, leading toasts with gravitas, while the youngest sits at the foot, tasked with grilling meat, pouring drinks, and managing the parade of dishes brought by waiters.

This etiquette, deeply ingrained in Confucian tradition, dictates that even billionaire chaebol leaders must adhere to age-based seating when entering a barbecue restaurant, a nod to the revered social order.

That fateful evening, the head seat was claimed by Chey Tae-won, the 65-year-old chairman of SK Group. To his left sat Lee Hae-jin, the visionary founder of Naver, and to his right, Koo Kwang-mo, the 48-year-old chairman of LG Group.

As the youngest at the table, Koo Kwang-mo, the third-generation heir to South Korea's fourth-largest chaebol, embraced his role as the 'maknae' (youngest member): he pulled tissues from a ceiling-mounted dispenser, placed them before each guest, poured water with precision, and then stood at the grill, deftly flipping pork belly with practiced ease.

Finally, Jensen Huang arrived, his entrance momentarily halted by the throng at the door. He entered, his black leather jacket still bearing the lingering heat of early summer. His seat was a guest spot, and upon sitting, he embarked on a crash course in Korean drinking etiquette.

Chey Tae-won tapped his spoon against a cup, mixing beer with soju to create a 'bomb shot.' As soon as the restaurant owner finished the demonstration, Huang eagerly grabbed the spoon and replicated the move. When Koo offered grilled meat, Lee Hae-jin taught him the art of wrapping kimchi and pork belly in perilla leaves. Huang followed suit, stuffing a large bite into his mouth and then, with gusto, picking up a chili pepper to dip in sauce.

As the meal drew to a close, Huang leaned down and inscribed on the table where he had sat: 'JENSEN WAS HERE,' followed by three emphatic words: 'LOVE LOVE LOVE.'

This was no ordinary dinner for the leader of the world's most valuable company by market cap; it was a meticulously choreographed performance, perfectly aligned with local cultural codes. In this script, the CEO of the world's highest-valued tech company was reduced to a foreign guest in need of instruction on how to eat barbecue.

After dinner, Huang stepped outside and made his way to a nearby fried chicken restaurant, where he began distributing honey-banana-flavored potato chips to the onlooking citizens. The next day, convenience store data revealed a staggering 776% surge in sales of that chip flavor compared to the same day the previous week.

Rewinding five days: in Taipei, Huang stood under the spotlight at the GTC conference, with a large screen behind him displaying Vera Rubin, RTX Spark, and AI Factory. He spread his arms wide, declaring, 'The era of Agentic AI has arrived.' Later that night, deep into his speech, he was photographed carrying a bag of fried dough twists at a night market.

Earlier still, Huang had been the last to join Trump's White House entourage, prompting jokes from mainland Chinese netizens: 'Don't go to impromptu dinners.' Huang hurriedly boarded a plane with just a backpack.

From his backpack at Alaska Airport to fried dough twists in a Taipei night market, and learning to wrap perilla leaves at a Seoul barbecue table, these three snapshots depict a 63-year-old man traversing three countries, appearing in three vastly different scenarios.

NVIDIA CEO Jensen Huang is becoming a globetrotting icon.

What is the essence of an icon? To be loved, chased, photographed—and to be guided by an unseen hand. Its path is not self-determined but set by festivals, crowds, and the team it represents.

In that Seoul barbecue restaurant, one of the world's most powerful tech leaders sat in a guest seat according to Korean rules, drank bomb shots, and learned to wrap perilla leaves with finesse.

Huang was complying with local cultural codes, adhering to supply chain priorities, and participating in a game he must win but cannot fully control.

The unseen hand guiding Huang has one end tied to NVIDIA's $5 trillion market cap and the other to the lifeblood of the global AI industry. He must keep moving, for if he stops, the crowd will disperse.

What is tech superstar Jensen Huang enduring, avoiding, selling, and fearing?

1. Commercial Radius: When a Person's Itinerary Becomes an Asset Pricing Model

Mapping Huang's flight paths over the past six months yields a power map of the global AI industry.

January: Shanghai, exploring a wet market in Pudong New Area; January: Shenzhen, dining at a hotpot restaurant in Futian CBD; May: Beijing, savoring Zhajiangmian (noodles with soybean paste) and fried squid, with Nanluoguxiang becoming Huang's 'pain street'; June: Taipei, delivering a two-hour speech at the GTC conference, announcing seven new chips and NVIDIA's entry into the AI PC market; June: Seoul, meeting with Hyundai, LG, SK Hynix, Samsung Electronics, and Naver, signing a multi-year HBM partnership with SK Hynix, and announcing a new AI R&D center in South Korea.

His itinerary rivals that of a head of state, with one key difference: heads of state travel for diplomacy, while Huang travels for market cap.

As is well known, NVIDIA's valuation model has undergone a transformation. Its P/E ratio has long exceeded 30x, implicitly assuming it is not just a hardware supplier but the sole builder of AI-era infrastructure.

This narrative requires constant fuel—new products, partnerships, markets, and concepts. Huang himself is the most direct conduit.

