President & Chief Investment Officer, Alphabet Inc. (Class A)
Google CIO Ruth Porat on AI Optimism, Data Centers, and the U.S.-China Energy Race
Ruth Porat, President and Chief Investment Officer of Alphabet and Google, lays out the case for AI optimism, a projected $4 trillion lift to U.S. GDP, Nobel Prize–winning breakthroughs like AlphaFold, and gains in cybersecurity and healthcare, in conversation with IBM Chairman and CEO Arvind Krishna.
From there, the conversation turns to one of AI's most contested questions: energy. Porat breaks down why data centers aren't the primary driver behind rising electricity prices, how Google is investing in added capacity and demand response to offset its footprint, and why the U.S. risks falling...
About Ruth Porat
Ruth Porat, President and Chief Investment Officer at Alphabet and Google, has recently discussed the economic and societal potential of artificial intelligence, as well as the energy infrastructure required to support it. In a CNBC-hosted forum with IBM Chairman Arvind Krishna, Porat said she believes AI could add approximately $4 trillion to U.S. GDP over several years and cited data that 80% of small businesses using AI report an uplift in revenue and profitability. She described AI as a tool that will reshape tasks rather than eliminate careers, and emphasized the need for job skilling efforts to ease workforce transitions. Porat also highlighted AI’s role in advancing science, pointing to Google’s AlphaFold as an example of open-sourced research used by 3.5 million scientists across 190 countries. On energy, Porat stated that data centers currently use about 4% of the U.S. grid, with projections rising to roughly 12%, but argued they are not the primary driver of rising electricity prices. She cited a Lawrence Berkeley National Labs study indicating that electricity prices have grown more slowly in states with data centers. Porat noted that China has built 80 times more transmission capacity than the United States in the last decade, and said that without sufficient energy infrastructure, the U.S. risks falling behind in delivering AI’s benefits. In other appearances, Porat discussed Google’s use of office food and “casual collisions” to foster innovation, and shared personal stories about her father’s background as a Holocaust survivor and physicist. She also addressed health disparities, stating that treatable cancer survival rates can be 50% lower depending on a patient’s postal code, and described Google’s work using AI to screen for diabetic retinopathy in the global south.
Transcript (6 segments)
A
Arvind0:00
Let's just think about all that we're thinking about nowadays. We're thinking about technology, we think about AI, we think about capital, we think about jobs. And if you look at Ruth's background, who better to sort of have this conversation with around this topic? And there's a lot being written nowadays about both AI and about technology, the outsized role that these are playing in almost every industry. So, Ruth, can I begin with why should we have optimism around these topics?
R
Ruth Porat0:35
Great. Well, first, it's so wonderful to be back here. I always love coming to New York, but looking around this room and seeing so many old friends, it's an absolute treat and great to be with you always, Arvind. So, the upside from AI, when we look at it, I think at a high level, there are four really profound areas that lead to optimism. Number one is the economic upside. In the US alone, it's estimated that the upside is about four trillion to GDP over the next several years. And we're already seeing with small businesses that are applying AI to whether it's how they reach out to customers, how they engage with customers, how they manage chain, etc. They're seeing 80% of them are seeing an uplift in revenue and profitability. The second that I think is really exciting is advancing science, a place you and I spend a host of time. And probably one of the best examples of that is a colleague of mine who won the Nobel Prize for something hopefully many have already heard about — AlphaFold, which is about protein folding, and it has been described as one of the greatest contributions to drug discovery in our lifetime. What they did is they said they wanted to take on a 50-year grand challenge on protein folding, which is sort of the insight to life. And when they started on that journey, somebody challenged the head of our AI lab, Demis Hassabis, and said, 'How can you take on this challenge? It takes four years for a PhD student just to diagram one protein. Just one, and there are 200 million out there. How is it possible?' And his answer was, 'Why not?' And they went on that journey, and within four years, not only had they cracked the code, but won the Nobel Prize on this. And if there's one thing I think I would want to leave anybody with that I quote a lot at Google, it's with AI, that question 'Why not?' is something we should each be asking ourselves about anything that to date has been intractable, because it lets you break through it. The others obviously around cybersecurity, where we both spend a lot of time. And the final area is really better delivery of healthcare, education, addressing food security. We were talking about a whole host of really important social issues that we can now address as a result of AI. That aggregate portfolio to me is the reason we should be optimistic about what AI, responsibly executed, can deliver for society.
