NVIDIA Conference: Jensen Huang Says AI Is Reinventing Computing—Why Enterprises Must Move Fast

NVIDIA NASDAQ: NVDA CEO Jensen Huang used a late-evening conversation with a partner to outline how he believes artificial intelligence is reshaping the computing industry and why enterprises should move quickly to adopt it, even if near-term return on investment is difficult to quantify.

Reinventing the computing stack

Huang said the industry is “reinventing computing for the first time in 60 years,” describing a shift from explicit programming—where developers write code and define variables—to “implicit programming,” where users state intent and the system determines how to achieve it.

He argued that the change extends well beyond processors to the broader computing stack, including storage, networking, and security. In that context, Huang referenced an ongoing partnership aimed at combining NVIDIA’s AI networking technology with Cisco’s Nexus control plane to deliver AI performance with Cisco’s “controllability, and security, and the manageability.” He added that the companies intend to take a similar approach to security.

Huang also said enterprise AI adoption lagged in part because early systems were more “interesting and curious” than truly useful. He described the next phase as “agentic AI,” where systems can use tools, perform research, incorporate memory, and rely on retrieval-augmented generation to ground responses in facts.

Advice for enterprises: experiment first, curate later

Asked what steps enterprises should take to get ready, Huang said he would avoid focusing too early on spreadsheet-style ROI calculations, arguing it is often hard to quantify the value of new technology at the beginning. Instead, he recommended identifying a company’s most impactful work and applying AI there, rather than starting with “peripheral stuff.”

He repeatedly emphasized the need to encourage broad experimentation, describing his own organization as having “a thousand flowers bloom” across many projects. Huang said innovation is not always controlled and suggested companies should let people try tools “safely,” then later “curate the garden” once they have learned what works. He noted that his company uses a range of AI offerings, naming Anthropic, Codex, and Gemini as examples.

• Start with the core: focus AI on the most essential work your company does.
• Encourage experimentation: allow teams to explore multiple tools and approaches early.
• Curate later: standardize platforms only after learning what fits, rather than committing too soon.

As an example of concentrating on core work, Huang pointed to efforts to “revolutionize the tools” used for chip and system design. He cited partnerships involving Synopsys, Cadence, Siemens, and Dassault, saying he wants those tool providers to have whatever technology they need to modernize workflows.

“Abundance of intelligence” and a new way to approach problems

Huang framed AI as dramatically reducing the “cost of intelligence,” enabling tasks that once took a year to be done in a day, an hour, or even in real time. He compared the pace of progress to Moore’s Law and said AI advancement implies a new “sensibility” where teams should assume capabilities that feel effectively unlimited—such as processing extremely large graphs in full rather than in pieces.

He argued that companies not adopting that mindset risk being outpaced by competitors or new entrants that treat speed, scale, and iteration as near constraints-free.

Physical AI, tool use, and an expanded market opportunity

On “physical AI,” Huang rejected the idea that AI will simply replace software tools, calling that view “illogical.” He said advanced AI systems—whether robots or digital agents—should use tools rather than reinvent them, comparing enterprise software to practical implements like screwdrivers and hammers. He cited “tool use” as a key direction for AI progress.

He also said the next generation of physical AI will need deeper understanding of the real world, including causality. As an illustration, he described a child’s intuitive understanding of domino effects and argued that such physical reasoning is not something large language models inherently possess today.

Huang said “digital labor,” such as self-driving cars functioning as “digital chauffeurs,” could be valued far beyond the underlying hardware over time. He also argued that AI exposes the technology sector to a market far larger than traditional IT, contrasting an IT industry of roughly a trillion dollars with a global economy of about a hundred trillion dollars.

He suggested many companies will seek to become “technology first,” emphasizing the advantage of working with “electrons, not atoms,” and said that AI makes it easier for organizations to translate domain expertise into software outcomes because they can express intent directly rather than relying solely on traditional coding skills.

Build vs. rent, on-prem needs, and “AI in the loop”

Huang urged companies to develop hands-on understanding of AI infrastructure rather than relying entirely on renting cloud services. He compared it to building a computer or learning to “lift the hood” and “change the oil” in an automotive context, arguing that the technology is important enough that organizations need tactile familiarity with how it works.

He also made the case for a hybrid approach—renting some resources while owning others—particularly for sovereignty and proprietary information. Huang said NVIDIA builds AI systems locally because it is not comfortable putting all of the company’s conversations in the cloud, arguing that a company’s “most valuable IP” may be its questions rather than its answers.

Finally, Huang challenged the idea that AI systems should always keep “humans in the loop,” calling that framing backward. Instead, he said companies should put “AI in the loop” so organizations can capture and compound knowledge over time. He predicted employees will have “lots of AIs” embedded in daily workflows and that those systems will become part of a company’s intellectual property.

About NVIDIA NASDAQ: NVDA

NVIDIA Corporation, founded in 1993 and headquartered in Santa Clara, California, is a global technology company that designs and develops graphics processing units (GPUs) and system-on-chip (SoC) technologies. Co-founded by Jensen Huang, who serves as president and chief executive officer, along with Chris Malachowsky and Curtis Priem, NVIDIA has grown from a graphics-focused chipmaker into a broad provider of accelerated computing hardware and software for multiple industries.

The company's product portfolio spans discrete GPUs for gaming and professional visualization (marketed under the GeForce and NVIDIA RTX lines), high-performance data center accelerators used for AI training and inference (including widely adopted platforms such as the A100 and H100 series), and Tegra SoCs for automotive and edge applications.

This instant news alert was generated by narrative science technology and financial data from MarketBeat in order to provide readers with the fastest reporting and unbiased coverage. Please send any questions or comments about this story to contact@marketbeat.com.