Every appearance by Huang accomplishes three tasks:

First, injecting certainty into the supply chain. At SK Hynix's booth, he wrote, 'Please produce more.' The value of this message lies in who delivered it. Capital markets interpreted it as: HBM orders are not slowing. The next day, SK Hynix's stock price stabilized.

Second, providing an anchor of confidence for capital markets. According to Caixin, South Korea's KOSPI Composite Index fell over 6% on the morning of June 5, with KOSPI 200 futures dropping more than 5% and triggering a five-minute trading halt; multiple South Korean chip giant stocks opened lower, with Samsung Electronics falling nearly 7% in early trading and SK Hynix dropping over 8%.

On June 8, during a press conference with SK Group chairman Chey Tae-won, Huang stated, 'If you're an AI stockholder, you should be happy. [Tech stock] prices are very low right now.'

A CEO hoping his words can stabilize another country's stock market.

Third, creating demand for next-generation products. At GTC, he devoted significant time to discussing Agentic AI, stating it requires Vera Rubin. Customers hadn't yet grasped what Agentic AI was but already felt the need to buy new chips.

This classic tactic of framing product roadmaps as technological trends contains a structural contradiction: the more successful he is, the more NVIDIA relies on him. The greater the reliance, the less his itinerary can pause.

Wherever Huang appears, opportunities arise; places he doesn't visit feel neglected. This 'person-centric' valuation logic means he must always be on the move, never absent from any market for too long.

The line blurs further when Huang personally recommends stocks. In Taipei, he said, 'Marvell could be the next trillion-dollar company.' The next day, Marvell's stock surged 32%, its largest single-day gain since 2000. In Seoul, he said, 'Buy Qualcomm stock,' and Qualcomm rose 3% pre-market.

Huang is not an analyst, does not disclose his holdings, and offers no risk warnings to retail investors following his advice. When a man worth nearly $200 billion and controlling global AI computing power says 'buy it,' it's no longer a suggestion but a signal.

Signals can be interpreted—or misinterpreted.

2. The Tightrope Walker Has No Safety Net

Huang likes to say, 'NVIDIA is a global company.'

Examining his travel map reveals a clear triangle: the United States (political protection), China (largest market), and South Korea (core supply chain). Huang must constantly move between these three points, as a fracture (rupture) at any one would cost NVIDIA dearly.

In the U.S., he is a meritorious outsider.

With a market cap exceeding $5 trillion, NVIDIA symbolizes U.S. tech hegemony, but Huang's personal position is more complex than outsiders realize.

During a 2025 podcast interview, the host asked him directly: 'Isn't selling AI chips to China like selling enriched uranium to the enemy?'

Most symbolically, Huang was not included in Trump's May 2026 delegation to China but was added via a phone call. This means that in Washington's corridors of power, Huang remains someone who must seek approval to enter.

His Chinese identity, technical background, and relatively low-profile political donations place him in a gray zone within U.S. politics: usable but not fully trustworthy.

In China, he is a wary partner.

Huang once said NVIDIA's market share in China's AI chip sector is near zero. The precision of this figure matters less than the trend: Huawei's Ascend is rapidly replacing NVIDIA, with domestic large models like DeepSeek running on Chinese computing power, and internet giants migrating training clusters to domestic chips.

It's not that NVIDIA's products are inferior but that customers cannot be sure they'll still be available tomorrow. When a product's supply depends entirely on geopolitical rivalry, the rational choice for businesses is to seek alternatives—a reality Huang understands well.

His acts of drinking douzhi (fermented mung bean juice), strolling through hutongs, and performing yangko dances at year-end parties have garnered billions of exposures on Chinese social media. But behind these gestures lies the harsh reality of losing the world's largest application market.

His affability feels like a valedictory performance: knowing he cannot change policy, he seeks to influence emotional ties to NVIDIA, leaving a door open for a possible return.

In South Korea, he is a client asking for favors.

The conventional view casts NVIDIA as the client and South Korean HBM manufacturers as suppliers. But a closer look at HBM4 capacity planning reveals a far more complex power dynamic.

SK Hynix's expansion won't come online until 2027, while NVIDIA's Vera Rubin is expected to ship at scale in late 2026. Each Vera Rubin chip requires stacking over a dozen HBM layers; without sufficient memory, even the most powerful GPUs are useless.

The capacity gap is not speculative but certain. Huang's message at SK Hynix's booth—'Please produce more'—was a plea, with a super-client using personal charm and influence to secure supply chain priority.

Geopolitics has torn the global tech supply chain into fragments. Huang stands in the cracks.

He cannot choose sides, as he needs every fragment. Instead, he must keep moving, using his feet to hold the pieces together and prevent the cracks from widening.

3. Huang's Non-Fiction Narrative

Huang is tech's master storyteller. His narratives share three traits: they are grand, self-consistent, and always position NVIDIA as the sole solution.

He declares, 'Agentic AI is here,' then tells you it requires Vera Rubin; he proclaims, 'Physical AI is next,' then introduces Jetson Thor; he asserts, 'Tokenomics is the future,' then emphasizes NVIDIA's performance-per-watt leadership.

