By Song Jia Nan | Tech & Markets Correspondent
On June 2, during a media briefing at the GTC Taipei 2026 event at Computex, Nvidia CEO Jensen Huang didn’t shy away from the ongoing industry chatter surrounding pay disputes at SK Hynix and Samsung Electronics. He made his stance on the matter pretty clear.
In Huang’s view, companies really “should reward their employees to the fullest extent possible.” While he was quick to note that commenting on how other firms handle their payrolls isn’t exactly his place, he did share a candid thought: “I do everything I can to pay my people as much as I possibly can. That doesn’t necessarily mean my approach is the only right one, though.”
Recently, South Korea’s memory chip sector has been riding the massive wave of global AI computing demand. Profits are soaring, and how top players split those gains has become a hot topic across the board. SK Hynix, for instance, struck a landmark deal with its labor union last September. They scrapped the old rule capping profit-sharing bonuses at ten times base pay and shifted to a model where 10% of annual operating profit goes straight into the bonus pool.
Looking at SK Hynix’s fiscal 2025 earnings report, the numbers are staggering: annual revenue hit 97.15 trillion won, with operating profit jumping to 47.21 trillion won (a 49% margin) and net profit reaching 42.95 trillion won (44% margin). These figures completely blew past their 2024 record highs. Revenue grew by over 30 trillion won year-over-year, and operating profits doubled, setting a new annual benchmark. Crunching those numbers, each of their 30,000-plus employees stands to pocket an average bonus of around 140 million won (roughly 650,000 RMB).
Then there’s the forward-looking chatter: Macquarie Securities projects that if SK Hynix’s operating profit hits 447 trillion won by 2027, that same 10% profit-sharing model split across roughly 35,000 workers could push average bonuses to nearly 6.1 million RMB per person.
SK Hynix has responded to the speculation by noting that since this year’s and next year’s annual results aren’t locked in yet, exact bonus figures remain unpredictable. Still, they confirmed the new framework is firmly in place, tying the annual performance bonus pool directly to 10% of operating profits.
On the flip side, Samsung Electronics has also seen its storage division print money. In Q1 2026, the tech giant dropped what many are calling its “strongest financial report ever.” Revenue soared to 133.9 trillion won, up 69% year-over-year. Operating profit exploded by 756% to 57.2 trillion won—easily surpassing the entire year of 2025. The semiconductor arm alone drove 53.7 trillion won of that profit, accounting for over 93% of the total. HBM revenue tripled, and the company’s market cap briefly crossed the $1 trillion mark.
But when profits skyrocket, workers often feel left out of the celebration. Dissatisfaction over bonus distribution quickly boiled over, centering squarely on the performance pay structure. Samsung’s two main labor unions pushed hard, demanding a payout equal to 15% of each division’s operating profit, plus a 7% base wage hike. They even drew a line in the sand: meet these demands, or face an 18-day strike starting May 21.
Samsung’s initial counteroffer was pretty cautious: 10% of operating profit for bonuses, a one-time special payout, and absolutely no long-term commitment to making it a permanent policy. After a grueling round of negotiations, union members ultimately voted to accept a provisional wage agreement, narrowly dodging a massive walkout. Under that temporary deal, Samsung agreed to a 6.2% bump in average annual pay across the board. For the critical semiconductor (DS) division, they also greenlit a “Special Management Performance Bonus” funded by 10.5% of operating profits—with no hard cap on payouts.
On a broader note, the rapid rise of AI has sparked endless debates about machines stealing human jobs. Huang shut that narrative down quickly. He pointed out that the era of “practical AI” is already here, where tokens have become the new unit of profit and AI acts as a GDP “generator.” Rather than shrinking the workforce, software engineering roles are multiplying. “People keep saying AI kills jobs—that’s absolute nonsense,” Huang said. “The reality is we’re hiring more software engineers than ever.”
Back in January, Huang made a similar point about blue-collar trades. With data centers popping up everywhere to train and run AI models, plumbers, electricians, and construction workers are suddenly commanding six-figure salaries. He emphasized that this tech boom will trigger one of the largest infrastructure buildouts in history, pulling in trillions of dollars in fresh investment. “We’re watching this sector absolutely thrive, with pay rates nearly doubling. Everyone deserves a shot at a solid, livable income,” he noted.
Beyond the pay discussions, Huang took time to unpack Nvidia’s latest hardware drop: the brand-new RTX Spark PC chip, unveiled just yesterday. This silicon is aimed squarely at the consumer PC space, officially challenging Intel’s long-standing dominance in desktop and laptop processors.
According to Huang, Nvidia is aggressively expanding its AI chip lineup for PCs. But the RTX Spark isn’t just another next-gen AI PC component. It’s a completely fresh computing architecture built for the upcoming age of AI agents. The chip’s strategy hinges on three pillars: serving as an endpoint for AI agents, powering a new Windows AI PC architecture, and acting as a unified cloud-to-edge AI compute platform. He also stressed that Nvidia plans to lean heavily on off-the-shelf ARM technology rather than spending resources on custom CPU cores.
When asked about the state of the global supply chain, Huang got real about current bottlenecks. Upstream constraints in wafer foundries and component manufacturing are still causing periodic supply hiccups across Nvidia’s chip portfolio. That said, he assured that current production schedules are strategically allocated to prioritize the initial rollout of these new PC chips and AI agent processors.
On the partnership front, Nvidia is doubling down on the humanoid robotics supply chain. They just announced a collaboration with China’s Unitree Robotics to launch the H2+, a next-gen humanoid robot reference design (also known as the Isaac GR00T system). The goal? Fast-track innovation across the global humanoid robotics sector.
Investors responded enthusiastically to Nvidia’s string of announcements. By the close of trading on June 1, shares jumped 6.26% to $224.36, pushing the company’s market valuation to a massive $5.43 trillion. After-hours trading did see a slight pullback of 0.2%, but the overall sentiment remains undeniably bullish.
