Taiwan Semiconductor Manufacturing Company (NYSE: TSMC) just announced capital expenditure of US$52 billion to US$56 billion for 2026, an increase of between 27% and 37% compared to 2025.
That’s great news for ASML Holding (NASDAQ: ASML), the Dutch company that sells the lithography machines TSMC needs for advanced chip production.
But here’s the thing: the investment doesn’t stop there.
Every dollar TSMC spends on fab equipment eventually flows through to cloud providers like Microsoft (NASDAQ: MSFT) and application companies like Meta Platforms (NASDAQ: META).
With all three companies reporting earnings tonight, investors have a rare opportunity to trace the AI money trail from silicon to software.
ASML: The Picks and Shovels
In any gold rush, it pays to sell the shovels.
ASML is the only company in the world that makes extreme ultraviolet (EUV) lithography machines — the US$400-million systems required to manufacture the most advanced chips.
Without ASML, there is no AI hardware revolution.
The numbers back this up.
For 2025’s third quarter (3Q2025), ASML reported net system bookings of €5.4 billion, marking the second consecutive quarter above €5 billion.
Of this, €3.6 billion came from EUV systems alone.
TSMC’s expansion plans provide further validation.
The Taiwanese chipmaker plans to build four new advanced packaging plants this year — two in Chiayi Science Park and two in Southern Taiwan Science Park.
More importantly, TSMC started mass production of its 2-nanometre process in 4Q2025, with plans to build new 2nm fabs in Hsinchu and Kaohsiung.
All of this requires ASML’s machines.
That said, there is one dark cloud on the horizon: China.
The country contributed 36% of ASML’s revenue in 2024, and management expects this to decline significantly in 2026 due to export restrictions.
But here’s the kicker: ASML’s management stated that 2026 total net revenue will not fall below 2025 levels — a remarkable statement given China’s outsized contribution.
TSMC’s capex increase, with 60% to 80% allocated to advanced process development, helps explain that confidence.
Microsoft: The Infrastructure Layer
If ASML makes the machines that make the chips, Microsoft runs the cloud infrastructure where those chips do their work.
For the first quarter of fiscal year 2026, Azure and other cloud services revenue grew 40%, driven by demand for AI services.
Microsoft now serves 80,000 Azure AI customers, including 80% of the Fortune 500.
CEO Satya Nadella isn’t betting everything on one AI model, either.
Azure AI Foundry now offers access to over 11,000 models — including OpenAI’s GPT-5, Anthropic’s Claude, and Google’s Gemini.
The message to enterprises is clear: build on Azure, and you’re never locked into a single provider.
This hedging strategy is working. Microsoft recently committed to investing up to US$15-billion in Anthropic alongside NVIDIA (NASDAQ: NVDA), signalling that it’s prepared to win regardless of which AI model ultimately dominates.
The key question for upcoming earnings: can Azure maintain its growth momentum as capacity constraints ease?
Meta: The Application Layer
At the end of the AI supply chain sits Meta, turning all that infrastructure into advertising dollars.
Meta’s AI-powered ad tools now generate a US$60-billion annual run-rate.
That’s not a projection — that’s real money flowing through AI systems today.
For 3Q2025, Family of Apps (Facebook, Instagram, Whatsapp) ad revenue surpassed US$50 billion, up 26% year-over-year.
CEO Mark Zuckerberg recently described three camps in the AI industry: optimists who see superintelligence emerging within two to three years, moderates who expect breakthroughs by the decade’s end, and pessimists who think it will take well into the 2030s.
Zuckerberg? He’s investing for the most optimistic scenario.
Unlike unprofitable AI startups that risk bankruptcy if they misjudge timing, Meta’s strong cash flow provides a cushion.
The downside of overbuilding — slowing down to let demand catch up — is acceptable in Zuckerberg’s eyes.
Watch for updates on Reels monetisation (now at US$50 billion run-rate) and the rollout of Business AI, which aims to turn every small business into an AI-powered selling machine.
Get Smart: Follow the Money
Here’s something everyone should remember: TSMC’s US$52-56 billion capex projection is a signal based on customer orders.
Those customers include the hyperscalers building AI infrastructure.
And those hyperscalers are seeing real demand from enterprises and consumers.
The AI supply chain is only as strong as its weakest link. ASML’s bookings, Microsoft’s cloud growth, and Meta’s ad revenue all need to deliver for the investment thesis to hold.
So far, the receipts are there.
But as always, the future remains uncertain.
Generative AI is reshaping the stock market, but not in the way most investors think. It’s not just about which companies are using AI. It’s about how they’re using it to unlock new revenue, dominate their markets, and quietly reshape the business world. Our latest FREE report “How GenAI is Reshaping the Stock Market” breaks the hype down, so you can invest with greater clarity and confidence. Click here to download your copy today.
Follow us on Facebook, Instagram and Telegram for the latest investing news and analyses!
After all, where the business goes, eventually the stock will follow.
