Nvidia's Jensen Huang Takes Seoul by Storm with Robot Partnerships as RTX Spark Rattles PC Rivals

CEO Jensen Huang visits Seoul for Physical AI deals as Nvidia unveils the RTX Spark desktop chip with MediaTek, shaking up Intel, AMD, and Qualcomm.

Nvidia is waging a two-pronged campaign to cement its AI dominance, and this week the action is split between the boardrooms of South Korea’s largest industrial conglomerates and the consumer-electronics battlefield of the Windows laptop. CEO Jensen Huang jets into Seoul on June 4 for a series of high-stakes meetings that could turn the chipmaker from a mere hardware supplier into the central infrastructure partner for the global robotics industry. The trip comes hot on the heels of the Computex keynote in Taipei, where Nvidia unveiled the RTX Spark – a desktop processor that threatens to upend the PC market and has already sent shockwaves through Intel, AMD and Qualcomm.

The Korean offensive is all about "Physical AI" – the large models that power humanoid robots and autonomous driving. Huang will sit down with the chairmen of SK Group, LG Group and Naver on June 5, followed by Hyundai Motor and the NCsoft founder two days later. The agenda: embedding Nvidia’s Omniverse, Isaac and Cosmos platforms into Korean smart factories and autonomous systems. The stock market has already taken notice: LG Electronics has surged 329% since the start of the year, and Hyundai Motor has climbed 144%. A closed-door roundtable with local AI startups is scheduled for June 8, likely to flesh out the next concrete steps.

Back in Taipei, the RTX Spark stole the show. Developed jointly with MediaTek, the chip marries Nvidia’s Blackwell GPU with a MediaTek CPU core on a single system-on-chip, sharing unified memory to eliminate the bottleneck that plagues local AI workloads. Nvidia claims 1 petaflop of FP4 performance, and laptops from Microsoft, Dell, HP, ASUS, Lenovo and MSI – all powered by the new silicon – will hit the market later this year. Morgan Stanley expects the high-end N1X variant to start at $2,899, while the pared-down N1 will come in at $1,799. The message is clear: big language models can now run entirely on-device, no cloud required.

The announcement hit Nvidia’s rivals like a sledgehammer. Intel lost nearly 4.7% on Monday, falling to $109.33; Qualcomm shed 8.8%; and AMD also slipped into the red. Arm Holdings, by contrast, jumped 16% on the licensing revenue every ARM-based chip brings. MediaTek itself gained over 5% in Taipei. Nvidia’s own shares added 6.3% in New York. In Frankfurt, the stock traded at €192.86 on Wednesday, up 0.66% – a modest gain that leaves it nearly 20% higher year-to-date but still about 5% below its 52-week high of €202.50.

The financial firepower underpinning both campaigns is formidable. Nvidia has guided for second-quarter revenue of around $91 billion, well above the consensus estimate of $86.84 billion. The first quarter delivered a record $81.62 billion, driven by 92% year-on-year growth in the data-center segment. A colossal $80 billion share-buyback program underscores management’s confidence, while the company’s credit default swaps trade at a risk level comparable to US Treasuries – a strong vote of confidence from institutional investors. Goldman Sachs reiterated its "Buy" rating with a $285 price target, pointing to the ramp-up of AI factories as a long-term growth driver.

Nor is Nvidia resting on its laurels in the data center. Huang confirmed that the next-generation Vera Rubin platform is already in full production, with first systems heading to customers such as Anthropic, OpenAI, xAI, Dell, Oracle and CoreWeave in the autumn. At the same time, the company deepened its ties with Marvell Technology, injecting $2 billion to integrate Marvell’s custom network solutions into Nvidia’s NVLink Fusion platform. Huang described Marvell as a potential "next trillion-dollar company".

Analysts are split on how quickly the RTX Spark will translate into revenue. Seaport Research’s Jay Goldberg called the chip technically promising but warned that Windows on ARM has not yet matched x86 systems, making broad adoption "multiple generations" away. Truist’s William Stein, however, maintained his buy recommendation with a $307 target, and Cathie Wood snapped up 300,017 Nvidia shares after the Computex keynote. The AI PC narrative, first floated by Microsoft in 2024, has so far delivered few real-world sparks – but Nvidia’s unmatched credibility in AI could finally ignite it.

For Intel, AMD and Qualcomm, the danger is not an overnight market loss but a slow erosion of the most lucrative segment: premium laptops where customers are willing to pay a premium for AI features. Nvidia now has its foot firmly in the door. Whether the giant fully steps through will depend on the next product generation. Meanwhile, in Seoul, Huang is laying the groundwork for an entirely different kind of AI revolution – one that takes place not on a desk, but on the factory floor.