The Next OpenAI: Is It ‘Hopeless’ For Nations With Limited Funds To Run In The AI Race?

OpenAI took the world by storm when it unveiled ChatGPT towards the end of 2022. It has since dominated the world of artificial intelligence, creating an entirely new wave of startups centered around AI.

OpenAI’s impact has not been limited to just AI – from startups to the biggest companies in the world like Apple and Microsoft to the stock market and more, the company’s technology has left a mark on everything.

However, this wouldn’t have been possible for Sam Altman and his team without the significant amount of resources and talent that were poured into creating what we now know as ChatGPT.

AI Is Expensive

While everyone can train their large language model, doing so is an expensive affair. From expensive GPUs like Nvidia’s H100 – which costs between $30,000 to $40,000 each – to power and more, it takes millions of dollars to just scratch the surface.

The next major barrier is talent: salaries of those working in AI are sky-high.

The median salary in the U.S. for a person in the AI industry stands at $326,100, according to data from AI Paygrades, while a software engineer earns between $9,590 to $25,000 in India.

For context, India's per capita GDP is $2,730, far behind the U.S. at $85,370 and China at $13,140. This resource gap significantly affects the ability of developing countries to create leading AI companies.

The problem of limited resources is obvious to Altman, too. When asked if countries like India can develop a ChatGPT competitor, the OpenAI co-founder was blunt. “The way this works is, we're going to tell you, it's totally hopeless to compete with us on training foundation models,” he said at an event in India in 2023.

This begs the question: will we ever see the next OpenAI emerge from the developing world and countries with limited funds?

Here’s What Nvidia Is Doing To Make AI More Affordable

In a conversation with Benzinga, Nvidia's Managing Director in South Asia, Vishal Dhupar, likened the advent of AI to an industrial revolution, highlighting its transformative potential. 

He explained that just as the industrial revolution was driven by steam power, today's revolution is driven by data and computing power. "AI is upon us, and the way I want you to think about it is an industrial revolution," he said.

According to Dhupar, Nvidia is actively working to reduce the total cost of ownership for AI technologies, aiming for what he termed as “literally zero intelligence cost.”

He spoke about the advancements in computing architecture, noting that the speed of computing is doubling every six months, scaling tenfold annually, and achieving a millionfold increase over a decade.

"Computing is now moving twice as fast every six months, 10 times every year, one million times every 10 years," he pointed out.

The Nvidia executive then spoke about the need for new infrastructure, transitioning from traditional data centers to "AI factories." This transformation is crucial for reducing costs and making AI more accessible. 

“At the end of the day, everybody understands that the total cost of ownership has to be the lowest and true to the new change that is taking place,” he remarked.

Catching Up Is The Difficult Part

OpenAI’s chief scientist Ilya Sutskever left the company after a failed coup against co-founder Altman. Months later, he announced a new AI startup, Safe Superintelligence, with no intent to turn a profit. Without a clear product roadmap, attracting investors and talent will be challenging.

Despite his announcement, no progress has been revealed. Even with funding, Sutskever faces nearly a decade of catching up with OpenAI and other established AI startups.

This situation parallels the challenges faced by developing nations like India, which are just beginning their AI journeys. In contrast, countries like the U.S. and China have advanced rapidly due to significant funding and investments in R&D.

However, not every expert is pessimistic about the chances of an OpenAI-like startup emerging again. Barring the issue of limited resources, there are chances of the “next OpenAI” emerging from anywhere – academia or the corporate world – according to Chris Shaw, CMO at NTT Research, “because it’s so broad and open for people to try things.”

For now, creating the next OpenAI might be out of reach of many, if not most, nations hamstrung by limited resources. However, with Nvidia working to bring down the cost of computing and more advancements in the field, experts think it’s a matter of when not if.

That's all for this week! If you found these updates useful, you'll like more from this newsletter. Get deeper dives, hot takes, and all the latest tech news delivered straight to your inbox.