That's when TSMC's Q1 2026 revenue confirmation landed, NT$1,134.10 billion, roughly $35.71 billion, the strongest March on record, sitting at the very top of its own guidance range and beating Bloomberg consensus. The market shrugged. The stock ticked up, then gave it back. And by Monday morning, the conversation had moved on to tariff noise and Fed speculation.

That's the part that bothers me.

Because what got buried in that Friday data dump wasn't a routine quarterly beat. It was more evidence that one company, a single foundry operating on a 36,000 square kilometre island, now sits at the centre of every technology that will define the next twenty years.

Here's the question nobody is asking out loud: if semiconductors are now the raw material that every government, defence contractor, and AI lab on earth cannot function without then TSMC isn't a tech stock. It's a toll booth on the most important road in the global economy.

Today we're going to talk about what that actually means for investors who get it now, versus the ones who'll figure it out later.

Berkshire-owned PacifiCorp won an Oregon appeals ruling that could sharply reduce its wildfire liability by undoing the class-action setup used in earlier cases. The company has already paid billions, but this decision could lower future damages materially, which is why the market sees it as a meaningful financial win for Berkshire.

The Fool lays out a familiar Palantir split: the bull case is its strong AI positioning, sticky government relationships, and growing commercial runway, while the bear case is that the stock is still extremely expensive and exposed to political, competitive, and demand-cycle risks. The takeaway is not that Palantir is broken. It is that a great business can still be a tough buy when expectations are already sky-high.

This MarketWatch piece appears aimed at investors nearing retirement who are feeling the stress of war-driven volatility. The central message is not to panic over headlines, but to lean on a sound financial plan and make sure the portfolio is built to handle shocks rather than react emotionally to them.

Last month, 48% of visitors to documentation sites across Mintlify were AI agents—not humans.

Claude Code, Cursor, and other coding agents are becoming the actual customers reading your docs. And they read everything.

This changes what good documentation means. Humans skim and forgive gaps. Agents methodically check every endpoint, read every guide, and compare you against alternatives with zero fatigue.

Your docs aren't just helping users anymore—they're your product's first interview with the machines deciding whether to recommend you.

That means:
→ Clear schema markup so agents can parse your content
→ Real benchmarks, not marketing fluff
→ Open endpoints agents can actually test
→ Honest comparisons that emphasize strengths without hype

In the agentic world, documentation becomes 10x more important. Companies that make their products machine-understandable will win distribution through AI.

Make Your Docs Agent-Ready

There used to be a simple framework for understanding geopolitical power. Find the resource every modern economy needed to function. Figure out who controlled it. That told you everything about leverage, alliances, and where the next conflict would happen.

For most of the twentieth century, that resource was oil. Nations bent foreign policy around it. Navies patrolled shipping lanes because of it. Governments rose and fell on access to it. The companies that sat at the chokepoints of extraction and refining didn't just profit from energy demand, they held structural leverage over entire economies.

Semiconductors are following the same script. Except the geology is more concentrated, the manufacturing knowledge took five decades to build, and it cannot be replicated in a decade regardless of how many billions a government throws at the problem. TSMC isn't just a chip manufacturer. It's the Saudi Aramco of the AI era.

The same concentration that makes TSMC the world's most powerful technology chokepoint also makes it the single largest geopolitical risk in any portfolio that holds it. That tension is the whole story, and every section below is about understanding it clearly enough to act on it with conviction.

The AI demand headline numbers get all the attention. These are the figures that actually tell you what's happening.

Q1 2025 revenue, just twelve months ago, was $25.53 billion, with net income up 60.3% year-over-year. Wind forward to Q1 2026: revenue hit $35.6 billion, up 35% year-on-year, beating analyst forecasts despite ongoing geopolitical tensions and trade uncertainty. That's not a company riding a wave. That's a company accelerating into one.

Then there's the profitability picture, which is where this gets genuinely remarkable.

Q4 2025 - TSMC's most recent full quarterly filing showed gross margin of 62.3% and operating margin of 54%.

Then there's the net profit margin: 48.3%.

