Morris Chang doesn't usually mince words, but his recent take on Intel is something else. The legendary founder of Taiwan Semiconductor Manufacturing Co. (TSMC) basically said out loud what half of Wall Street has been whispering for years. Intel is in a mess. Not just a "bad quarter" kind of mess, but a deep, structural identity crisis.
During a press conference in late 2024 for the launch of his second autobiography volume, Chang took a direct swing at the American chip giant. He pointed out that TSMC founder criticizes Intel business strategy for one glaring reason: they’re trying to do everything and succeeding at almost nothing right now. Honestly, when the guy who invented the modern foundry model tells you your strategy is broken, you probably should’ve listened three years ago.
🔗 Read more: Mark Cuban Explained (Simply): Why the Healthcare System is Broken
The Pat Gelsinger Era: Eloquence vs. Execution
The big news recently was the sudden departure of Pat Gelsinger. He was supposed to be the savior. The guy came in with a "five nodes in four years" plan that sounded great on a PowerPoint slide. Chang noted that Gelsinger's eloquence is what got him the job in the first place. He could talk the talk. But the strategy he pushed—restarting Intel’s foundry business to make chips for everyone else while still designing their own—has left them stuck.
Intel is currently without a permanent CEO and, more importantly, without a clear direction. Chang’s critique was sharp: Gelsinger seemed to care more about becoming a foundry than winning the AI race. Think about that for a second. While Nvidia was becoming the most valuable company on the planet by focusing on AI, Intel was busy pouring billions into factories to compete with TSMC's decades-old lead. It's a classic case of chasing the wrong rabbit.
Why the Foundry Model is Harder Than It Looks
Most people think building a chip factory is just about buying expensive machines from ASML. It isn't. It’s a service business. Chang built TSMC on the idea of being a partner, not a competitor.
- Conflict of Interest: Intel wants to make chips for Apple or Nvidia. But Intel also competes with them. Would you give your secret blueprints to a guy who’s trying to put you out of business? Probably not.
- The Yield Problem: TSMC has mastered the art of "yield"—getting the most working chips out of a single silicon wafer. Intel has struggled here for years. Even their upcoming 18A node, which they've hyped as a TSMC-killer, is facing skeptical looks from big customers.
- Culture Shock: Moving from an Integrated Device Manufacturer (IDM) to a foundry is like a restaurant that only cooks its own recipes suddenly trying to become a ghost kitchen for 50 different chefs. It’s a total mess of logistics and ego.
The AI Opportunity Intel Let Slip
You've gotta wonder how Intel missed the AI boom this badly. Chang mentioned that Intel basically stood on the sidelines. Instead of pivoting hard into AI accelerators or training chips that could rival Nvidia, they stayed obsessed with their foundry ambitions.
💡 You might also like: Vanguard Admiral Shares: What Most People Get Wrong About These Low-Cost Funds
Intel's Gaudi AI chips haven't exactly set the world on fire. In fact, they recently admitted they won't even hit their modest sales targets for the Gaudi 3. When you compare that to the billions flowing into TSMC because every AI player has to use them, the contrast is staggering. Intel isn't just losing; they're becoming irrelevant in the only sector that currently matters for growth.
A "Hostile" Relationship and Revoked Discounts
There’s some spicy history here, too. Chang didn't hold back on his personal feelings toward Gelsinger, calling him "hostile" and "very discourteous." This isn't just old-man grumpiness. It had real business consequences.
A few years ago, Gelsinger went on a press tour suggesting that Taiwan wasn't a "stable place" for chipmaking, trying to lobby for U.S. government subsidies. TSMC didn't take that well. They reportedly yanked a massive 40% discount Intel was getting on 3nm wafers.
"TSMC deals with Intel the way Intel deals with TSMC," Chang said. Basically, if you want to play dirty in the press, don't expect a "friend" price on the most advanced tech in the world. That loss of a discount hit Intel's margins hard right when they were already bleeding cash.
Can Intel Actually Turn it Around?
The road to 2026 looks brutal for Intel. They've received billions in CHIPS Act money, but government cash can't buy a working business strategy. Chang's advice was simple but tough: the board needs to figure out what the company is before they hire another CEO.
✨ Don't miss: USD vs Polish Zloty: What Most People Get Wrong About the Exchange Rate
If they keep trying to be a top-tier chip designer and a world-class foundry, they might end up being neither. There is serious talk now about a joint venture or even spinning off the manufacturing arm entirely. It’s a "better the devil you know" situation, but the status quo is clearly a sinking ship.
Actionable Insights for the Industry
If you're watching the semiconductor space, here's what actually matters moving forward:
- Watch the 18A Node: This is Intel's last stand. If they can't get a "whale" customer (like Apple or Nvidia) to commit to 18A by the end of 2026, the foundry dream is likely dead.
- AI Pivot: Whoever takes the CEO seat needs to prioritize AI over everything else. The PC market is flat; the data center is shifting to GPUs and custom silicon.
- Trust over Tech: Intel needs to prove they can be a neutral partner. They might need to legally separate their factory business from their design business to ever win over competitors.
- Yield is King: Don't listen to marketing about "nanometers." Look at the yield rates. If they can't produce chips efficiently, they can't compete on price, and TSMC will keep its 90% market share of advanced nodes.
The semiconductor world doesn't have much room for second place. As Morris Chang pointed out, Intel is currently in the same predicament it was in four years ago—no leader, no clear plan, and a very fast-moving target in the rearview mirror.