These agents become less like a PC and more like R2-D2
At Computex in Taiwan, Nvidia unveiled RTX Spark — a chip co-built with Microsoft that carries the weight of a bold proposition: that personal computing, largely unchanged in its essential character for forty years, is ready to be remade around artificial intelligence that lives on the device itself rather than in distant servers. The announcement positions AI agents not as novelties but as the next layer of the human-machine relationship, capable of anticipating needs rather than merely responding to commands. Yet between the ambition of the engineers and the habits of ordinary people lies a familiar distance — one measured not in transistors, but in trust.
- Nvidia's Jensen Huang declared the PC due for its first reinvention in four decades, staking the claim on a single chip packing a Blackwell GPU, 20-core CPU, and 128GB of unified memory capable of one petaflop of AI performance.
- The RTX Spark is designed to run AI agents — from companies like Anthropic and Nous Research — natively inside Windows, bypassing the cloud and the technical friction that has kept local AI out of reach for most users.
- A September rollout through Dell, HP, Lenovo, ASUS, and others signals serious industry commitment, but undisclosed pricing and premium specifications suggest these machines will arrive with a cost barrier attached.
- A survey of over 2,000 Australians reveals the gap: only one in ten currently use AI agents on dedicated devices, fewer than a third are comfortable with AI managing more of their lives, and battery life still outranks AI features as a purchase priority.
- Analysts point to privacy and security as the stubborn core of consumer hesitation — and while local processing may ease some fears, the harder challenge of winning genuine trust remains unresolved as the chip's launch date approaches.
Jensen Huang took the stage at Computex in Taiwan and announced what Nvidia is calling the reinvention of the personal computer: a new chip named RTX Spark, built in partnership with Microsoft, designed to bring AI agents directly into Windows machines without relying on cloud servers. The chip houses a Blackwell graphics processor delivering one petaflop of AI performance, a 20-core CPU, 128 gigabytes of unified memory, and 70 billion transistors in total. Huang positioned it as a direct challenge to Apple, Intel, and AMD in the race to embed AI into consumer hardware.
What sets RTX Spark apart is its ambition to make AI agents genuinely accessible. Today, running agents locally on a PC requires navigating open-source tools and managing real technical complexity. Nvidia's vision is for agents from Anthropic, Nous Research, and others to run natively, integrated into the Windows taskbar, capable of completing tasks autonomously on a user's behalf — something Huang compared, in spirit, to the helpful droids of Star Wars.
The rollout begins in September, with ASUS, Dell, HP, Lenovo, Microsoft Surface, MSI, Acer, and GIGABYTE all preparing RTX Spark devices. Pricing has not been disclosed, though the chip's specifications and the current cost pressures around AI memory suggest premium price points. Microsoft CEO Satya Nadella framed the collaboration as delivering "unmetered intelligence to every home and every desk."
The distance between that vision and current reality is significant. A Telsyte survey of more than 2,000 Australians found only one in ten using AI agents on dedicated devices, with Apple silicon capturing over 40 percent of that small group. Fewer than a third of Australians over 16 said they were comfortable with AI taking on more of their daily lives, and battery life still ranked above AI features when people described what they wanted in their next computer.
Telsyte analyst Alvin Lee identified trust, security, and privacy as the primary obstacles to broader adoption, noting that local processing might offer some reassurance — though the market remains unconvinced. The RTX Spark is a genuine technical achievement and a serious industry bet. Whether it changes how people actually live with their computers depends on something no specification sheet can guarantee: the slow, difficult work of earning trust.
Nvidia's Jensen Huang stood before the crowd at Computex in Taiwan on Monday and made a bold claim: the PC, as we know it, is about to be reinvented for the first time in four decades. The vehicle for this transformation is a new chip called RTX Spark, developed in partnership with Microsoft, designed to bring artificial intelligence agents directly into Windows machines without requiring a connection to distant cloud servers.
The RTX Spark is built on formidable specifications. It houses an Nvidia Blackwell graphics processor capable of delivering one petaflop of AI performance—a measure of raw computational power—alongside a 20-core central processor and 128 gigabytes of unified memory. The entire assembly contains 70 billion transistors. Huang called it "the most amazing chip the world has ever built," and positioned it as a direct challenge to competitors like Apple, Intel, and AMD who have been racing to embed AI capabilities into consumer hardware.
