AI capable of matching or even surpassing human intelligence remains a major goal pursued by the technology industry. But in the view of Nvidia CEO Jensen Huang, the milestone of artificial general intelligence (AGI) itself has become much less important.
During an earnings call held on Wednesday local time, Huang said: “For many tasks, we can say that we have already achieved AGI. I think all of these so-called milestones... at this point, don't really mean much anymore.”
OpenAI CEO Sam Altman recently told TIME magazine that OpenAI will achieve AGI internally by the end of this year. By OpenAI's definition, AGI refers to “highly autonomous systems that outperform humans at most economically valuable work.”
Huang did not provide a specific definition of AGI during the earnings call. However, he believes AI is no longer simply waiting for humans to enter prompts before executing tasks. Today, a growing number of AI programs known as “agents” can operate autonomously and achieve “recursive” improvement by repeatedly performing tasks.
Huang said: “I think there are three things that matter most to the entire industry right now. First, AI is doing work that has productive value and practical use. Second, AI is generating tokens that can produce profits. Third, if we have more computing power, we can generate more profitable tokens, allowing all services to earn more revenue. We are now at this stage.”
Huang's emphasis on productivity and profitability may also be a response to recent concerns about an AI bubble. As the technology industry invests in AI data centers on an unprecedented scale, concerns are growing about whether this boom contains a bubble. At the same time, the two major AI companies OpenAI and Anthropic still need to prove that they can achieve sustained profitability.
Critics also argue that large language models can never truly achieve AGI because they lack persistent memory, continue to struggle with understanding logic, and remain prone to “hallucinations” that produce incorrect information.
For Nvidia at least, however, the AI boom has already delivered enormous tangible profits. Nvidia generated $96.2 billion in revenue in the second quarter of fiscal 2026, up as much as 106% year over year (note: approximately RMB 648.178 billion at the current exchange rate).
Nvidia also said that without the memory supply shortage triggered by the AI boom, the company could have grown even faster. The company expects the memory shortage to last at least until early 2028 and to remain a bottleneck constraining business growth.
Nvidia CFO Colette Kress said during the call: “We are experiencing an extreme memory pricing environment. The increase in memory prices has exceeded our previous expectations, and prices will rise further next year.”
This is clearly not good news for Nvidia's graphics card business. Recently, Nvidia graphics cards have seen a round of price increases, influenced by rising memory prices and other factors.
In fact, Huang had expressed similar views before. Earlier this year, he said on Lex Fridman's podcast that, in some sense, the technology industry had already achieved AGI.
He said: “I think it's now. I think we have achieved AGI.”
Huang gave the example that today's AI programs can develop an application and make it go viral quickly, attracting billions of users, even though such applications may soon lose their popularity.
However, he also stressed that AI's capabilities still have clear limits. “But if you ask 100,000 agents like this to build an Nvidia, the probability of success is still zero,” he said.