01What the Nvidia GTC Conference Announced
The Nvidia GTC Conference is Nvidia’s annual developer event where CEO Jensen Huang lays out the company’s vision for AI. This year, Huang spoke to 30,000 people in San Jose and hundreds of thousands online. He announced new chips, AI agent software, autonomous car deals, and a plan to build data centers in space.
The key message: AI is becoming as essential as electricity or the internet, and Nvidia is building the backbone for it.
In this episode, Julian Goldie breaks down the nine major announcements from the conference. He explains what they mean for regular people, not just tech insiders. From a new supercomputer chip called Vera Rubin to a robotaxi partnership with Uber, the news covers hardware, software, and real-world deployments that will affect your job and life within the next few years.
02Vera Rubin: The New AI Supercomputer Chip
Vera Rubin is the new AI supercomputer platform from Nvidia. It is named after the astronomer who discovered dark matter. The chip is built for agentic AI—AI that can plan, reason, and take actions on its own.
Previous chips like Blackwell were built to train large models. Vera Rubin is different. It is designed for AI that runs for hours or days without stopping.
The chip uses a 3-nanometer process from TSMC and has 336 billion transistors. It includes HBM4 memory with 22 terabytes per second of bandwidth. That means it can handle much more data at once without slowdowns.
Vera Rubin can deliver 700 million tokens per second. For comparison, the previous Hopper generation delivered only 2 million tokens per second. The first Vera Rubin system is already running in Microsoft’s Azure cloud.
03How to Get Started with NemoCore for AI Agents
NemoCore is the enterprise version of OpenCore, an open-source platform from developer Peter Steinberger. OpenCore lets anyone create AI agents that can take actions on your computer. It became popular quickly, but it had security risks.
NemoCore solves that by making it safe for companies to use. Here is how to get started:
- Visit the Nvidia developer portal and look for NemoCore documentation.
- Download the NemoCore software stack for your cloud provider.
- Set up access permissions to control what your AI agents can see and do.
- Test a simple agent, such as one that reads emails or updates a spreadsheet.
- Monitor agent activity with NemoCore’s built-in security logs.
- Deploy the agent at scale for your team or company.
04Using NemoCore and Vera Rubin for AI Tasks
In this episode, Julian explains how NemoCore and Vera Rubin work together. NemoCore lets you create AI agents that can communicate with outside systems. Vera Rubin powers those agents with fast token generation.
For example, you could build an agent that answers customer questions, checks inventory, and places orders—all without human help.
Nvidia also showed how they use Grok LPUs alongside Vera Rubin. Grok handles the fast responses, called decode. Vera Rubin handles the initial work, called pre-fill.
Together, they deliver 35 times more performance for real-time tasks. Julian says this means AI will become dramatically more capable in 2026. Companies can now run complex agents that were too slow or expensive before.
05Results and Comparisons from the Conference
Jensen Huang shared several big numbers at the GTC conference. Last year, Nvidia projected $500 billion in chip orders through 2027. This year, Huang doubled that forecast to $1 trillion.
He said demand is only constrained by how fast Nvidia can make chips. Customers are lined up with money ready to buy, but supply cannot keep up.
Nvidia also showed performance comparisons. Vera Rubin delivers 700 million tokens per second, compared to 2 million on older Hopper chips. That is a 350x improvement.
AWS is deploying more than 1 million Nvidia GPUs this year. Huang also announced partnerships with BYD, Hyundai, and Nissan for self-driving cars. Uber plans to deploy Nvidia-powered robotaxis across 28 cities by 2028.
06Tips, Limits, and Mistakes to Avoid
Julian Goldie gives several warnings in this episode. First, do not ignore the robotics and autonomous vehicle news. Many people focus only on the chip announcements.
But the Uber deal and car partnerships mean millions of driving jobs will change soon. Second, do not assume AI agents are only for tech companies. NemoCore makes them accessible to any business.
Julian also notes that Nvidia’s supply constraints are real. You cannot just order a Vera Rubin system and get it tomorrow. Plan ahead if your business depends on Nvidia hardware.
Finally, avoid the mistake of waiting. Julian says the time to learn about these tools is now, not after your competitors have already deployed them.
07Who the Nvidia GTC Conference Matters For
This conference matters for every company and every worker. Jensen Huang said every company needs a strategy for OpenCore and AI agents, just like they needed a strategy for the internet. If you run a business, you need to understand how NemoCore can help you automate tasks.
If you work in driving, logistics, or transportation, the robotaxi announcements affect your industry directly.
Julian explains that the gap between people who understand these tools and those who do not is widening every month. The people who learn now will build things no one has imagined yet. The people who ignore it will be caught off guard.
This is not a future trend—it is happening right now with systems already running in the cloud.
08What Comes After Vera Rubin: Feynman in 2028
Nvidia already has plans beyond Vera Rubin. The next architecture is called Feynman, named after the physicist Richard Feynman. It is on track for 2028.
Feynman will include a new GPU called Rosa, named after Rosalind Franklin, who discovered the structure of DNA. The CPU is designed to move data efficiently across the full AI stack.
Julian says this shows Nvidia’s long-term thinking. They are not just releasing one chip at a time. They are building a complete ecosystem that gets better every few years.
Each new generation makes AI faster, cheaper, and more capable. The companies that build on Nvidia’s platform today will benefit from these upgrades automatically.
09Why This Conference Changes the AI Landscape
Julian says this GTC conference felt different from previous ones. The hardware, software, partnerships, and use cases are all coming together at the same time. AI is no longer a single breakthrough app.
It is becoming central infrastructure, like electricity or the internet. Nvidia is building the power grid for AI.
The key question for you is what you will do with that power. People who figured out the internet early built businesses that changed the world. People who figure out AI agents early will do the same.
The conference showed that the tools are ready, the demand is real, and the time to start is now.
