CEOInterviews.AI
Start App
Scott Galloway
Founder; Host of The Prof G Pod; Co-host of Pivot and Prof G Markets, Prof G Media

The State of the AI Market with Scott Galloway

📅 Oct 18, 2026 AI:ROI Conversations with Section 32 MIN 16 VIEWS 31 SEGMENTS · 2 SPEAKERS
We’ve been hearing about an AI bubble that may or may not pop, so we check in on that at each conference. And to lead us as always we hosted Section founder and Pivot co-host Scott Galloway - along with Section CEO, Greg Shove. Join the next live event at sectionai.com/events

What Scott Galloway said

Written from the verified transcript and checked against it. Every figure links to the moment it was said.

Scott Galloway expressed anxiety about AI safety, questioning whether AI CEOs' warnings are genuine or a form of regulatory capture to slow competitors. He criticized the lack of government oversight, comparing it to a scenario where the CDC would embed scientists in frontier labs. He argued that regulation is a feature, not a bug, and that China regulates AI more than the US, with 75% of its population feeling positive about AI. He opposed government stakes in companies like OpenAI, calling it socialism, and advocated for regulatory guardrails instead. He predicted China will engage in AI dumping, making open-weight models 80-90% as good at 10% the price, potentially crashing US GDP and capex. He advised CEOs to separate existential risk from business value, focus on practical AI adoption, and reward employees who use AI.

Key takeaways

  1. Galloway believes AI CEOs' warnings may be regulatory capture to slow competitors, not genuine existential threats.
  2. He predicts China will flood the market with open-weight AI models, potentially crashing US GDP and capex.
  3. He advises CEOs to separate existential risk from business value and focus on practical AI adoption.
  4. He says 75% of China's population feels positive about AI versus 25% in the US due to regulation.

Numbers and commitments

FigureWhat it refers toTypeAt
92% of US GDP growth from AI metric 2:49
72% of earnings from AI metric 2:49
75% of China's population positive about AI metric 9:42
25% of US population positive about AI metric 9:42
60-70% of token usage in US from open-weight Chinese models metric 22:18
98% of revenue captured by US models metric 22:18
99% of shareholder value captured by US models metric 22:18
80-90% as good as frontier models at 10% of price metric 22:18
10% of price for open-weight models metric 22:18
30 months for China to do to Silicon Valley what Tokyo did to Detroit timeline 25:52

Chapters

  1. 0:00Anxiety about AI safety
  2. 2:49Regulatory capture and government oversight
  3. 8:42China's AI regulation and public sentiment
  4. 9:42Empathy for AI CEOs and need for guardrails
  5. 17:55Opposition to government stakes in AI
  6. 22:18China's AI dumping and market bifurcation
  7. 27:13Advice for CEOs on AI adoption
  8. 30:15Practical AI implementation and rewards

