Former president Donald Trump reiterated his frustration with Taiwan over the weekend when he appeared on the Joe Rogan Experience podcast and accused the country of stealing America’s chip industry.
Trump criticized the U.S. CHIPS Act and said he would implement tariffs on chips from Taiwan if elected president. Such tariffs would impact the global leader in chip building, Taiwan Semiconductor Manufacturing Company, whose customers include companies like Nvidia and Apple.
Shares of Taiwan Semiconductor closed down 4.3% on Monday.
“You know, Taiwan, they stole our chip business… and they want protection,” Trump said during the appearance. The podcast was published on Saturday evening.
Every hyperscaler, like Amazon, Google and Microsoft, working on their own in-house chip is fabbing with the Taiwanese company. UBS analysts estimate over 90 percent of the world’s advanced chips are manufactured by TSMC. Intel and Samsung are among the companies trying to compete but have faced a series of setbacks.
Given the broader geopolitical concerns surrounding Taiwan and the risk of a China invasion, pressure has been growing on U.S. companies to build an alternative to TSMC in the U.S.
Intel, which has emerged as a poster child for the U.S. CHIPS Act, has faced many challenges. “We want to get leading edge infrastructure built here in the U.S., and to be honest, from a policy standpoint, it really shouldn’t matter all that much who is building it,” Bernstein analyst Stacy Rasgon told CNBC.
Rasgon added that the idea that Taiwan had stolen our chips industry was “ridiculous.”
Taiwan Semiconductor Manufacturing Company is on tap to receive nearly $7 billion from the U.S. Commerce Department to build its Arizona foundry as part of the CHIPS Act. On the company’s earnings call two weeks ago, TSMC CEO CC Wei said its Arizona plant was making progress, with volumes expected to ramp in 2025.
The U.S. Commerce Department funds have yet to be allocated to TSMC or other major semiconductor firms. Sources say funds are expected to be allocated by the end of this year as long as specific milestones are met.
Trump also suggested foreign companies shouldn’t be able to enter the U.S. and use government money. “That chip deal is so bad,” he said. “We put up billions of dollars for rich companies to come in and borrow the money and build chip companies here. They’re not going to give us the good companies anyway.”
Mizuho analysts recently wrote that a Trump win would be bad for Taiwan Semiconductor. Analysts at Citi are debating how much tariffs could increase the costs across the chip supply chain. They add that tariffs wouldn’t be easy for governments to navigate. “[Tariffs] would require complex audits across thousand of devices, which contain a variety of chips,” the Citi analysts wrote.
Markets have been keeping a close eye on the risk surrounding Taiwan, given how dependent Silicon Valley is on TSMC’s chips. Earlier this summer, when Trump made similar comments about Taiwan, the SMH ETF lost $675 billion in market cap in one week. TSMC fell over 10%.
U.S. companies that either have fabs or are in the process of building them, like Intel, Global Foundries and Texas Instruments, outperformed on the expectation that if Trump wins, he’ll favor the domestic players.
However, a broader trade war could also challenge the sector. “[Under a Trump presidency], there are potentially big tariffs against China, which, as we have seen before, will elicit a China reaction as we saw with Micron,” said Patrick Moorhead, CEO of Moor Insights & Strategy and a top-ranked tech analyst said to CNBC.
But experts warn that if Vice President Kamala Harris wins the election, it’s not an “all clear” for the semi trade. Some of the harshest export controls on China implemented under the Biden administration dramatically impacted how much Nvidia and other semiconductors can sell into the country. Pre-export controls, Nvidia’s business in China generated over 25 percent of total sales. China now accounts for less than 10% of Nvidia’s revenue.
Every weekday, the CNBC Investing Club with Jim Cramer holds a “Morning Meeting” livestream at 10:20 a.m. ET. Here’s a recap of Friday’s key moments. 1. The S & P 500 and Nasdaq Composite pushed higher Friday, buoyed by strength in Big Tech names like Club holding Amazon . The e-commerce giant reported a blockbuster earnings report Thursday evening, highlighted by growth in its cloud computing unit. Shares are up more than 10%. Friday also marks the last trading session of October. Next week, Club names Eaton, DuPont, and T exas Roadhouse will all release quarterly results. 2. Don’t own any Apple stock? New investors should consider starting a position if shares continue to fall on Friday, advises Jim Cramer. Apple stock should be up, given its stellar quarterly earnings report Thursday evening, which showed that the iPhone maker’s shares have more room to run. Apple posted strong guidance and saw huge revenue growth in its crucial high-margin services unit. “Let it come in. And then, if you don’t own any Apple, then you buy,” Jim added. “We have the same attitude as we’ve always had, which is own it, don’t trade it.” 3. Investors should also consider buying more Nike and Boeing following a stretch of underperformance for the Club names. Nike and Boeing shares have each declined roughly 7% over the past month. “We have a bad market for anything other than tech,” Jim said Friday. The footwear giant and aircraft maker have both been unfairly punished, and their turnaround stories remain strong. “These are companies that have vastly improved,” he added. We bought some Nike stock on Friday morning. 4. Stocks covered in Friday’s rapid fire at the end of the video were: Chevron, Reddit, and Netflix . (See here for a full list of the stocks in Jim Cramer’s Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
As the tech industry’s giants race to build out AI infrastructure — Microsoft spent $34.9 billion in just one quarter while Meta plans to spend up to $72 billion this year — they may not be the only ones footing the multi-hundred-billion-dollar bill.