Individually, these narratives have technical merit. Together, they form a precise closed loop: Huang defines the future, NVIDIA provides its sole infrastructure, markets price that future, and NVIDIA uses the proceeds to develop the next generation, defining an even more distant future.

This is an obvious strategy. In an AI era rife with uncertainty, following a leader feels safer than independent judgment. Every monopolistic company behaves this way.

The question is whether this loop can run indefinitely.

The first narrative worth examining: AI PCs.

NVIDIA and MediaTek's RTX Spark aims to redefine PCs. But supply chain forecasts suggest about 10 million devices based on this chip will ship over the next two years—a fraction of the 250 million annual PC market.

Renowned analyst Ming-Chi Kuo noted in a report: 'Neither current PC sales nor excitement reflect rigid demand for on-device AI computing.' Most users still access cloud AI via browsers; local large model scenarios remain niche, appealing only to developers and enthusiasts.

The second narrative: humanoid robots.

Huang's GTC appearance, standing amid nine humanoid robots, was visually striking. But industry reports highlight severe training data scarcity for robots, with Sim-to-Real technology still at a tipping point. Mature solutions for tactile sensors in dexterous hands and high-degree-of-freedom joint drives remain elusive.

NVIDIA provides training platforms and computing power—essential infrastructure. But platform maturity ≠ industry maturity. Smartphone chip adoption waited for iOS and Android ecosystems to explode. The 'iPhone moment' for humanoid robots remains distant.

More critically, NVIDIA faces a structural issue in this space: if robots proliferate, will value accrue to hardware (chips) or applications (operating systems, software, services)?

History shows that in mature tech ecosystems, hardware eventually commoditizes, with profits shifting to software and applications. NVIDIA's current strategy seeks to dominate both hardware and platforms, but its ability to hold this position depends on building an irreplaceable software ecosystem, as it did in data centers. In robotics, this ecosystem is still nascent.

The third narrative: Tokenomics.

'Computing power equals revenue; Tokens per watt equals profit margin.' This formula holds perfectly in NVIDIA's world because its business is selling computing power.

But for AI-using enterprises, what matters is application value, ROI, and problem-solving—not Token consumption.

Framing a business model as a universal economic law is brilliant marketing. NVIDIA currently controls Token pricing as the sole supplier. But if alternatives emerge—even theoretically, such as in-memory computing, photonic computing, or more efficient distributed architectures—Token prices will face competitive pressure.

A deeper issue is reflexivity.

Huang declares AI will explode, capital rushes in, computing power is purchased, his chips sell, and he can then say, 'See, I was right.'

This is not a prophecy but rather a self-fulfilling cycle. The cycle itself does not pose a problem—except that it obscures a crucial factor: end-user demand.

Capital can indeed purchase computing power, but this power must be harnessed for applications. If the monetization of applications lags behind the expansion of computing capabilities, overcapacity will inevitably ensue.

The dot-com bubble of 2000 serves as a cautionary tale. Cisco was selling routers at a breakneck pace, and telecommunications companies were laying down extensive fiber networks. However, internet revenues failed to meet expectations. When the bubble burst, Cisco's stock price plummeted from $80 to $13, and it took more than a decade to recover.

Huang often cites Cisco as a cautionary example, cautioning, 'I don't want NVIDIA to go through that.' What he doesn't mention is that Cisco's CEO, John Chambers, was also a tech luminary, frequently appearing in the media and at summits, proclaiming, 'The internet will change everything.' He was correct—just off on the timing. And when timing is off, everything else falls apart.

Today's Jensen Huang and the John Chambers of the past share a common risk: they are both staking their credibility on endorsing future timelines.

When credibility is deeply intertwined with a narrative, once the timeline of that narrative is proven incorrect, the collapse of credibility can be far more devastating than a mere decline in performance.

IV. Conclusion

Jensen Huang appears in Taipei night markets, clutching twisted dough sticks in his hand. He turns up at a Seoul barbecue restaurant, scribbling 'JENSEN WAS HERE' on the table. He is spotted at an Alaska airport, waiting to board with a backpack slung over his shoulder.

Each photo conveys the same message: I am just an ordinary individual.

But what the photos fail to convey is that a single word from this man can send a company's market value skyrocketing by 30%, his itinerary influences the pricing of assets worth trillions, and his narrative shapes the future direction of an entire industry.

The person most adept at defining the future is nonetheless consumed with anxiety over every supply chain detail in the present.

The term 'roving mascot' encapsulates praise, jest, and an indescribable sense of weariness. Jensen Huang has defined this era, yet he is also confined by it. He crafted the narrative of AI, only to find himself enslaved by that very narrative.

The tide continues to rise, and everyone is cheering. But no one knows when the tide will cease to rise, and no one knows if, after the ebb, the figure in the leather jacket will still be standing.

When the fate of an industry is tethered to one person's schedule, is this a triumph for the industry or a lurking peril?

All rights reserved by RoboIsland. No reproduction without permission.