Sit with that number for a second.

Nearly half of every dollar TSMC earns in revenue becomes profit. That is not the margin profile of a manufacturer. That is the margin profile of a company with no real competition at the frontier. For context: Apple's net margin runs around 26%. ExxonMobil, the benchmark for resource monopoly dominance, typically operates in the 10–15% range.

TSMC controls 72% of the global semiconductor foundry market. For advanced nodes sub-7nm, where the chips powering AI infrastructure, defence systems, and next-generation smartphones actually get made, that share climbs north of 90% by most industry estimates.

For full-year 2026, TSMC is guiding for close to 30% revenue growth in US dollar terms, with Q1 2026 gross margin guided at 63–65% and operating margin at 54–56%.

These are not cyclical numbers. These are the economics of structural monopoly.

A foundry is a factory that manufactures chips designed by other companies. TSMC doesn't design what's inside your iPhone or your Nvidia GPU, Apple and Nvidia do that. TSMC builds it. And at the most advanced level of chip manufacturing, the frontier where performance and power efficiency actually matter, TSMC is the only factory on earth that can do it at scale.

That's not a competitive advantage. That's a different category of thing.

When you are the sole manufacturer of something the entire global technology stack depends on, you don't negotiate on price, you set it. And when you raise it, your customers have nowhere else to go. Apple pays. Nvidia pays. AMD pays. The cost gets passed downstream. Which is exactly what the margin data above is showing you in real time.

Here's the part most headlines skip entirely.

TSMC unveiled record capex of $52–56 billion for 2026, up roughly 30% from the $40.9 billion spent in 2025. And that's before you account for the Arizona buildout in full.

TSMC has committed $165 billion to its US expansion: six fabrication plants, two advanced packaging facilities, and a research and development centre spanning a 2,000-acre campus outside Phoenix. For scale, $165 billion is larger than the entire annual GDP of Hungary, and larger than the combined annual capex of Intel, Samsung, and GlobalFoundries combined. This is the largest foreign direct investment in American history.

That is not a capex cycle. That is a generational infrastructure bet.

The complication: building in the United States costs four to five times what an identical plant costs in Taiwan, due to higher labour costs, regulatory requirements, and the sheer challenge of standing up a semiconductor workforce from scratch in a geography where that knowledge base doesn't yet exist. That cost differential doesn't disappear into TSMC's margins. It gets priced into the wafer. Which gets priced into the chip. Which eventually gets priced into your iPhone.

The fear isn't that AI demand collapses. The fear is sequencing.

The bill for this infrastructure buildout arrives years before the US fabs generate meaningful returns. Between $165 billion in Arizona commitments and annual capex running at $52–56 billion, TSMC faces a multi-year window where spending peaks, depreciation charges accumulate, and US construction costs keep compounding, all before a single advanced chip rolls off a Phoenix production line at scale. If margins soften during that window, and the stock is already trading near all-time highs, there's no obvious floor while investors wait for the returns to catch up.

Semiconductor fabs typically take three to five years from groundbreaking to full production. TSMC broke ground in Arizona in 2021. The math on when those assets start contributing meaningfully to free cash flow is not comforting if you're looking at a three-year time horizon.

This is the reason the April 16 call matters more than the revenue number that's already been reported. Revenue is backward-looking. Margin trajectory and capex revision language is what tells you whether the sequencing risk is getting better or worse.

TSM closed Friday at $370.60, up from a prior close of $365.49, with a 52-week range of $145.84 to $390.21. The stock has nearly tripled off its 52-week low, and yet it hasn't broken out to new all-time highs.

That divergence is informative. The market is acknowledging the fundamental story while staying cautious ahead of the full April 16 earnings call and its margin detail. This isn't distribution. It's a stock waiting for permission to move.

Per Bloomberg consensus data, the average 12-month analyst price target sits at $437.87, with a high estimate of $600. Eighteen analysts currently rate TSM a Strong Buy. Zero rate it a Sell.

