5 Key Takeaways from NVIDIA GTC DC 2025
In this blog
- Takeaway 1: Agentic AI is here — and evolving rapidly
- Takeaway 2: Quantum computing is closer than expected
- Takeaway 3: DOE is powering infrastructure modernization
- Takeaway 4: AI RAN showed real-world impact for communications and defense
- Takeaway 5: Adoption requires training, alignment and collaboration
- Bringing it all together
- Download
NVIDIA's GTC conference in Washington, D.C., offered an up-close look at how the next wave of AI-driven innovation stands to impact organizations across the public and private sectors. From agentic AI to advances in high-performance and even quantum computing, the event made clear that AI's future is accelerating faster than many expected.
Here are our top takeaways.
Takeaway 1: Agentic AI is here — and evolving rapidly
GTC highlighted a shift from static AI models to agentic systems capable of reasoning, executing tasks and synthesizing insights across diverse datasets. These agents are purpose-built, infused with domain expertise and operate in near real-time, integrating new data and adapting to workflows dynamically.
A prime example of agentic AI that was on display is our Multi-Intelligence Toolkit, an application built within the WWT AI Proving Ground. The application leverages industry-leading models within an agentic AI framework to execute a multistep reasoning workflow.
Additionally, we showcased Kendra, a digital human developed via WWT's AI Proving Ground, which helps users access information about their Department of Veterans Affairs (VA) benefits. Explicitly trained on benefit data, Kendra can operate across channels — from virtual interactions to phone support to responding to questions in person at VA offices — delivering context-aware assistance while learning from each interaction.
"AI building AI" was a key theme at GTC and the conference highlighted how nontechnical staff are increasingly able to build and deploy AI agents using natural language tools. This democratization of AI development empowers teams to create agents that have a direct impact on mission and business outcomes without having to involve AI specialists.
Takeaway 2: Quantum computing is closer than expected
A surprise at the conference was how much quantum computing was spotlighted — not just as a research topic, but as a near-term priority. Multiple quantum vendors showcased solutions on the exhibit floor, including a full-scale replica of a quantum computer.
NVIDIA emphasized that its NVIDIA NVQLink™ technology could help push quantum into production faster than the industry anticipates. The company also announced 27 new quantum partnerships and highlighted Department of Energy (DoE) labs that are advancing quantum initiatives. For federal agencies exploring post-quantum cryptography and other emerging security areas, the message was clear: progress is accelerating.
Takeaway 3: DOE is powering infrastructure modernization
The Department of Energy (DOE) featured prominently across discussions of both quantum and high-performance computing. NVIDIA spotlighted seven new supercomputers acquired by the DOE to drive U.S. innovation and underscored the ongoing modernization of data infrastructure to support AI workloads.
As capital investments continue to optimize around reduced AI processing costs — falling from $20 to 40 cents per million tokens — NVIDIA positioned its new NVIDIA Blackwell platform as the most cost-effective solution on the market.
Takeaway 4: AI RAN showed real-world impact for communications and defense
A standout demonstration came from Dr. May Casterline who introduced NV Radar, a toolkit that allows organizations to embed AI into radio frequency missions. NVIDIA's $1 billion investment in Nokia and partnerships with Dell and T-Mobile signal how AI RAN will play a major role in advancing 6G innovation.
The applications of AI RAN in 6G mirror WWT's federal research efforts and communications projects, including those supporting defense and intelligence operations.
Takeaway 5: Adoption requires training, alignment and collaboration
A recurring theme at GTC was that adoption hinges on organizational readiness, not just infrastructure. Effective adoption requires deliberate training programs, role-specific guidance and continuous upskilling.
Adoption also depends on bridging domain expertise with AI capabilities. Whether it's healthcare, government or energy, success comes from connecting subject matter experts with technical teams to co-develop solutions. Organizations that invest in these connections will see faster adoption, higher ROI, and lower failure rates compared with isolated or purely top-down AI initiatives.
Bringing it all together
GTC DC 2025 reinforced that AI innovation isn't slowing — it's expanding into new domains like quantum computing, 6G and large-scale infrastructure modernization. The event underscored the importance of continued collaboration with NVIDIA and the broader ecosystem to accelerate federal missions and enterprise outcomes.
The bottom line? AI's evolution is here, and it's reshaping every sector.