The field of artificial intelligence (AI) is all about the 95% number. Similarly, an influential and perhaps overblown research study from MIT in August 2025 found that 95% of generative AI pilots fail. luckDiane Brady spoke with PwC Global Chairman Mohamed Khande in Davos, Switzerland, about six months later, and the numbers were stubbornly high: 56% of CEOs surveyed said they were getting “nothing” from their AI efforts.
According to Nvidia CEO Jensen Huang, the solution is peace, love, understanding, and good parenting skills. The $4 trillion market cap man arrived at the Cisco AI Summit with a message that sounded more like a fusion of 1960s counterculture and modern parenting than Wall Street rigor. “Let a thousand flowers bloom.”
In a conversation with Cisco CEO Chuck Robbins, Huang discussed the tensions faced by business leaders who feel the pressure to implement AI but fear that they won’t see immediate, quantifiable results. When Robbins asked about the first steps companies should take, Huang denied immediately fixating on spreadsheets.
“We get questions like ROI,” Huang says. “I won’t go there.”
Instead, he advocated a philosophy of abundance and messy experimentation, explicitly likening corporate innovation to raising children. He argued that asking engineers to prove financial success before trying new AI tools is as stifling as asking a child to justify a hobby with a business plan.
“I want the same thing for my company that I want for my children: to explore life,” Huang explained. If children say they want to try something, they should say yes, he added. At home, we don’t ask questions like, “What is the return on investment here?” How does this lead to financial success? How can you prove to me that it’s worth it? “I never do that at home. But I do it at work.”
Innovation requires therapy, not control
Huang acknowledged that this approach requires executives to relinquish some control, which they may feel uncomfortable with, but argued that creativity and innovation bring value to the approach. “We have a lot of different AI projects that are out of our control, which is great,” he said, noting that innovation doesn’t always happen when you have control. “If you want to be in control, first of all, you have to go to therapy. But secondly, it’s an illusion. You can’t control it. If you want your company to be successful, you can’t control it. If you want your company to be successful, you can’t control it.”
Mr. Hwang argued that to be successful, leaders must seek to: influence Rather than controlling them, their company. The logic behind “making a thousand flowers bloom” is risk management through diversification. While this approach would “make the garden cluttered,” it would avoid the mistake of putting resources into the early stages of a technology shift when the “winning” tools are not yet clear, or “putting all the wood behind one arrow,” he said.
raise the hood
While advocating a relaxed approach to ROI, Huang was adamant about the need for “tactile understanding.” He urged leaders not to rely solely on cloud rentals or finished products.
Computers are everywhere these days, he said, but even if you build your own, you’ll understand better, in the same way that a good car owner keeps a close eye on the engine instead of taking an Uber. “Open the hood, change the oil, understand all the components,” he said. “Create something. You might discover that you’re actually insanely good at it. You might realize that you need that skill.”
He emphasized that because AI technology is essential to the future, companies need to build some infrastructure on-premises to truly understand how the “components” work. This concerns data privacy and what Huang calls the most valuable intellectual property: questions. “My most valuable intellectual property is not my answers…they are my questions,” Huang said, noting that while answers are a commodity, smart questions are invaluable.
luck I recently visited KPMG’s Lake House in Orlando, Florida. There, the company was rolling out an AI training framework, first to interns and then to the entire company. “Think, Be Fast, Check” is how employees are trained to take advantage of AI, and the first and last points emphasized that they cannot be taken for granted.
From explicit to implicit
The urgency of the experiment stems from a fundamental “reinvention of computing,” Huang said. The industry is moving away from “explicit programming,” where users write specific lines of code, to “implicit programming,” where users state their intent and AI finds a solution.
In this new world, Huang points out, “typing is a commodity.” The real value lies in the expertise needed to guide the AI. “You tell the computer what you want, and the computer works to figure out how to solve the problem.”
Mr. Huang concluded by turning the common ethical narrative of “people in circles” on its head. He said the goal should be “AI in the loop.” By integrating AI into every process, companies can understand the “life experiences” of their employees and turn everyday tasks into lasting corporate intellectual property. In other words, they will bloom a thousand flowers, but only if they have the right curiosity, the right questions, and the right support from above to think freely.
Regarding this story, luck Journalists used generative AI as an investigative tool. Editors verified the accuracy of the information before publication.
