Earlier this month, Marketplace launched our AI series looking if the industry, with its hundreds of billions of dollars spent specifically on data centers, becomes too big to fail. “Marketplace” host Kai Ryssdal also visit a Los Angeles data center and “Marketplace Tech” host Meghan McCarty Carino examined people’s negative feelings about the future of AI.
To conclude this part of the report, Ryssdal and McCarty Carino spent the day together on the ground in Silicon Valley to explore the economics of AI – and the trade-offs that come with it.
Despite the AI boom, McCarty Carino, who lives and grew up in the Bay Area, noted that this doesn’t look like a boom time for the local economy.
“There are some metrics you can look at, like the number of tech jobs in the Bay Area is still lower than it was during the pandemic,” she said. “A number of people have told me that the tech job market here hasn’t been this brutal since the dot-com bubble collapsed. »
Small Fish AI
According to analysis According to Pitchbook and the National Venture Capital Association, AI startups, of which there are thousands, generated $222 billion in venture capital investments in 2025. That means AI startups received about 65% of all venture capital dollars last year.
Ryssdal and McCarty Carino visited one of these AI startups, called TinyFish AI. It received about $47 million in venture capital funding last year. That’s small compared to companies like Anthropic and Open AI, but for AI to return the way investors expect, it will take an entire ecosystem of AI companies.
Nair told Ryssdal and McCarty Carino that the data centers needed for AI to work would be like a new form of “invisible fracking.”
Andie Corban/Market
“The problem is that as AI becomes more powerful and value-driven, the chasm between large and small companies widens,” said Sudheesh Nair, co-founder and CEO of TinyFish. “We’re intentionally trying to see if it’s possible for us to bridge that gap.”
TinyFish creates AI software that makes it easier to display information from small, independent hotel websites in search results on platforms like Google Hotels or Expedia.
“I would say we are the layer of infrastructure that we want others to build applications on,” Nair said.
Nair acknowledges that the deployment of AI across the economy – by startups like his and others – will have consequences, particularly in places where water and electricity are cheaper for data centers.
“Small communities are going to be really hit by this,” he said. “A bit like invisible fracking.”
He believes the value of AI will eventually reach them, but it could take years. “For example, if we find a new cancer drug,” he said.
Digital Real Estate Data Center
After TinyFish, Ryssdal and McCarty Carino visited one of the defining images of the AI economy today: a data center. This one, still under construction, belongs to Digital Realty. With over 300 data centers worldwide, the company is one of the largest data center colocation providers. In Silicon Valley alone, Digital Realty has 15 data centers, including the one in Santa Clara visited by the “Marketplace” team.
Digital Realty expects the new Santa Clara center to be operational and powered by early 2027.
Andie Corban/Market
Digital Realty broke ground on the facility in 2021. It spans 430,000 square feet, with an empty, identical 12-megawatt data hall on each of its four floors. Chief revenue officer Colin McLean told Ryssdal and McCarty Carino it is expected to become operational in 2027.
“Obviously, we have to work very closely with municipalities and the utility network when delivering electricity,” McLean said. “It’s a challenge if you don’t have electricity for a data center, so we have to wait for the right time.”
McLean said power will be “the problem” in the future when it comes to data centers. At Carnegie Mellon and North Carolina State University study found that data center growth could increase electricity bills by 8% nationally by 2030. In some communities, such as parts of Virginia that are home to many AI data centers, these costs could increase by more than 25%.
Although that facility is not yet operational, McLean said the company begins leasing data center space to customers 12 to 14 months before getting power.
“There’s sort of an insatiable demand for data center capacity,” McLean said. “We’re seeing a continued trend of increased consumption. Data is becoming more and more valuable, it’s almost like the new oil in terms of the value proposition.”
Digital Realty’s data centers can be used for AI or more traditional colocation purposes – think of an app or a bank that needs data center space close to its customers.
Ryssdal asked McLean about his concerns about the amount of electricity and water demand that data centers bring to communities.
“We understand and we want to be good partners for the community,” McLean said. “We believe data centers make a tremendous contribution to the digital and local economy, not only in the form of taxpayer dollars, but also in the form of jobs and how we meet infrastructure needs. »
Click the audio player above to hear the full episode.