The consumer is increasingly facing AI-soaked subscription tiers as tech firms attempt to monetize their huge investments and bundle their software offerings with difficult-to-separate-out AI tools that can make it tough for customers to opt out — and more expensive if they don’t.
One example is Microsoft 365, which now includes Copilot AI features in many of its tiers. The company recently introduced Microsoft 365 Premium at $19.99 per month, which bundles Copilot Pro features with Microsoft 365. Previously, Copilot Pro cost $20/month on top of existing subscriptions, and to use it in desktop Office apps, customers also needed a separate Microsoft 365 Personal ($6.99/month) or Family ($9.99/month) plan — bringing the total to roughly $27–$30/month. The base Microsoft 365 subscription itself has become increasingly essential as Microsoft de-emphasizes standalone Office purchases and makes cloud integration more central to document workflows.
Similar bundling of AI tools is becoming the norm from Alphabet to Adobe.
For instance, in March 2025, Google Workspace added its Gemini AI assistant into Business and Enterprise plans with price increases of about $2–$4 per user per month — roughly a 16%–33% jump depending on the tier — and with AI features that, in most cases, can’t be removed or opted out of. For a 50-person company on Business Plus, that means an additional $2,400 annually.
Adobe rebranded Creative Cloud All Apps to Creative Cloud Pro starting mid-2025, with prices increasing from $59.99 to $69.99 per month (or $659.88 to $779.99 annually) — a $10-per-month hike linked to expanded generative AI capabilities such as unlimited standard image and vector generation.
Whether the customer wants the AI or not isn’t really the point, according to experts — it’s the cachet that costs.
“AI is all the rage right now and that buzz fuels what marketers call perceived value bias,” said Elizabeth Parkins, professor of practice at Roanoke College. “When something’s labeled ‘AI-powered,’ people assume it must be smarter or more useful, even if it barely changes their experience. That sense of progress makes the extra subscription feel justified — until consumers start asking whether they’re paying for innovation or just the illusion of it,” she added.
Microsoft, Adobe, and Google did not respond to requests for comment.
Fred Hicks, assistant vice president and chief information officer at Adelphi University, said the companies are adding the extra charges to help pay for multi-billion-dollar data centers and their insatiable appetite for energy.
“The cost of running GPU clusters and power consumption is so high that baking it into subscriptions is how they can recoup costs. We have seen software licensing turn into subscription models,” Hicks said, citing Microsoft and Adobe Creative Cloud as examples of the approach that was adopted before the gen AI boom. “This creates a funding model of constant income over a single-cost perpetual license. AI subscriptions follow the same philosophy,” he said.
Personalized AI can pay off over time
Hicks says that AI baked into everything is going to be ubiquitous at some point in the near future because firms that do not have it will lose an edge in the market. For consumers, paying the price may pay off over time due to the personalization that can be fine-tuned by AI if it is in your life on a regular basis.
“Personalization using the same AI model trains on your habits and preferences. It will become more accurate in personalizing the user’s needs. This requires a long-term engagement and subscription,” Hicks said.
But over-subscription will become an issue, like it already is with streaming services, prompting consumers to review what they really need and purge at least some subscriptions for cost savings. That may be easier said than done though when it comes to software.
“Debundling AI subscriptions from other services will almost be impossible. Google and Microsoft now include basic AI with many of their application subscriptions. Higher tiers are required for deeper integration, increasing costs,” Hicks said.
Chris Sorensen, CEO of PhoneBurner, a U.S.-based SaaS company, says that a quiet but significant shift is underway.
“AI itself isn’t only improving products but redefining how pricing structures work. Companies like Adobe, Microsoft, and Google are using AI ‘enhancements’ to justify recurring revenue where one-time licenses used to suffice,” Sorensen said, adding that subscription models make sense as they create predictable income but do hide incremental costs.
