- If you're looking for the next big thing to invest in, you might be thinking you should embrace the brave new future of artificial intelligence, and you would hardly be alone.
From 2022 to '23, private investment in AI technology increased eight-fold.
- [Julia] And last year, NVIDIA, a company hat makes the semiconductor chips that AI relies on briefly became the most valuable company in the world as investors rushed to ride the AI wave to huge investment gains.
- But anytime the price of an asset inflates that fast, it can start to look like, well, a bubble.
So are the billions of dollars that companies and investors are pouring into AI technology paying off?
- Or could the bubble be about to burst?
(bubble pops) (light bright music) - If you used AI models like ChatGPT, or Midjourney, or at least read that AI Overview that appears at the top of every Google search now, you might believe that AI technology is about to be a part of all our lives.
- [Julia] At least that's what tech companies like Google, Meta, and Amazon are betting on.
This year, these companies are projected to spend 185 billion on developing AI tools, and thousands of startups are popping up to race these tech giants to build the model that will change the way we work, create art, and communicate with each other.
Their investors are optimistic that the rise of AI will pay off big.
- But for others, all the hype around artificial intelligence reminds them of another tech boom that happened in the '90s.
In 1994, the same year that the first online shopper used a browser to order a Sting CD, a young Jeff Bezos set up the book selling website that would eventually become Amazon.
- As everyday people started using their dial up connections to send emails, and access chat rooms, and even shop, online marketplaces like eBay and Expedia soon followed, hoping to monetize the internet.
These early dot coms raised millions from investors and burned through it just as quickly, hiring employees, building servers, and taking out flashy Super Bowl ads to grow fast and beat out the competition.
- But when many of these startups failed to turn a profit by the year 2000, investor enthusiasm cooled.
Over the next two years, the NASDAQ stock market index fell 77% from its peak and dot com companies that had made initial public offerings without ever being profitable began to file for bankruptcy.
- Even though the dot com bubble burst, its technology still changed the way we live now.
You may not log onto mp3.com to buy your favorite artist's new single, but you do shop online, right?
- The early dot com industry experienced what economists call a market correction.
The companies were valued too high, too fast, so a drop in the market brought prices back down to reality.
From there, companies that couldn't turn a profit failed and companies that could make money continued to grow at a slower steadier pace.
- [Julia] Still, it left behind technology infrastructure and employee knowledge that got reused in the next wave of tech startups and consumer behavior eventually followed.
Now, one in every five retail transactions happens online.
- The dot com bubble and its aftermath may give us a preview of what could happen with AI.
Amid all this FOMO-driven investment, AI companies will have to figure out how to turn a profit in order to merit their sky-high valuations.
But in contrast to scrappy web startups, this may be even more challenging for AI tech.
- [Julia] AI companies are expensive to operate.
They require massive data centers to process all of that data that large language models, text to image generators, and other AI tools use to generate results.
Think racks and racks of all those NVIDIA chips.
ChatGPT alone runs on tens of thousands of them, and at $30,000 a pop, they don't come cheap.
- [Philip] These data centers also require lots of electricity.
ChatGPT, for example, uses as much electricity as 17,000 US households per day.
Even without considering the environmental cost of this energy demand, that makes for a pretty hefty electric bill.
- And then there are the legal issues.
Increasingly, media companies like the "New York Times" and Getty Images are suing AI companies for using their content to train AI models without their permission.
And the American government, fearful of how the technology could be used militarily, is putting restrictions on who manufacturers like NVIDIA can sell their chips to.
The result for AI companies may be less revenue from overseas and higher costs on things like legal fees and licensing agreements.
- So AI models are backed by big investments to match their big overhead.
But are they bringing in big profits?
- Not yet.
While one of AI's biggest players, OpenAI, is expected to bring in $3.4 billion in revenue this year, they're still spending more than double that on training their model.
- Tech giants are seeking deep-pocketed business clients, promising that AI will increase efficiency and productivity and even replace human workers.
But businesses are approaching AI tentatively.
In a McKinsey survey, 63% of respondents said inaccuracy of AI tools poses significant risk to their business.
- Until AI models can reliably avoid producing factual errors or freaky images of eight-fingered hands, many businesses are hesitant to fully rely on AI for their most important functions.
- There are some early signs that the market correction is coming for AI.
Even though NVIDIA stock continues to soar in 2024, major players like Meta and Microsoft saw stocks dip this year as investors doubted their all-in approach to AI.
Newer AI startups like Babylon Health and Stability AI have gone from billion dollar valuations to mass layoffs and bankruptcy filings.
- As we saw in the dot com boom, a few AI companies could become the next eBays and Amazons, but others may go the way of mp3.com.
- But when the dust settles, the remaining skilled workers and tech infrastructure could spell a future for AI that doesn't have to live up to quite so much hype.
- [Both] And that's our two cents.
(bright music)
Passport
Follow Us