What distinguishes RTX Spark from existing approaches is its ambition to make AI agents practical and accessible on personal computers. Today, AI agents can technically run locally on some Windows devices, but doing so typically demands technical expertise—users must navigate open-source platforms like OpenClaw and manage installation and safety concerns themselves. Nvidia's vision is different: agents from companies like Anthropic and Nous Research would run natively on RTX Spark machines, integrated directly into the Windows taskbar, capable of autonomously completing tasks on behalf of users. Over time, Huang suggested, these agents might evolve into something less like a traditional computer and more like the helpful droids from Star Wars—always present, always working, anticipating needs rather than waiting for commands.
The rollout begins in September, when manufacturers including ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI will begin shipping laptops and desktops equipped with RTX Spark. Acer and GIGABYTE will follow. Pricing has not been announced, but the chip's advanced specifications and the current surge in memory costs driven by AI demand suggest these machines will command premium prices. Microsoft CEO Satya Nadella framed the partnership as a step toward delivering "unmetered intelligence to every home and every desk," language that signals the companies see this as foundational infrastructure for the next era of computing.
Yet there is a gap between vision and reality. A survey of more than 2,000 Australians released this week by analyst firm Telsyte found that only one in ten are currently running AI agents on dedicated devices. Among those who do, more than 40 percent chose Apple silicon—a share far exceeding Mac's overall market penetration. The barriers to broader adoption are clear and persistent. Less than a third of Australians over 16 expressed comfort with AI managing more of their daily lives. Privacy and security concerns dominate the hesitation; people worry about what happens when they grant an autonomous agent deeper access to their personal data and digital lives.
Alvin Lee, a senior analyst at Telsyte, noted that trust, security, and privacy remain "the primary barriers to full AI autonomy." He suggested, though, that local processing—computation happening on the device itself rather than in remote data centers—might ease some of those concerns by giving users a sense of control. Yet even this argument faces skepticism in the market. When Telsyte asked Australians what matters most in their next computer purchase, battery life and raw performance still dominated. Only one in three said advanced AI features were essential. Around 40 percent expressed interest in hypothetical devices from AI companies like OpenAI, Anthropic, and xAI, but those products do not yet exist.
The RTX Spark launch represents a genuine technical achievement and a significant bet by two of the world's largest technology companies. Whether it reshapes how people actually use computers depends on whether Nvidia and Microsoft can convince consumers that local AI agents solve problems they recognize having, and that the privacy trade-offs are worth making. The chip arrives in September. The harder work—winning trust—has only just begun.
Notable Quotes
Microsoft and Nvidia are going to reinvent the PC using Nvidia's chip technology and a Windows platform for agents— Jensen Huang, Nvidia CEO
Our goal is to deliver unmetered intelligence to every home and every desk with Windows. RTX Spark marks a real breakthrough towards that vision— Satya Nadella, Microsoft CEO
The Hearth Conversation Another angle on the story
Why does Nvidia think the PC needs reinventing now, after forty years of incremental change?
Because AI has shifted what a computer can do. For decades, a PC was a tool you controlled—you clicked, typed, navigated. Now there's software that can act on your behalf, autonomously. That changes everything about the relationship between person and machine.
But AI agents already exist. Why does RTX Spark matter if people can already run them locally?
Accessibility. Right now, running an agent locally requires technical skill. You need to know about open-source platforms, installation, security. RTX Spark puts agents in the Windows taskbar—the same place you've always launched applications. It removes the friction.
The survey shows Australians don't trust AI with their data. How does on-device processing solve that?
It doesn't solve it entirely, but it shifts the equation. If the agent runs on your machine instead of sending everything to a cloud server, you're not handing your data to a distant company. You retain some control. That matters psychologically, even if the technical security questions remain.
Is this actually a reinvention, or just a more powerful chip?
It's both. The hardware is more powerful, yes. But the real shift is architectural—the idea that your personal computer becomes a hub for autonomous agents working on your behalf, rather than a tool you manually operate. That's a different mental model.
Why would someone want an AI agent managing their life?
Efficiency, mostly. Imagine an agent that schedules your calendar, filters your email, handles routine tasks. But also—and this is what Huang emphasized—it frees you from the constant small decisions. Over time, these agents could become genuinely useful in ways we haven't yet imagined.
The survey says only one in ten Australians use AI agents. Isn't that a sign the market doesn't want this?
It's a sign the market hasn't yet seen a compelling reason to adopt it. RTX Spark is betting that when agents are easy to use and run locally, adoption will follow. That's an assumption, not a certainty.