On the chart, TSM is consolidating in the upper half of its 52-week range, holding above the $360 level that previously acted as resistance and has since flipped to support. The line to watch is $390. A clean weekly close above that level, ideally on volume, is the signal that the next leg higher has started. Below $360 with momentum, and the correction that the sequencing risk suggests becomes a real conversation.

The April 16 call is the catalyst in either direction. If gross margins come in at the top of the 63–65% guided range and capex language stays controlled, the bull case accelerates hard. If management flags any softening in customer ordering patterns or revises capex upward again, the move down will be fast.

Revenue is already confirmed. What moves TSM and every semiconductor name tied to it is Q2 guidance, updated margin commentary, and whether TSMC signals another capex revision upward. Listen specifically for how management characterises Arizona cost progress. Any language suggesting construction costs are running above internal projections is the tell.

Equipment installation at TSMC's second Arizona fab is targeted for Q3 2026, with high-volume 3nm production planned for 2027, accelerated by a full year from the original 2028 schedule. Whether that acceleration holds matters not just for TSMC's own P&L, but for Apple, Nvidia, and AMD, all of which have built supply chain timelines around those dates. Any further slippage has second-order effects across the entire semiconductor ecosystem.

More than 80% of TSMC's manufacturing capacity remains in Taiwan. Most serious investors already know this. What they don't always have is a framework for sizing what it actually means for a position.

Here's how I think about it: geopolitical risk of this kind is not a binary. It doesn't price in gradually and then spike on the day, it reprices in sudden, discontinuous jumps when the perceived probability of a bad outcome shifts. The practical implication is that TSM will trade at a persistent discount to what its fundamentals alone would justify, because the market can never fully trust that the discount is temporary. That's not a flaw in the thesis. It's the cost of admission.

The question isn't whether to hold a company with 80% of production on a contested island. The question is whether your position size reflects the possibility that a tail event even a low-probability one would mean not just a drawdown but an extended period of manufacturing disruption with no clear recovery timeline. If you wouldn't be able to hold through that scenario, your position is too large. Full stop.

The single largest bull assumption baked into TSM's current valuation is that hyperscaler AI spending doesn't slow. Watch what Nvidia, Microsoft, and Meta say in their own upcoming earnings calls, they are the leading indicators for TSMC's forward demand pipeline. If more efficient model architectures begin reducing raw compute requirements per dollar of output, TSMC's fastest-growing segment absorbs that directly.

The moat is real. The margins are real. The demand is real.

But here's the thing that the bull case tends to gloss over: a company with 90% advanced-node market share, 48% net margins, and zero credible competitors trades at roughly 22x forward earnings, a discount to the S&P 500 average. The market is not ignoring the fundamentals. It's pricing in the island.

That gap between what the business is worth on its own merits and what the geopolitical overhang implies is, depending on your view of cross-strait risk, either the most interesting entry point in large-cap tech or exactly the right discount.

My view: it's the former, conditionally. The condition is sizing. TSM at 3–5% of a diversified portfolio, held with a five-year minimum horizon and full awareness of the Arizona sequencing drag, is a compelling position. TSM chased at all-time highs with a short time horizon and no real framework for the geopolitical exposure is a different trade entirely, and one that will test you at exactly the wrong moment.

The April 16 call either gives the stock permission to move or it doesn't. Either way, the long-term thesis doesn't resolve on a single earnings print. Go in clear-eyed or don't go in.

Those numbers aren't a CX issue — they're a design issue. Gladly's 2026 Customer Expectations Report breaks down exactly where AI-powered service loses customers, and what the architecture of loyalty-driven CX actually looks like.

Get the report

The tell that TSMC hasn't fully arrived at consensus yet is in the multiple. A company with 90% advanced-node market share, 48% net margins, and zero credible competitors trades at roughly 22x forward earnings, a discount to the S&P 500 average. The market is not ignoring the fundamentals. It's pricing in the island.

That gap between what the business is worth on its own merits and what the geopolitical overhang implies is, depending on your view of cross-strait risk, either the most interesting entry point in large-cap tech, or exactly the right discount.

Knowing which one you believe is the actual work.