HigherIQ — Saturday, January 17, 2026

Good morning, it's Saturday, January 17th, 2026. Markets are closed for the long weekend, which gives us time to step back and look at the bigger picture on two stories that will shape the year ahead. Let's start with the week's close. The major indexes finished Friday essentially flat — the S&P at 6,940, Nasdaq down six basis points, Dow off about two-tenths of a percent. But here's what's interesting: while the broader market drifted, chip stocks had a very different week. TSMC, Broadcom, and AMD all climbed, with Broadcom up two and a half percent on Friday alone. The surface explanation is TSMC's blowout fourth-quarter earnings. The real story is much bigger. On Thursday, the US and Taiwan finalized a trade deal that represents the most significant reshoring of semiconductor manufacturing in decades. The numbers are staggering: Taiwanese chip and tech companies have committed to at least 250 billion dollars in direct US investment, with another 250 billion in credit guarantees to support the broader supply chain. In exchange, the US is cutting tariffs on Taiwanese goods from 20 percent down to 15 percent, with zero tariffs on generic pharmaceuticals, aircraft components, and certain natural resources. This isn't just a trade deal. This is industrial policy at a scale we haven't seen since the original CHIPS Act, and it fundamentally changes the math for anyone building or investing in semiconductors. Here's why it matters for you. If you're invested in chip stocks or tech broadly, this deal provides something the market has been desperate for: clarity. For the past year, companies like TSMC have been navigating a fog of tariff uncertainty, unsure whether building in the US would be rewarded or punished. Now there's a framework. Companies building new US chip fabs can import up to 2.5 times their planned capacity tariff-free during construction. Once the factories are running, they can still import 1.5 times their US production without paying duties. That's a massive incentive to keep building. TSMC is already responding. The company boosted its 2026 capital expenditure budget to as much as 56 billion dollars — up nearly 40 percent from last year — and announced it's accelerating construction on its Arizona fab cluster. Commerce Secretary Howard Lutnick said the administration's goal is to bring 40 percent of Taiwan's semiconductor supply chain to American soil. For context, the US share of global chip fabrication dropped from 37 percent in 1990 to under 10 percent in 2024. This deal is an attempt to reverse that trend. But there's a geopolitical dimension that matters just as much. This deal makes Taiwan even more strategically central to the US economy at a moment when China is watching closely. Beijing called the agreement "economic plunder" and slammed Taiwan for signing anything with "sovereign connotations." The more the US relies on Taiwanese investment and expertise, the higher the stakes become if cross-strait tensions escalate. For investors, that's both opportunity and risk — chip stocks may have more runway, but they're also increasingly tied to geopolitical variables that are hard to model. Now let's talk about AI, because there's a major business model shift happening that most people missed this week. OpenAI announced it will start testing ads in ChatGPT for the first time. Ads will appear at the bottom of responses for users on the free tier and the cheapest subscription level. They'll be labeled, they won't appear on sensitive topics like health or politics, and OpenAI says it won't sell user data to advertisers. Two years ago, Sam Altman called advertising a "last resort." Now it's happening. And the reason comes down to simple math. OpenAI has an infrastructure roadmap reportedly on the order of 1.4 trillion dollars over time. That's trillion with a T. Meanwhile, the company generated roughly 20 billion dollars in revenue last year. Even with generous margins, that doesn't pencil out. The gap between what these AI systems cost to build and what they currently earn is enormous, and something has to give. This is the tension that will define AI in 2026. The technology is transformative, but the economics are brutal. Training runs cost hundreds of millions of dollars. Inference at scale requires massive data centers. And users have been trained by years of free internet services to expect powerful tools without paying much — or anything — for them. Ads are the obvious pressure release valve. If you can't charge users more, monetize their attention instead. But this creates a new set of questions. Will users trust AI responses that are subsidized by advertisers? Will the quality of answers change if engagement becomes the metric that matters? And what happens when the same companies building AI are also the ones deciding what information you see and how it's framed? This connects directly to what's happening in Davos this week. The World Economic Forum kicks off Monday, and one of the central themes is the "responsible deployment" of generative AI. A WEF survey of 4,600 young people across 144 countries found that two-thirds of respondents between 18 and 30 fear AI will reduce entry-level job opportunities over the next three years. That's not abstract anxiety — it's the generation entering the workforce right now looking at these systems and wondering where they fit. Meanwhile, the companies building AI are under intense pressure to make the economics work, which pushes them toward aggressive monetization, which accelerates adoption, which increases the pressure on jobs. It's a feedback loop, and we're only in the early innings. For anyone thinking about career positioning, the actionable insight here is straightforward: the value of skills that AI can replicate is declining, while the value of skills that complement AI — judgment, relationship-building, creative direction, domain expertise — is increasing. The young people worried about entry-level jobs aren't wrong to be concerned, but the response isn't to avoid AI. It's to become the person who knows how to use it better than everyone else in the room. A few more stories worth noting before we close. President Trump is heading to Davos next week and has already signaled he'll use tariff threats to pressure Denmark on Greenland. He said he "may put a tariff on countries if they don't go along with" his push to acquire the Arctic island. Denmark and the US agreed earlier this week to form a working group on the issue, but Trump hasn't ruled out more aggressive measures. For anyone in import-dependent businesses or investing in European markets, the tariff-as-diplomacy playbook is expanding beyond trade into territorial disputes. On the Fed front, Kevin Hassett's chances of replacing Jerome Powell as Fed Chair dimmed after Trump said he wants to keep Hassett in his current role as National Economic Council director. That boosts the prospects of former Fed Governor Kevin Warsh, though any nominee may face hurdles from Republican Senator Thom Tillis, who has vowed to block confirmations until the criminal probe of Powell ends. Markets are watching this closely — whoever leads the Fed will navigate the tension between Trump's desire for lower rates and the institution's independence. And one lighter note: Dos Equis is bringing back "The Most Interesting Man in the World" after a 10-year hiatus. The original actor, Jonathan Goldsmith, now 87, will reprise his role starting with a College Football Championship ad on Monday. It's a nostalgia play for a brand that saw US sales drop 8 percent last year, but it's also a reminder that sometimes the best marketing strategy is admitting your replacement idea didn't work. That's your Saturday briefing. A landmark chip deal reshaping where semiconductors get made. An AI business model shift that signals how these companies will actually pay for what they're building. And a week ahead in Davos where all of these threads — trade, AI, geopolitics — will converge in one very expensive ski resort. Stay curious, stay informed, and we'll see you on Sunday.