Dan Gallagher
Investors remain as fixated as ever on reading the tea leaves for signs of AI demand. They should look carefully at the companies that face the highest stakes if they get it wrong.
Taiwan Semiconductor Manufacturing is one of those. The company better known as TSMC is the world's largest producer of the type of advanced semiconductors that power everything from AI servers to smartphones to personal computers.
TSMC's business is lucrative, but also expensive to operate. Proof of that: While its customer list includes the world's richest companies ( Nvidia, Apple and Google), they have studiously avoided getting into chip manufacturing themselves.
How expensive? TSMC used its last earnings report three months ago to project capital expenditures between $52 billion and $56 billion this year. That was well above the $44 billion Wall Street had been expecting and more than 60% higher than what TSMC's capex bill has averaged over the past five years.
Runaway capital spending is hardly unique in the AI race. But if companies like Amazon, Microsoft and Google overshoot a bit in their data-center build-outs, they at least have the comfort of knowing they eventually can make use of that capacity.
Underused chip-production facilities have high fixed costs and can immediately weigh on the bottom line. "If we didn't do it carefully, that would be a big disaster to TSMC for sure," its chief, CC Wei, said in the company's last earnings call.
TSMC is widely expected to at least maintain that capex target in its first-quarter report, due early Thursday morning New York time. And that spending hasn't been worrisome to investors so far. TSMC's stock price is up 25% this year -- far ahead of Nvidia's 4% gain.
Like Nvidia, TSMC's position as a key player in the AI build-out has conveyed unmatched pricing power that is boosting its profits. Analysts expect TSMC to report gross-profit margins of 65% for the first quarter. This would be their highest level in more than 20 years, according to data from S&P Global Market Intelligence. And that is even with weakness in other markets that TSMC serves such as smartphones.
Robust AI demand isn't likely to last forever. The chip business is notoriously cyclical. But those cycles are why TSMC's investment plans are such an important early indicator. Unused chip-production tools make for expensive paperweights.
This is an edition of the WSJ AI & Business newsletter, a weekly digest to help you make sense of AI's impact on business with news, insights and data from our global team of technology journalists. If you're not subscribed, sign up here.
We're Using So Much AI That Computing Firepower is Running Out
The artificial-intelligence gold rush is rapidly drying up the supply of the one resource that AI developers can't do without: computing power. The capacity crunch has caused consternation among power users, forced companies to scuttle products and led to reliability problems. Companies have been scrambling to secure the availability of computing capacity needed to serve a growing base of customers who are also significantly increasing their AI use.
The Number
Net-advertising revenue Emarketer expects Meta Platforms to generate this year, exceeding what the market-research firm expects for Google.
What the Humans Are Saying
AI in Charts
Meta Platforms' big AI moment is overshadowing its "big-tobacco" moment. The release of its new Muse Spark AI model helped the company's stock fully recover from a selloff following a pair of surprise jury verdicts in March. Those cases found the company responsible for harms done to users of its social networks. This brought new uncertainty to the business that is paying for Meta's expensive AI makeover.
AI in the Wild
A software bug that was capable of crashing an operating system used by firewalls, servers and network appliances went undetected for over 27 years. Last month, it was caught by Mythos, the latest AI model from Anthropic that has spooked the White House, banking executives and cybersecurity professionals around the world.
Welcome to the bug armageddon. AI models like Mythos and others are finding bugs in older software at a rate never seen before. While most of the coding issues may be minor, their sheer volume has amplified the risk that smaller software developers will become overwhelmed with reports of bugs such as the one Mythos found.
Other Highlights From the Week in AI
-- CoreWeave, Anthropic form an AI cloud agreement. -- Sam Altman attack suspect charged with attempted murder. -- Anthropic adds Novartis CEO to board of directors. -- How to switch AI chatbots -- and why you might want to.
About Us
WSJ AI & Business is a weekly look at AI's transformation of the business world. This newsletter was curated and edited by Dan Gallagher and Asa Fitch. Reach them at dan.gallagher@wsj.com and asa.fitch@wsj.com (if you're reading this in your inbox, you can just hit reply). Got a tip for us? Here's how to submit.
(END) Dow Jones Newswires
April 14, 2026 12:22 ET (16:22 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments