This is going to transform the traditional app-centric PC to a real useful Agentic AI personal computer
At Taipei's Computex gathering, Nvidia CEO Jensen Huang unveiled the RTX Spark chip — a device designed to bring artificial intelligence out of distant data centers and into the personal computers people use every day. Built over three years in partnership with Microsoft and MediaTek, the chip signals a broader philosophical shift: from AI as a cloud-dependent service one queries, to AI as an autonomous agent that thinks and acts locally, on your machine. Analysts are calling it a threshold moment comparable to the iPhone or ChatGPT, and the markets moved accordingly — rivals fell, partners rose, and the industry began recalibrating its sense of what a personal computer is for.
- The announcement carries the weight of a paradigm shift — Nvidia is not merely releasing a chip but declaring that the app-centric PC, built for human-initiated actions, is giving way to a machine designed for autonomous AI agents that act without waiting to be asked.
- Markets responded with immediate conviction: AMD and Intel dropped nearly 4 percent, Qualcomm fell almost 7 percent, while Microsoft — Nvidia's partner — surged 3.1 percent, signaling investor belief that competitive lines in personal computing are being redrawn.
- The companion Vera CPU, already adopted by OpenAI, Anthropic, and SpaceX, targets a $200 billion market and represents Nvidia's strategic pivot from data center dominance toward the everyday devices in people's homes and offices.
- Qualcomm's CEO independently framed 2026 as 'the year of agents,' underscoring that the urgency around local, always-on AI is industry-wide — edge computing is becoming not a preference but an architectural necessity.
- Huang used the Computex stage to pledge roughly $150 billion annually in Taiwan investment, weaving together technological ambition and geopolitical reality in a region whose chip supply chain has become a matter of global consequence.
Jensen Huang took the stage at Taipei Music Hall in his signature black leather jacket to announce something he framed as the end of the old personal computer. The occasion was Computex, the industry's largest gathering, and the announcement was the RTX Spark — a chip built to run artificial intelligence directly on laptops and desktops, without routing tasks through distant cloud servers. Three years in the making, developed alongside Microsoft and Taiwan's MediaTek, it represents Nvidia's vision of a computer that doesn't wait to be asked but acts on your behalf.
Analysts at Counterpoint Research called the moment historic. Co-founder Neil Shah compared it to the arrival of the iPhone or ChatGPT — one of those rare threshold events when an entire category of technology shifts its fundamental purpose. The RTX Spark, he argued, would transform the traditional PC into what he called a real agentic AI personal computer, eventually present in every home.
Huang devoted much of his keynote to two products: the RTX Spark and the Vera CPU, a processor purpose-built for AI agents. The Vera has already drawn early adopters — OpenAI, Anthropic, and SpaceX — and Huang has described it as a gateway to a $200 billion market and Nvidia's next major growth driver. The implication is significant: the company that built its dominance on data centers is now betting on the machines people carry and keep on their desks.
The market read the room quickly. Shares of AMD, Intel, and Qualcomm fell sharply in pre-market trading; Microsoft climbed 3.1 percent. Qualcomm's own CEO, Cristiano Amon, had separately framed 2026 as 'the year of agents,' describing a world where AI moves past answering questions toward acting independently and continuously — a shift that demands entirely different hardware architecture than today's user-initiated devices.
Huang, who was born in southern Taiwan, used the occasion to announce plans to invest around $150 billion annually in the island, calling it the epicenter of the AI revolution. The speech came weeks after he had joined President Trump on a corporate visit to Beijing, a reminder that Taiwan's place in the global chip supply chain carries consequences far beyond technology. When asked whether AI would displace software engineers, Huang was blunt: 'Complete nonsense,' he said, arguing the technology would drive hiring rather than eliminate it. Whether that confidence proves warranted depends on how quickly the RTX Spark and Vera deliver on their considerable promise.
Jensen Huang stood on stage at Taipei Music Hall in his signature black leather jacket, ready to announce that the personal computer was about to change in a fundamental way. The Nvidia CEO had come to Taiwan for Computex, the industry's largest gathering, to unveil the RTX Spark—a chip that would let laptops and desktops run artificial intelligence directly, without constantly reaching back to distant servers in the cloud. It was, he suggested, the beginning of the end for the old PC.
The RTX Spark represents three years of work between Nvidia and Microsoft, a collaboration aimed at what Huang called "reinventing the PC" for an era when AI is no longer a tool you query but an agent that acts on your behalf. The chip was developed with help from Taiwan's MediaTek and is designed to do something previous generations of consumer hardware could not: run AI agents locally, meaning the thinking happens on your machine, not somewhere else. Analysts at Counterpoint Research saw the moment as historic. Neil Shah, the firm's co-founder, compared it to the arrival of the iPhone or ChatGPT—a threshold moment when an entire category of computing shifts its purpose. "This is going to transform the traditional app-centric PC to a real useful Agentic AI personal computer," Shah said, "which will eventually be in every home in coming years."
Huang spent much of his keynote on two products: the RTX Spark and Nvidia's Vera CPU, a processor built specifically for AI agents. The Vera has already attracted early adopters—OpenAI, Anthropic, and SpaceX are among the first to use it. During an earnings call in May, Huang had described the Vera as a gateway to a market worth $200 billion, and he called it "our new major growth driver." The shift signals where Nvidia sees the future: not in the data centers that have made the company dominant, but in the machines people use every day.
The market reacted swiftly. Shares of AMD and Intel, Nvidia's traditional rivals in the PC chip space, fell nearly 4 percent in pre-market trading. Qualcomm dropped almost 7 percent. Apple slipped 0.6 percent. Microsoft, Nvidia's partner in this effort, jumped 3.1 percent. The message was clear: investors believed this announcement would reshape the competitive landscape.
Huang was not alone in seeing 2026 as a turning point. Cristiano Amon, CEO of Qualcomm, had made similar remarks ahead of Computex, framing the year as "the year of agents." The shift he described is profound: AI is moving past being a tool that answers questions when you ask it toward becoming something that acts independently, that runs continuously, that makes decisions without waiting for human instruction. This requires a different kind of computer. Today's devices were built for user-initiated actions—you open an app, you type a command, something happens. An always-on autonomous agent needs different architecture entirely. "All of these devices today, they have been built for actions initiated by the user, not by the agents," Amon said. Local computing, what the industry calls edge AI, becomes not optional but necessary.
Huang, who was born in Taiwan's southern city of Tainan, used his time on stage to emphasize his company's commitment to the island. He announced plans to invest around $150 billion a year in Taiwan, calling it "the epicenter of the AI revolution." The timing was notable: the speech came roughly two weeks after Huang had accompanied President Donald Trump on a visit to Beijing, part of a corporate delegation meeting with Chinese leader Xi Jinping. Taiwan's role in the global chip supply chain has become a matter of geopolitical consequence, and Huang's words were as much about that reality as about technology.
When asked about concerns that AI would displace software engineers, Huang was dismissive. "Complete nonsense," he said. The technology would make workers more productive, he argued, which would drive hiring rather than eliminate jobs. "The number of engineers, software engineers, is actually increasing. People talk about AI reducing jobs—complete nonsense. It's causing more software engineers to be hired." Whether that optimism holds depends on what happens next: whether the RTX Spark and Vera actually deliver on their promise, whether the market adopts them as broadly as Nvidia hopes, and whether the shift to local, autonomous AI agents unfolds as quickly as the company's leadership believes it will. Computex runs through June 5, and the industry will be watching closely.
Citas Notables
This is going to transform the traditional app-centric PC to a real useful Agentic AI personal computer which will eventually be in every home in coming years as private edge AI agents become pivotal.— Neil Shah, Counterpoint Research co-founder
All of these devices today, they have been built for actions initiated by the user, not by the agents.— Cristiano Amon, Qualcomm CEO
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter that AI runs on your laptop instead of in the cloud? Isn't it just moving the same computation to a different place?
It's more than that. If your AI agent has to ask a distant server for permission every time it wants to act, it's slow and it's dependent. But if the thinking happens locally, the agent can act in real time, without latency, without needing a connection. It becomes something that can actually run your life—manage your calendar, filter your email, make decisions—without constantly phoning home.
And why is 2026 the year this becomes possible? What changed?
The chips got powerful enough. For years, consumer hardware was too weak to run meaningful AI models. But now—with the RTX Spark, with better processors—you can fit real intelligence into a device you hold in your hands. That's the threshold moment.
The stock market reacted immediately. AMD and Intel dropped. Why are they vulnerable?
Because they've been building chips for a world that's ending. They optimized for traditional computing tasks. Nvidia is saying the future is different—it's about AI agents, about edge computing, about a completely different architecture. If Nvidia is right, the old playbook doesn't work anymore.
Huang mentioned a $200 billion market. Is that real or aspirational?
It's a projection, but it's grounded. If every PC, every device becomes an AI agent platform, the market for the chips that power that is enormous. Whether Nvidia actually captures it depends on execution and adoption. But the size of the opportunity is real.
What does it mean that Huang visited Beijing with Trump just before this announcement?
It means Taiwan's role in the global chip supply chain is now inseparable from geopolitics. Huang is signaling that Nvidia is committed to Taiwan, that the island is central to the future. But he's also navigating a world where the U.S. and China are competing for technological dominance. Every move carries weight.