"We Will Go Bankrupt in 30 Days"... Overcoming Countless Failures to Become a World-Class Company [Heart of Innovation, Corporate Research Center 15]

The Bitter Taste of Failure: An Essential Nutrient for Top Companies
Google Publicly Shares Its List of Failed Services
Sharing Failure Experiences at Hyundai's "Buck-up Day"

Editor's NoteThere is a growing call for a new driving force to revitalize the Korean economy, which is losing momentum. Corporate research institutes once served as the heart of our economy. However, there is now little public interest in what these corporate research centers are focusing on. IBM, the leader of the U.S. IT industry, managed to stage a comeback after its "lost decade." A key factor in this revival was the rebuilding of its research center. In the past, founders of Korea's leading conglomerates all emphasized "technology for the nation." Within the industry, there are increasing arguments that the founding spirit should be revisited and that a Korean-style research and development (R&D) DNA should be reintroduced into corporate research institutes. Asia Economy is launching a four-part series exploring research centers of global corporations, proposing the role of technology and policy directions for this era.

Jensen Huang, CEO of NVIDIA. Photo by AFP Yonhap News

Jensen Huang, CEO of NVIDIA, has repeatedly told his employees, "Our company will go bankrupt in 30 days." Although NVIDIA now stands at the center of the artificial intelligence (AI) ecosystem and is hailed as a success story, he has consistently emphasized the company's many failures along the way to keep this reality in mind.

After its founding in 1995, NVIDIA faced bankruptcy due to the failure of its first product, the NV1. However, the company managed to recover thanks to the bold and seemingly reckless development of a new product, the NV3, which received a positive response. In early 1998, during its initial collaboration with Taiwan's TSMC, a mistake in the manufacturing process led to more than half of the finished products being scrapped. NVIDIA narrowly survived this crisis by selling a partial stake to its partner. Even after these events, NVIDIA continued to confront the risk of failure head-on, and as a result, it has emerged as a top predator in the AI ecosystem.

Exhibition work from the photo exhibition "Moments of Failure Captured in Everyday Life" held in 2023 at the KAIST main campus in Daejeon. KAIST

Cho Seongho, a professor in the Department of Computer Science at KAIST, has attracted attention for his unique focus on studying corporate failure cases. Professor Cho stated, "The greatest successes can only be achieved through bold challenges, and such challenges are only possible after accumulating experiences of failure." He added, "Not only NVIDIA, but also renowned scientists and entrepreneurs such as Thomas Edison and Elon Musk, have all experienced failures behind their successes."

In June 2021, Professor Cho established the "Institute for Failure Studies" and became its director, with the aim of encouraging a culture that embraces failure. The institute seeks to foster an academic environment where research failures are not feared but seen as opportunities for new challenges, and, in the long run, to provide a safety net for members of society who have not yet succeeded so they can rise again. This year, the institute is exploring failures unique to the AI era, such as strange translations and bizarre image generation, and is organizing exhibitions and competitions to encourage AI and human failures to serve as mirrors for each other.

Google Answers, operated from 2002 to 2006, now states that it no longer accepts questions.

NVIDIA is not the only company that remembers its failures. Google, for example, maintains a website called the "Google Graveyard," where it publicly lists, year by year, products and services that failed to reach commercialization. The site includes information such as "Google Answers," a service that operated from 2002 to 2006 allowing users to submit questions and receive answers, and "Google Hire," an employment service for small businesses that was discontinued in 2020. Instead of assigning blame for failures, Google accumulates and studies failure data, turning it into an asset.

Hyundai Motor Group's Institute of Advanced Technology Development (IATD) held a weekly event called "Buck-up Day," where failure cases are shared. Hyundai Motor Company

Among Korean corporate research institutes, Hyundai Motor Group's Institute of Advanced Technology Development (IATD), which was previously its new technology research center, fostered a culture of sharing failures and trying again. At the weekly "Buck-up Day," researchers shared experiences of failed projects and discussed ways to prevent similar failures in the future. A researcher who first proposed Buck-up Day explained, "I heard that Israeli startups often have gatherings to share failure cases. Inspired by the word 'beogeopda' (meaning 'overwhelming' or 'burdensome'), we created Buck-up Day so that at least our colleagues could avoid repeating the same mistakes."