Consumer pushback is emerging
“Many consumers are starting to notice this shift. After a while you start to notice that you are paying $10 here and $20 there for features not being used and not actively opted into,” Sorensen said — and that extra revenue which may benefit companies for now could be in for more pushback in the future.
“Some pushback is emerging, particularly in creative and productivity communities, but I do believe this model is only going to grow,” Sorensen said. He thinks what is likely to happen is that companies will build up “AI premium intelligence” tiers which will eventually turn software ownership into perpetual rental.
Tien Tzuo, founder and CEO of Zuora, an enterprise software company that provides a platform for businesses to launch, manage, and monetize subscription-based services, says that AI-infused products and price hikes are an increasingly vexing problem for consumers.
“All companies are layering AI into their products, but it’s the largest companies like Adobe, Microsoft, and Google who are often hiking prices without clear justification. Other companies like Zendesk are taking a more transparent and customer-friendly approach by correlating AI pricing to outcomes, so you only pay when an AI resolves a ticket,” Tzuo said.
If consumers balk enough at paying more, especially if use cases prove to be underwhelming to many, there will be a future where AI is pay-as-you-go, Tzuo said.
“We’re seeing an explosion of interest in usage-based pricing for AI, where customers have control over what they pay for based on what they consume,” he said. “AI is shifting what that ‘value’ means, and paying based on usage helps companies prove it. How often you use a product or see a result should speak for itself,” Tzuo added.
The increased AI bundling is simply an extension of what has been happening online for years, according to Ananya Sen, assistant professor of information technology and management at Carnegie Mellon University’s Heinz College. “The issue is how you can subscribe in one click, but to unsubscribe you have to make a phone call. In some sense what we are seeing with AI products is a continuation of that but maybe at a greater scale,” Sen said.
Since many people don’t understand AI products as well as existing software tools, they may not know what they are opting into, but one irony is that subscriptions today may end up being more secure than is in the best interest of consumers. That’s as a result of behavioral psychology, he says.
“There is an inertia once you opt in — it’s harder to opt out, you have to make an active choice. Companies are banking on and exploiting this,” Sen said. “And when it comes to AI products, it is a fast-evolving space. It is hard for a normal online consumer who uses these different tools to keep track — it becomes a bandwidth issue, your mental bandwidth and attention,” he added.
Sen says consumers need to be active in their own subscription ecosystem. “There has to be some responsibility on the consumer side. But it is a two-way street. These small dollar values add up,” Sen said.
Many consumers still have one advantage: they remain on free basic versions of software. But companies will do what they can to migrate more to subscription products. “Even when you think of the big players, a large proportion of the users use the basic free version. Even for the most prominent players, it is hard to get people to convert to a subscription,” Sen said.
Reddit CEO Steve Huffman stands on the floor of the New York Stock Exchange (NYSE) after ringing a bell on the floor setting the share price at $47 in its initial public offering (IPO) on March 21, 2024 in New York City.
Spencer Platt | Getty Images
Reddit‘s stock popped more than 12% Friday after the company surpassed third-quarter estimates and signaled strong advertising growth.
The social media platform’s revenues surged 68% from a year ago to $585 million, beating an LSEG estimate of $546 million. Earnings per share totaled 80 cents, surpassing an estimate of 51 cents.
Reddit also released a better-than-expected sales outlook for the fourth quarter. The company projects between $655 million and $665 million, topping the $638 million forecast from Wall Street.
These results “speak to the company’s continued progress across its ad and platform initiatives,” wrote Morgan Stanley analyst Brian Nowak. “We see a long runway for growth across both active advertisers (+75% y/y in 3Q) as well as greater penetration within existing advertisers.”
Reddit said that nine of its top 15 advertiser verticals grew more than 50%. Nowak highlighted Reddit’s ongoing investments in automation, which are improving return on advertising spending.
Read more CNBC tech news
The number of daily active users jumped 19% from the year-ago period to 116 million, surpassing Wall Street’s 114 million estimate.
Reddit has attracted more people to the platform from Google, and generated advertising dollars from those who create accounts. The company makes more money off of logged-in users, and has raised concerns as of late that AI chat apps, including ChatGPT and Google’s AI Overview could impact new user growth.
“I’m looking forward to continuing to work on these things with these partners, but they’re not a major traffic driver today,” CEO Steve Huffman said during the earnings call. “But I think there’s plenty of opportunity there as we continue to work together.”
The company’s daily active unique users rose 7% from last year to 23.1 million, but lagged the 12% growth seen in the second quarter. Globally, logged-in users grew 14% to 50.2 million.
Reddit’s other revenues, which include data licensing partnerships with Google and OpenAI, grew 7% from a year ago to $36 million.