Trump's Tariffs Return, Kalshi's Midterms Hub, and Elon's AI "Odyssey"
Kara Swisher & Scott Galloway
24 Jul 20265 min read1h 05m
TL;DR
Trump's revived tariffs — now invoking the 1930 Smoot-Hawley Act to hit Canada with 50% duties — are pushing inflation above 4% and costing the average U.S. household $1,000–$1,200 a year, while driving Canada toward closer ties with China. Meanwhile, Chinese open-weight AI models have surged from 10% to 58% of global token usage since January 2025, and Galloway argues China has already won the volume race even as American frontier models retain a premium-tier advantage.
Key Moments
Scott Galloway
“tariffs will take approximately uh $1,000 toll on every household this year. It'll go to $1,200 on every household average if these tariffs go through. And about 90% of uh those tariffs are being borne by US consumers.”
Galloway quantifying the direct consumer cost of Trump's tariff policy
“In January of 2025, which wasn't that long ago, about 10% of token usage was coming from Chinese openweight models. As of today, it looks like it's 58%.”
Galloway presenting data on how fast Chinese AI models have taken over global token usage
“I would argue that AI, I mean, on a volume level that China has already won. And it's not a race to see who has the smartest model. It's a race to see who can put the most AI in the most people's hands.”
Galloway reframing the US-China AI competition as a distribution race, not a capability race
“The kids whose parents take it away are more depressed because they're ostracized and isolated from everybody else. So unless it's a collective universal ban, it doesn't work.”
Galloway citing NYU research by Adam Alter on why individual parental bans on social media backfire without a collective policy
Pivot is a twice-weekly tech and business podcast from New York Magazine and the Vox Media Podcast Network. Hosts Kara Swisher and Scott Galloway analyze the week's biggest stories at the intersection of technology, business, and politics. This episode was recorded live at the CAO Summit in Los Angeles.
Takeaways
1
China already dominates AI by volume Chinese open-weight models went from 10% to 58% of global token usage between January 2025 and mid-2026. Galloway argues the AI race isn't about model quality — it's about distribution, and on that metric China has already won. American frontier models survive as a premium tier, like BMW vs. Toyota, but the mass market belongs to DeepSeek and its peers.
2
U.S. benefits more from Canada trade than Canada Galloway breaks down the trade asymmetry: U.S. exports to Canada — iPhones, financial services, advertising — carry 40-60% gross margins at 30x P/E multiples, generating ~$1,500 in shareholder value per $100 shipped. Canada's timber and unrefined oil exports carry 20-30% margins at 10x multiples, generating $2-3 in value per dollar. By this logic, Canada has far more economic justification to impose tariffs on the U.S. than vice versa.
3
Individual social media bans backfire without collective policy NYU research cited by Galloway found that children whose parents individually ban social media become more depressed — not less — because they are socially ostracized from peers who remain on the platforms. The intervention only works as a universal, policy-level ban. France (under-15) and Australia (under-16) have passed blanket bans; both hosts expect similar legislation to reach the U.S. soon.
4
Tariffs cost average U.S. household $1,200 Galloway estimates current tariffs levy a $1,000 annual toll per U.S. household, rising to $1,200 if the new Canada tariffs fully take effect, with roughly 90% of the cost borne by American consumers. With inflation already running above 4%, he traces two of those four percentage points directly to energy prices and tariffs. The Smoot-Hawley Act invoked to justify Canada tariffs was, until now, considered a cautionary relic studied in grad schools.
5
Prediction markets beat polls on accuracy Kalshi has never gotten a Fed rate decision wrong and is reportedly 78% more accurate than traditional polls at predicting election outcomes. Galloway attributes this to the 'wisdom of crowds' effect — people reveal true preferences when money is at stake, not when answering surveys. Polling firms are going out of business as a result, even though election wagering is only 1-2% of Kalshi's revenue.
6
AI companies' political donations are self-protective, not regulatory Anthropic doubled its political spending to $40M backing a pro-AI-regulation PAC, but both Swisher and Galloway dismiss it as theater — comparing it to Zuckerberg's years of calling for regulation while lobbying against it. Galloway frames it as buying access: a ticket into the Washington networking conference rather than genuine policy advocacy. Elon Musk has separately committed $90M to the midterms.