A
Arvind2:57
I do think that people should note what Ruth began with. Four and a half trillion, even if out there in a few years, let's call it GDP here is 40. A 10-point increase, which you can reinvest back in the country and in an individual company or into the citizens — it's massive. If you go look at developmental economics, that 10 points is all the difference between breakaway and kind of not being. So, while there is a lot of optimism, there's also people out there that are worried about some of the downsides of AI. One of the big ones that gets talked about all the time is, 'Hey, it's causing my energy bills to go up,' or all the data centers are fundamental cause of this. Could you help demystify that a little bit, Ruth, because I know you guys do a lot of work on this topic, and you have made investments.
R
Ruth Porat3:51
Well, I like that you started with the upside because if the upside's not worth fighting for, I think we all just go back into our corner. And I think the upside is worth fighting for and we can responsibly protect against the downside. And to your question, we can't have the upside of AI without the energy to power it. I think there are a couple of pretty critical issues. When I was in New York as a banker, one of the core things we all learned is if you really want to address a risk issue, you have to go at the root cause of the risk issue. And with respect to energy, it's that our country is underinvested in energy infrastructure for far too long and it's caught up with us. That being said, data centers are in the spotlight and appropriately because of the growth and the expected growth, but today, just to put some numbers around it, data centers use about 4% of the grid. Expected to go to ballpark 12%. But just to scale it. What I find really important also for context, and I'll get to how do we mitigate against it, Lawrence Berkeley National Labs did a study a number of years ago and they said for states that have data centers, electricity prices actually go up more slowly than for states without data centers. Why? Because data centers and the large workloads help amortize the fixed costs over a larger base. So, data centers through 2024 have actually helped keep electricity prices growing at a slower rate. What we have done because of this very important question, and I'm wanting to make it clear that when we show up, it's a net benefit to the community — it's a couple of things. One, we're committed to adding capacity, energy capacity wherever we invest. So, you've probably hopefully seen a lot of announcements. We're investing in nuclear, batteries, solar, wind, etc., so that we're bringing additional capacity online. Second, with technology, we've done something where on really hot or cold days, we can move our workloads out of the way. It's called demand response. Across the country, we've created the equivalent of one gigawatt of incremental capacity by saying we'll get out of the way when the community needs it. We then say not only are we paying for the energy we use, but we'll pay for any infrastructure investment needed to support that growth whether or not we use it. And on top of that, we build an energy affordability fund in the communities we go into, which works with local organizations, and they go into homes and help weatherize homes. So, all of that leakage that would otherwise show up as an energy cost on your bill drops immediately. And then, to build this energy infrastructure is actually a job creator. There are hundreds of thousands of jobs that are unfilled and will continue to be unfilled if we're not trained. So, a year and a half ago, we started an electrician training program. And whenever I show up in a town and we're going to start a data center, one of the really important elements is we have the IBEW or the electrician training alliance with us. We have somebody who's talking about the jobs that are created. And I was in Missouri a couple of weeks ago, and one of the many beautiful stories — I went to a woman who's going through the training program. I'll never forget her name. It's Danielle. Her husband had recently passed away, and she has two children. And she said, being a waitress is what she had been doing can't cover the cost for her kids and it doesn't provide the security. This is a secure job. It's a career. And it's enabling people to stay in their communities and have careers that they're really proud of. As you were saying, when we show up and do any kind of build, the multiplier effect in the community is profound. For every job we create, third parties have said there are nine additional jobs created in the community. And so, it's a job creator as well. So, when we're protecting on the downside, it goes into capacity, affordability, and investments in the community.
A
Arvind7:54
We'll come back to jobs. Let's talk a little bit more about the capacity point, because the point you were making is for probably two or three decades, there hasn't been very much investment in the grid. I live in the Northeast. I can tell you that electricity prices here have very little to do with data centers, and they're outrageous, mostly because the supply isn't there. And so, I think that now that there is attention on the topic, because there is demand, if we get more capacity — to the point you were making about states with more data centers — it all comes down to if there is more capacity, people are going to figure out more productive ways to build the capacity. They're going to get it out there, and as long as you have the fixed load, because that's what a data center gives you, then that takes away a certain amount of the fixed cost, and so there is then more supply in the community. I mean, like you play out.
R
Ruth Porat8:44
Well, I absolutely agree, and there's another element of it, which is we've underinvested across the country collectively in the grid. There's 100 gigawatts of energy that is available if our country just modernizes our grid, according to the Department of Energy. There is a substantial amount of capacity waiting to get on the grid, but we as a country have not built the transmission lines. China has built 80 times more transmission capacity in the last 10 years than the United States. 80 times. China has a clear focus that to win in technology — the US may be ahead in chips and in models, and we at Google believe we will continue to be ahead in chips and models. But you can't deliver that without the energy that's needed, and China has been focused and relentless adding sources of energy as you just said, and this energy infrastructure, the transmission capacity. And I think collectively as a national security matter,