Questions asked in this interview

6
  1. 0:02What has you so anxious and what can't you figure out in this moment?
  2. 2:34Related to that, Scott, is this idea that, well, you know, they're running out of compute and the capital to fund compute, so the slowdown is good for them, you know?
  3. 8:07Like, what's your sense of what the companies could be doing themselves?
  4. 21:21Do you think that could be contributing to the faster adoption of enterprise AI even inside of China?
  5. 24:41Yeah, and the data seems to support that, by the way, right?
  6. 26:37As a CEO today, what's your advice in terms of pushing the internal adoption of these technologies on a population of employees who are waking up every morning and reading that these things might kill us? What do you say?
Interviewer 0:02 ↗
You joined us on stage for our customer event in person. This event, Scott, has customers and prospects from all over the world. There's probably two or three thousand people joining us right now. You said two things on stage yesterday that I hadn't heard you say maybe ever, or certainly in a while. One was you're very anxious. That I've heard you say before. You said you're very anxious, and the second thing you said was you're not sure what's going on. And related to AI, AI safety, and all the news we've been absorbing the last few days, the fact that AI CEOs are telling us that their products might kill us and at the same time telling us we should buy shares in their IPOs. What has you so anxious and what can't you figure out in this moment?
Scott Galloway 0:55 ↗
Well, some of it might be I'm just getting old and I'm getting depressed and more anxious. Some of it might just be age-related, but I'm boasting right now. I get asked to go on cable news every day and the brief is the following: What's going on with AI? And I write back, you know, I don't know. I can list a range of possibilities. But the thing that's creating the most anxiety, there's really two levels of anxiety here. One is it's very believable that these are somewhat responsible individuals who have real concerns and think, you know, I have enough money. I do have some fidelity to the nation and humanity. There's some things we don't understand and it makes sense to slow down. Okay, that's possible. It's possible that they've decided that the most elegant form of regulatory capture would be to use recent incidents as an opportunity to create regulation where the federal government tells them to tap the brakes, it makes everybody else slam the brakes on. And that there's only a small number of companies who have the resources, the connections to elegantly comply with this regulation, thereby slowing down a lot of people. A lot of people would argue that a lot of the regulations implemented by the FDA are a function of essentially sequestering the market to a small number of well-capitalized, well-connected players. That's a possibility. It could be that this is more catastrophizing and fundraising. I mean...
Interviewer 2:34 ↗
Right. Related to that, Scott, is this idea that, well, you know, they're running out of compute and the capital to fund compute, so the slowdown is good for them, you know? Tap the brakes for a few months and because the data centers are running into headwinds.
Scott Galloway 2:49 ↗
So I can sound like a good citizen while slowing down, letting my capacity catch up. And despite all the threats that they supposedly are concerned about, they haven't slowed down Goldman or JP Morgan calling their high-net-worth clients telling them to get out their checkbook for their IPO. So they don't seem to be... The concern has not leveled to the threat level where they would dare do anything to slow down this liquidity event that their Series A through F investors are expecting. So there's concern and there's concern. What I think is causing the most anxiety is the following. I think Americans have taken for granted that we have thoughtful people who are at somewhat of an arm's distance who are charged with protecting a tragedy of the commons, who when threats like this arise send competent people in to make decisions and potentially implement some sort of guardrails called the government and regulators. What's happened here is tantamount to the following. If a scientist from Moderna, some wunderkind scientist from Moderna quit and said, 'I think the COVID vaccine we're distributing has a 10% likelihood that it turns into an aggressive pathogen that kills the host and that is the most infectious thing ever, and I think there's a one in ten chance it could kill all of humanity.' That would get a lot of news. What would get more news is if the next day the CEO of Moderna came out and said, 'You know what? You may have a point.' And then would the government, would the president go, 'Oh, shut the fuck up, Moderna. Keep selling your shit and making more money and take the stock price higher. I'm smarter than you. I know more about pathogens and epidemiology. Trust me on this. I'm a smart businessman. Keep pumping out those vaccines.' So, the fact that no one is home, the fact that there's no adult supervision... Think about, I think the CDC is arguably one of the best brands in history and I still believe that the US government over the last 80 years is the best performing organization in history despite how much we like to shitpost it. It's turned back fascism, invented vaccines, the leather bomber jacket, DARPA, internet, GPS, and food stamps. The US government is the best performing organization in history. The Center for Disease Control, what is the greatest threat to the control of disease right now? The greatest threat is a supervirus being developed in one of these frontier labs. The CDC under any responsible, competent administration that didn't have a head up their ass, a fucking conspiracy theorist running it, would already have scientists, technologists, and all embedded at these frontier labs asking every model, 'Please, using all the detergents and liquids and things I might have in my garage, please mix up a bio weapon and then sing what it says.' And these people should have access to any room. You do not ever say no to these people. They have subpoena power. They're on a direct line to the FBI. And America at least