Nik_Maestro

Q4 was when Amazon asked investors for patience. Q1 was when it gave them evidence. The $200B AI capex reset looked frightening because free cash flow was going to disappear. The question was whether Amazon was building ahead of demand or into demand. Q1 points to the latter. AWS accelerated to 28%. Backlog surged to $364B. Trainium commitments reached $225B. The chip business crossed a $20B run-rate. But evidence is not cash. FCF fell to $1.2B, and Amazon raised $53B of long-term debt. That is the new Amazon debate. The Foundry has answered back. Now it has to pay for itself. nikhs.substack.com/p/amazon-1q26-…

The market still wants to analyze Alphabet as one company. That is the mistake. Alphabet is now four businesses running on one AI stack: 1. Search compounds intent 2. Cloud compounds compute demand 3. YouTube compounds attention 4. Waymo compounds autonomous miles That is why the 1Q26 print matters. Search grew 19%. Cloud grew 63% with 32.9% margins. Cloud backlog reached $462B. YouTube is layering subscriptions on top of living-room scale. Waymo crossed 500K autonomous rides per week. The old question was: will AI kill Google? The new question is much more interesting: can one AI infrastructure layer power four separate business-model upgrades? That is what I unpack in the full piece: Alphabet’s 1Q26: Four Markets, One Grid. nikhs.substack.com/p/alphabets-1q…

Meta’s Q1 was not a bad quarter. It was a hard-to-value quarter. The ad business is accelerating because of AI. The capex bill is exploding because of AI. Both are true. That is why the stock fell on 33% revenue growth. The real question is no longer “is AI helping Meta?” Q1 says yes. The question is whether Meta’s AI dividend from ads can outgrow the AI tax required to build a consumer agent platform. That is what I wrote about here. nikhs.substack.com/p/metas-1q26-e…

The Microsoft AI debate is being framed as: Great quarter vs scary capex. I think that misses the point. The real debate is whether Microsoft owns the next work surface. If enterprise context becomes the new OS, Microsoft wins. If the agent becomes the new OS, Microsoft may become essential middleware around someone else’s platform. That is a very different kind of victory. nikhs.substack.com/p/microsoft-3q…

Good point. The tweet oversimplifies Broadcom's role. Even if Google goes COT on the TPU silicon, that's only one piece of what AVGO provides. The networking ASICs, optical components (including Tomahawk/Jericho switches), SerDes IP, power management, and system-level integration work are all separate revenue streams that COT on the compute chip wouldn't touch. The real question is how much of AVGO's Google-related revenue is specifically tied to the TPU chip design vs. the broader infrastructure stack.

Rumor: Google to shift to a COT model starting with TPU v9, downgrading the role of MediaTek/Broadcom. This would be a major negative for Broadcom and MTK. According to this rumor, starting with TPU v9, Google plans to move to a model similar to AWS’s Annapurna, where Google directly places wafer orders with TSMC. In other words, Google would be doing COT directly. What does that mean? Google would become TSMC’s direct customer, while Broadcom and MediaTek would be downgraded to IP support / design service roles. Ownership of the chip and control over supply would shift toward Google. In other words, Broadcom/MediaTek’s per-chip margin would fall, if this is true. Instead of recognizing the full chip ASP as revenue, the model would shift toward design service fees / margin. That makes me more worried about Broadcom than MediaTek. $AVGO

Looking beyond the Nifty 50? Join us for an exclusive webinar with Mirae Asset Investment Managers on 29 April, Wednesday to explore how investors can capture India’s next leaders early, beyond the Nifty 50. 🕔 5:00 PM SGT/HKT | 2:30 PM IST | 1:00 PM GST Featuring Neelesh Surana, CIO, @MiraeAsset_IN , with Ashwin Patravali from @kristal_ai . Register now: us06web.zoom.us/webinar/regist… #IndiaInvesting #Midcaps #WealthManagement #Investing #Kristal #MiraeAsset

