If human-level AI doesn’t happen soon, Big Tech faces a financial bill.
To make up the gap between investments and profits in the industry, artificial intelligence companies may need to make good on their promise to develop artificial general intelligence (AGI) in the near future.
Unfortunately, there is still no scientific evidence that AGI – human-level or more rational machines – is even possible.
Growth market
According to analysts, the current AI market is largely expected. OpenAI is one of the few most profitable creative AI enterprises, and the gap between its revenue (about $3.4 billion according to data) and its next closest competitors is huge.
This equates to a capital shortfall or negative flow of around $600 billion, according to Sequoia Capital's latest analysis.
Sequoia mentions that the figures are based on Nvidia's GPU usage estimates. With this in mind, the above figures may be slightly skewed where global industry costs are concerned.
Basically, the analysis suggests that AI companies would need to generate more than half a trillion dollars in revenue to justify current spending — a figure that grows year over year.
Where is the product?
While current investor and corporate interest in generative AI technology may have boosted the market to all-time highs, including Nvidia's brief tenure as the world's most valuable company by market value, many analysts question when the real AI products or services will be. This is to maintain the growth rate.
So far, it's hard to argue that generative AI has found a legitimate use case that will lead to significantly expanding returns for those who have invested in it.
ChatGPT may be the industry's flagship product, but there's little reason to believe it will suddenly explode into the mainstream.
Simply put, if OpenAI's 10-figure profit margin were to hold the largest market share, it would take decades to reach the $600 billion revenue mark. Generative AI has yet to achieve the same value proposition as machine learning, but investments continue to improve at the VC, government and enterprise levels.
This could very well suggest that the AI market will soon enter an “AGI or bust” era, with companies like OpenAI and Anthropic making the right bets on the viability of delivering machine learning. Also people.
On the downside, companies at the heart of the generative AI sector could face tough times for earnings. If the market fails to provide enough evidence to put Nvidia's position at $3 trillion or so in the near future, that $600 billion hunger for the industry could expand to a point of no return.
But, on the positive side of things, if the industry actually creates AGI, there will be no point of return. And Nvidia is also key to this scenario.
Sequoia Capital also noted that Navi is preparing to train its new Blackwell-based chipset (called the “B100”) for generative AI. The B100 is said to be 2.5x faster than the current industry standard for training models (Nvidia's H100) and only 25% cheaper.
If the experts believe that it was possible to figure out AGI before Nvidia's latest and greatest chip came out, it should follow that it will be easier with hardware that shows a 150% increase in power and efficiency.
RELATED: SoftBank Lost 99% When Dotcom Bubble Bursts, Now It's All In on AI