would have some reasonable certainty that there's a fucking adult in the room. It's like you find out your kid is dealing meth or you find out the neighbor's kids are dealing meth. And you find out the adults are giving them the credit card and want them to continue to sell more. So, is this an existential threat? Is this fundraising? Is this catastrophizing for the IPO? Is it thoughtful CEOs trying to slow down? I don't know. What I do know is I have absolutely no confidence that we have the domain expertise and skills in the White House and that what we have is a group of people who see an opportunity to get a vig and take what has happened over the last 40 years to an unhealthy extreme. And that's the following. The story of corporate America, the story of technology over the last 40 years is let's accrete a massive amount of wealth to a relatively small number of people and let's socialize all of the risk and all of the downside. So, socializing the anxiety of an existential extinction event in exchange for hopefully continued shareholder growth where basically the administration has gone all in on AI. 92% of our GDP growth from the cap back, 72% of earnings are from this. That's all they care about. He's like, I can't afford the S&P to go down. So, okay, if it's a 10% risk, it's probably not going to happen before I die in suite 409 at Mar-a-Lago. I'm just not that concerned. So, anxiety from the unknown, but my real anxiety here is that we have taken for granted that there needs to be adult supervision with subpoena power that is thoughtful and is trying to prevent or looking after the well-being of American society. And you get the sense that that just isn't present anywhere.
Interviewer 8:07 ↗
Did you have any empathy, Scott, for the CEOs and management teams of these AI companies, the frontier labs, meaning they're being told by the government to beat China, they're being told by the investors to beat each other and return capital, every employee walking around the hallways there at these companies, you know, has been dreaming of their IPO the last six, seven years. This is all pretty new stuff moving fast. These are young management teams. Where do you come out on? Can their boards be more effective here in addition to oversight and regulation? Like, what's your sense of what the companies could be doing themselves?
Scott Galloway 8:42 ↗
I can't help it because I'm reading the comments, but I will come back to your question. Someone just said snarkily, 'Let's just give up the goose to China.' Don't be such a fucking idiot. China regulates the fuck out of AI. You are not allowed to have an AI bot pretend to be a person for more than a certain amount of time because they don't want our kids developing, their kids developing psychoses. They do not let kids use AI after certain hours. If you're a company using AI to develop anything resembling a health product, you have to have a CCP regulator in there. To believe that if we regulate these companies and give some surety that there's not a mass extinction event is somehow going to let China win is just stupid and a lazy talking point. From a public point of view, we... Yeah, your point of view is we cannot socialize the risk and we can still get this sort of technological, business, and competitive upside. Like, it's not one or the other. We could do both.
Regulation, going back to the empathy, I do feel for these guys. They don't know what rules they're playing by. If Ford, if there wasn't regulation from the EPA saying you cannot pour mercury into the river, Ford would be presented with a terrible choice of either I'm at an economic disadvantage and have lower profits and less money to put into R&D because General Motors and Stellantis are pouring mercury into the river saving a lot of money on waste disposal. Setting out reasonable guardrails helps companies scale. Everybody gets to play by the same rules. They know the rules they're playing by. These guys don't know. They have no sense of what the rules are. The Biden administration probably was a mix of two things: overregulation in some areas of crypto and too little regulation in others. So they didn't know, am I supposed to be doing this or could I go to jail? The EU and Canada, in my view, overregulate. But there is a sweet spot. You do need some guardrails. Pesticides and fossil fuels, we're net gainers from them. We should have them. But we have emission standards and we have an EPA. We took 30 years to regulate tobacco. We took 20 on opiates. It'll probably take 20 on social media. But regulation in most democracies and capitalist societies is a feature, not a bug. Oh, when I get into this plane, I know the FAA has reasonably certified there is an acceptable risk that it won't crash. Oh, when my doctor prescribes Lexapro, I have reasonable certainty that because of the FDA, it takes 10 years to A/B test and do blind studies that I will not have psychosis taking this thing. That is key to scaling a business. It's key to industry. And the notion that if we immediately do anything that China is going to take over and get ahead of us, China is regulating AI more than us. And 75% of their population feels good about AI versus 25% here because they think, 'Okay, there's parents home. Somebody's actually looking at the existential threats.' I do, to a certain extent, feel for these guys because they're being pulled in so many directions. Okay, I have people in my company signing letters saying that we're going to destroy humanity. I have every investor who I play golf with and hang out with and go to Davos with hoping that his septic tank of a portfolio gets saved when he goes public because his $3 million investment he put in the B round is going to be worth $3 billion, which will save his ass and turn him from a goat to a hero in the world of venture capital. Probably got hundreds, if not thousands, of employees who have worked their ass off, have nothing but student debt, barely can pay their rent on their $220,000 a year salary in San Francisco, and can't help but look at the $14 million in equity they have at the IPO, and they've already told their husband to go shopping for a house. So, these guys are being pulled in so many directions. And regulation that said, 'This is how capitalism works. There's a set of rules from the government. They're pretty clear. Right? You put out a car, you have to test it a bunch of different ways. You want to put out a drug, you got to make sure it doesn't start killing people or paralyzing them, right? You are not allowed to acquire your biggest competitor such that you have monopoly power.' And then every CEO says, 'Okay, I get the rules. I understand the rules. Now, I'm going to go so fucking hard and try and beat everybody and engage in this wonderful full-body contact sport called capitalism competition. And I get I don't have to pretend that I'm a thoughtful guy or that I'm saving the whales.' It is just ridiculous. I love how the press and the public keep waiting for the better angels of CEOs to show up. These guys would sleep with their sister for a nickel. And guess what? That's their job. And then we tax them along the way, make sure they're not doing too much stupid shit that will hurt the populace, and then we use that money to pay for the Navy and food stamps. And guess what? It's an imperfect construct, but it is the least imperfect construct in the world. To a certain extent, these guys are doing their job. We're not doing our job. We're not electing people who have the domain expertise, the leadership, or the will to go in and regulate this industry. There is a vision of an amazing AI where it's like, 'Okay, you got to test. We're going to assemble a group of social scientists, technologists, philosophers, and you have to submit every single product, every single iteration to this blue-ribbon panel, and they test on it a million ways around the most existential threats, whether or not it depresses kids, whether or not it hacks their credit card information, whatever it might be, convinces kids to hate their parents,' and we say, 'Okay, you can release this.' That's just like the bare minimum. The bare minimum. And then if we had a thoughtful government, they might use this as an opportunity to kiss and make up with China and go to China and say, 'You know what? You don't want your kids killing themselves. You don't want a middle power coming up with a nuke or a bioweapon. There's things we compete on, but there's a lot of overlap. We coordinate with a lot of our adversaries and a lot of our enemies on bioweapons. We have Interpol. We have the IAEA. We have universal electronic settlement payments where we all agree. We used to fly over Russian airspace because we saw it as mutually beneficial, so we shared airspace and we said even though I fucking hate you and you fucking hate me, we're not going to shoot down each other's 787s.' This is an opportunity for global cooperation around a very, very dramatic, powerful technology. And instead, what we're doing the equivalent of is if this were 1946, Andreessen Horowitz has gone to Oppenheimer and said, 'I got an idea. I'm going to fund something called Oppenheimer.co and we're going to get Maytag and General Motors and we're going to convince them to build nuclear bombs at scale. And then I know the president, I've given him money, we've given him shares, we're going to fund his son's golf course, and we're going to sell nukes into Procter & Gamble in the UAE.' What do you think Pakistan would have paid for a nuclear bomb in 1946? And they couldn't have bought just one because India would have bought two. And then the very talented biz dev person from Oppenheimer.co would have gone to India and gone, 'Guess what? Pakistan just put in an order for a dozen.' Can you imagine the margin Oppenheimer.co could have got on nuclear weapons in 1946? These guys claim this is more powerful than nuclear fission. Would we have let Oppenheimer sell bombs and then let JP Morgan and Goldman take them public? Now, I'm not suggesting we nationalize AI. And by the way, I think Truman and FDR, if they were president, would have already nationalized these things. I'm suggesting that if we are to believe these CEOs that this is more powerful than nuclear fission, then would we let Oppenheimer sell nukes without a lot of regulation? So, I do feel for these guys. They're getting pulled in a number of directions. The public is ignorant. They keep waiting for Mark Zuckerberg to become a good person. Don't hold your breath. He's doing his job. We're not doing our job. We're not electing talented, thoughtful people with subpoena power to go in and say, 'Until you can convince me someone's not going to make a bio weapon with kitchen supplies, you are not releasing this into the public.'
Interviewer 17:41 ↗
Thought I'm suing Facebook. It seems simple when you frame it up, Scott.
Scott Galloway 17:45 ↗
Which is giving the guardrails so they can go faster, actually, so that the business can thrive and then tax them appropriately. That'd be great if we...
Well, okay, so let me go to the other side. This fucking idea of taking a stake in OpenAI? When the government starts taking stakes, okay, that is the definition of socialism. Socialism, which the Republicans have weaponized against Democrats. Socialism is who controls the means of production. When the Trump administration takes a stake in Intel and a golden share in US Steel, that is the definition of socialism. And what that does is it puts AMD and other chip companies at a disadvantage because the government now has a vested stake in elevating Intel above other companies. In addition, it's a terrible idea. Whenever the government falls under the delusion that they're good at business, you end up with a warehouse full of unsold DeLoreans in Ireland or you end up with Air France. There are a small number of things the government does better than anybody. They're really good at the Navy. You need one Navy. They're really good at Veterans Affairs, or there's one company that should do Veterans Affairs. And there are also companies that should be in the government that are privatized. We should not have privatized and made for-profit prisons because that creates an incentive for them to come back to prison and an incentive for overcrowding and an incentive for laws where you steal your third antenna and you go to jail for the rest of your life. We should not have monetized health care. Greatest source of anxiety in America right now is that if your wife gets lung cancer, there's a 30% chance you're going bankrupt. So, there are in fact some things that should go under government control for scale and for empathy and for cost control. 