TSMC and ServiceNow’s AI Moats Are Ticking—But Valuations Leave No Room for Error

The best investment thesis starts with the simple question: Is the product people actually want? For the AI infrastructure story, the answer is a clear yes. The demand isn't just on paper; it's in the numbers and the real-world buildout.

Look at the scale. The global semiconductor industry is projected to hit US$975 billion in annual sales in 2026, a historic peak. And the engine is unmistakable: generative AI chips will approach US$500 billion in revenue in 2026, or roughly half of global chip sales. That's not a rumor; it's the financial bedrock of the boom. This isn't about hype-it's about factories running flat out to meet orders.

The proof is in the performance of the key players. TSMC, the world's largest chipmaker, has seen its earnings grow 51% in 2025. Its stock has rewarded that growth with a 104% gain over the past year. For ServiceNow, the enterprise software giant, the story is similar. Last quarter, its subscription revenue grew 21% year over year, and its current remaining performance obligations (cRPO) climbed 25%. These aren't accounting tricks; they're signs of customers locking in multi-year contracts for real business utility.

The bottom line is that the AI infrastructure buildout is real and driving tangible growth. The parking lot at the data center is full, and the orders for the chips and software to run it are piling up. The stocks have already priced in a lot of that optimism. The common-sense takeaway is that the demand is there, but the risk now is that the prices reflect perfection, leaving little room for error.

Kick-the-Tires: Product Quality and Brand Loyalty

The real test of a stock isn't just its growth rate; it's whether the company's product is so good and so entrenched that customers can't easily walk away. For both TSMC and ServiceNow, the evidence points to powerful moats, but they are not built of stone-they are made of technology and trust, and those can erode.

Take TSMC. The company isn't just a leader; it's become a virtual monopoly for manufacturing advanced chips. This isn't a title it earned through marketing. It's the result of years of relentless R&D and the sheer scale of its fabs, which have allowed it to achieve the high yields needed for cutting-edge nodes while competitors struggled. That technological edge translates directly to pricing power. When the world needs the most advanced chips for AI, TSMC is often the only game in town. This isn't just a business advantage; it's a fundamental bottleneck in the supply chain. The moat here is deep and wide, built on physics and engineering.

ServiceNow's moat is different. It's built on the sheer complexity of its platform and the cost of replacing it. The company positions itself as the AI control tower for business reinvention, the central nervous system that unifies data and workflow. For an enterprise, ripping out ServiceNow isn't like switching email providers. It's a multi-year, multi-million-dollar project involving years of custom logic, audit trails, and security protocols. As the evidence notes, the product is not something that can be easily ripped out and replaced. This creates immense customer stickiness, turning subscriptions into a reliable, long-term revenue stream.

Yet, even the strongest moats face pressure. The biggest threat to ServiceNow is the rise of AI-native startups. A recent report highlighted a rapid funding round for Serval, a Sequoia-backed startup, valuing it at $1 billion. The signal is clear: investors see a gap. These new entrants are betting that AI-native architecture can replace legacy ITSM systems. The challenge for ServiceNow is real. As one observer notes, legacy ITSM workflows are dying, and the company must prove its AI offering, Now Assist, is not just an add-on but the future. The startup threat introduces a new variable: the cost and risk of migration. While incumbents have the edge in security and compliance, the startups must demonstrate they can deliver a better employee experience and a clear ROI to justify the upheaval.

The bottom line is that both companies have formidable, product-driven advantages. TSMC's monopoly on advanced manufacturing is a physical moat. ServiceNow's platform lock-in is a digital one. But neither is invincible. The growth for the next two years depends entirely on their ability to keep kicking the tires on their own products. TSMC must maintain its technological lead, while ServiceNow must aggressively self-disrupt to stay ahead of the AI-native challengers. The moats are strong, but they require constant maintenance.

Valuation and the Margin of Safety

The growth stories are clear. The moats are deep. But the real question for any investor is whether the price paid today leaves room for the inevitable bumps in the road. After a strong run, the valuations of these two stocks are now a tight fit against their futures.

For TSMC, the math is straightforward. Wall Street expects the chipmaker to produce 21% revenue growth next year. That's impressive for a company of its size. Yet the stock trades at a discount, valued at the low end of where the big tech companies normally trade. In other words, it's being priced like a solid, reliable engine rather than a rocket ship. This offers a margin of safety. The risk is that the AI boom, which is the entire engine, slows. The industry is heavily reliant on that spending, and if demand corrects, the growth story could falter quickly.

ServiceNow presents a different picture. The stock has been through a brutal year, with its price declining 50% over the past year. That's a massive haircut for any company. Yet even after that drop, its valuation remains elevated. The stock now trades at a P/E multiple of 24.2x, which still represents a premium for its growth rate. The market is saying it believes in the AI transformation, but it's not giving the company a cheap ticket. The risk here is execution. The company must prove that its AI platform, Now Assist, can drive the next leg of growth and fend off the new competitors. Any stumble in that rollout would pressure the premium valuation.

The bottom line is that valuations are reasonable, but they leave no room for error. For TSMC, the discount to peers is a buffer, but the entire bet is on AI demand staying hot. For ServiceNow, the steep decline has reset expectations, but the current multiple demands flawless execution to justify it. In both cases, the growth stories must play out perfectly. There's no margin of safety for a misstep.

Risks and What to Watch

The investment thesis for these two stocks is built on strong fundamentals, but the real test is what happens next. The best way to judge the setup is to watch for specific, tangible signs of trouble or strength. For TSMC, the immediate focus is on execution. The company just posted one of its strongest quarters in history, but the question is whether it can keep that momentum. Investors should watch for continued record quarterly results, particularly on the margin front. Any signs of capacity constraints or yield issues in its advanced fabs could start to pressure those expanding profits. The moat is wide, but even a physical bottleneck can slow a train.

For ServiceNow, the watchpoint is all about adoption. The company has a large pipeline of big deals, but the real proof is in the conversion. The market needs to see that its AI product, Now Assist, is being widely adopted by customers and is driving higher-margin, sticky revenue. The threat from startups like Serval, which just raised a $1 billion valuation, is a constant reminder that the company must prove its AI offering is not just an add-on but the future. The bottom line is whether the product can convert its deal strength into the kind of reliable, long-term income that justifies its premium.

The key watchpoint for both companies, and for the entire AI infrastructure story, is the health of the broader spending cycle. The semiconductor industry is navigating a high-stakes paradox, with soaring demand but also a risk of correction. If AI spending slows, it could quickly turn the current boom into a bust for both chipmakers and the software that runs on them. The stock market has already priced in a lot of optimism, so any stumble in the cycle would be a major headwind.

The bottom line is that investors must stay vigilant for these specific signals. The growth stories are real, but they are not guaranteed. The next two years will be defined by quarterly results, product adoption rates, and the overall temperature of the AI budget. Keep your eyes on those numbers.

The Takeaway

The bottom line is this: both TSMC and ServiceNow have passed the basic common-sense test. Their products are in real-world demand, and their brands have built powerful, sticky moats. You can kick the tires on both and find solid engineering and entrenched software. But here's the catch-their valuations now leave no room for a stumble. The market has priced in a perfect AI boom, so any hiccup in execution or a slowdown in spending would be punished.

The decision comes down to patience. Can you hold through the volatility while waiting for the AI growth story to fully play out? For TSMC, the answer hinges on its ability to keep manufacturing the world's most advanced chips without a glitch. For ServiceNow, it's about proving its AI platform, Now Assist, converts its massive pipeline into reliable, high-margin revenue. This isn't a "set and forget" investment. It's a bet on a multi-year buildout that requires watching quarterly results and product adoption closely.

For a "keep it simple" approach, these are two of the clearest plays on the AI boom. TSMC is the indispensable factory; ServiceNow is the essential software. But simplicity doesn't mean safety. The setup demands a long-term view and nerves of steel. If you can stomach the volatility and believe the AI infrastructure story will hold, these are stocks to hold. If you're looking for a margin of safety, you might want to wait for a clearer signal of a slowdown before buying.