45% of every insurance dollar to health insurance goes to profit and administration. So, as inefficient as government is, if we effectively went single payer and had one source of insurance, you got 45 points of margin to play with. But the notion that the government should get involved in technology companies, and that's the problem. Donald Trump thinks that he's trying to bring a business perspective to government. No. We make a lot of money along the way creating an operating system of regulation and inspiration and sometimes systemic funding for the entire sector and great universities and laws and an ecosystem that supports entrepreneurs, and then we tax the fuck out of them along the way. We're going to tax your payroll taxes. We're going to tax you when you go buy a home. We're going to tax your employees when they go buy a cup of coffee, when they sell their shares. If they sell them after a year, it's 23.8 or 37 within a year or 51 or 52 if you live in Oakland or Short Hills, New Jersey. We make money along the way. Our job is to create infrastructure such that you don't do anything too damaging. It creates a level playing field and then we let you go at it and fight like fucking gladiators. But when we start picking winners and losers, as Obama did with Solyndra, it just doesn't goddamn work. So, the notion that we might benefit from the government getting involved in an ownership stake, no, no, no, no, no. Get involved at a regulatory standpoint for safety, helps them scale, helps give CEOs clarity around what they can and cannot do. But we're in the worst of all worlds: a total lack of government oversight that creates tremendous anxiety and real risk, but at the same time a guy who thinks he can pick winners and losers in specific sectors and somehow we're going to benefit from that.
Interviewer 21:21 ↗
Yeah, Scott, as you know, we've got a bunch of global clients at Section, and what I'm beginning to hear just in the last few months from those customers who have operations in China is that their AI transformations are moving faster in China than in the rest of the world in terms of their business units or their subsidiaries in China are sort of ahead of other regions. And of course the easy answer is, well, that's just China, they can tell people to use it, you know, and if they don't, they'll just fire them. But what's your take on that? Do you think there's any sort of lesson there around perhaps what you talked about earlier around China's regulating AI, 75% of the population is more optimistic about it? Do you think that could be contributing to the faster adoption of enterprise AI even inside of China? Because we're off to a pretty slow start here in the US, meaning large organizations really actually adopting and using these technologies to any economic value or gain.
Scott Galloway 22:18 ↗
But we have captured... So I've read that 60 or 70% of the token usage now in the US is from open-weight Chinese models, but that 98% of the revenue has been captured by US models and 99% of the shareholder value has been captured. I see the market as bifurcating, and that is we're becoming German automobile companies, more prestige, better models, higher margin, and they're becoming Toyota and Kia. Now the biggest news today was Carney has basically said, all right, I'm in an abusive relationship, I'm going to pull up Hinge and I'm swiping right on 27 EU countries. They're creating alliances to go around who has typically been their biggest customer and their biggest supplier. If I were advising Xi, I would do the following: subsidize open-weight models, flood the market, and engage in AI dumping similar to what they tried to do with steel in the '80s. And if all of a sudden all of these companies go, 'You know what? The open-weight models are 80 or 90% as good as the frontier models at 10% of the price,' and all of a sudden OpenAI announces its first flat-to-down quarter, you're going to see GDP and capex in the US crash. You're going to see S&P throw up. I think economic warfare, I would absolutely go after the soft tissue of America if I were China and start engaging in AI dumping. So, but these open-weight models, the real fear and the real potential of AI is the following. The University of California San Francisco can reduce sepsis death, which is the number one killer in hospitals, because they are so outstanding in what they do, and they have so many resources and such advanced technology. A rural hospital in Arkansas registers a lot of death from sepsis. AI may be able to implement fairly inexpensively a series of practices where the AI immediately looks at the chart and the makeup of the patient and says the following three things need to be done right away. It might level up a lot of different, you know, the one-person startup that can create a $10 million company, right?
Interviewer 24:41 ↗
Yeah. Yeah, and the data seems to support that, by the way, right? The US new business formation data is really encouraging. And I think it's not a coincidence. It is what you're saying, that one- or two-person, you know, AI-powered services startup.

11 more exchanges in this transcript

Sign in free to read the rest of this interview. No card required.

Sign in to read the full transcript

Cite this transcript

APA, MLA, BibTeX
APA

Galloway, S. (2026, October 18). The State of the AI Market with Scott Galloway [Interview transcript]. AI:ROI Conversations with Section. CEOInterviews.AI. https://ceointerviews.ai/interview/3086079/

MLA

Scott Galloway. "The State of the AI Market with Scott Galloway." AI:ROI Conversations with Section, 18 Oct. 2026. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/3086079/.

BibTeX
@misc{galloway2026_3086079,
  author       = {Scott Galloway},
  title        = {The State of the AI Market with Scott Galloway},
  howpublished = {Interview transcript, AI:ROI Conversations with Section. CEOInterviews.AI},
  year         = {2026},
  month        = {oct},
  url          = {https://ceointerviews.ai/interview/3086079/},
  note         = {Speaker-attributed transcript with timestamps}
}