Intel is not becoming the “Western TSMC.” At least not yet. What it may be becoming is something less glamorous but more achievable: a vertically integrated x86 server CPU company with a growing advanced packaging business, selective foundry exposure, and strategic backing from customers and governments that need domestic leading-edge capacity. That’s a very different bull case from the one investors were sold under Pat Gelsinger. And after Q1, it might be the more realistic one. nikhs.substack.com/p/intels-1q26-…

GE Vernova’s Q1 looked like a blowout quarter. But the more interesting story is not the print itself. Orders are accelerating faster than revenue. Customers are pre-funding future capacity. Electrification is starting to look less like a side segment and more like a platform. And every turbine shipped during today’s scarcity window creates decades of service revenue after the bottleneck fades. The scarcity may be temporary. The annuity it creates may not be. I wrote about why Q1 strengthened the GE Vernova thesis: and why, the stock now demands a much higher burden of proof. nikhs.substack.com/p/ge-vernova-q…

Lam Research is usually framed as a great semiconductor equipment company in a great cycle. But this quarter hints at something more interesting: AI may be increasing Lam’s value per wafer while also improving its operating model. That’s the real story in the update. nikhs.substack.com/p/lam-research…

ServiceNow may be winning the strategy and losing the stock. 1Q26 strengthened the case that ServiceNow is becoming the governed execution layer for enterprise AI. But the market has moved on from that narrative. It wants proof in organic growth, cleaner numbers, and margins that hold. That is the tension at the center of this quarter: product metrics say the model is getting bigger; financial metrics say not yet. I wrote up what changed, what still matters, and which signals could decide where ServiceNow goes from here. nikhs.substack.com/p/servicenow-1…

Everyone is watching oil. Almost no one is watching what oil + gas do to food. The real second-order risk is not at the pump. It is in fertilizer, planting decisions, crop yields, and finally your grocery bill. #OilPrices #FoodInflation #MarketPulse #Commodities

A quarter with strong growth, strong cash flow, and improving retention still was not enough for the stock. Why? Because Netflix is no longer judged like a turnaround. It is judged like a premium compounder. My take on why Q1’26 was strategically stronger than it looked, and why the ad business matters more than the headline miss. nikhs.substack.com/p/netflixs-1q2…

@kristal_ai 10 year from dreaming to democratizing private banking

A decade ago, a simple question sparked a journey. Today, we’re proud to share that @kristal_ai has been awarded: 🏆Best Wealth Management Firm/MFO for Digital Innovation and 🏆Best Wealth Management Firm/MFO for Digital Investment Advice at the PWM Wealth Tech Awards 2026. As Kristal enters its second decade, this recognition feels especially meaningful. What makes moments like this matter is not just the award, but the journey behind it the team that kept building, the clients who placed their trust in us, the partners who supported us, and everyone who believed that wealth management could be more accessible, more intelligent, and more human. To our teams across Singapore, Hong Kong, India, and the UAE: thank you for building with conviction. To our clients: thank you for trusting us with what matters most. You have always been the reason. And to @FT_PWM , thank you for this recognition. We’re honoured to be acknowledged for work we care deeply about. More than a decade in, we’re still building and we’re even more excited for what comes next. #KristalAI #WealthTech #PWMAwards #Innovation #WealthManagement #DigitalInvestmentAdvice

TSMC’s 1Q26 print mattered for more than the numbers. The key development was strategic, not just financial: management is adding net new N3 capacity across multiple geographies, despite a long-standing rule against expanding older nodes past target capacity. That says a great deal about AI demand, TSMC’s role in the stack, and the durability of scarcity. My latest asks the next question: whether that strategic deepening will accrue to shareholders at the rate many assume. nikhs.substack.com/p/tsmc-1q26-th…

Everyone still talks about ASML like it’s a tool vendor. I think that framing is getting stale. The real shift is that ASML is no longer just selling machines, it’s increasingly monetizing output, throughput, and an installed base that looks more infrastructure-like than cyclical. That changes how the business should be valued. My latest note breaks down why the old demand debate is over, why the installed base matters more now, and why the biggest remaining variable is policy, not end demand. nikhs.substack.com/p/asml-1q26-ea…

$ASML Q1 EARNINGS • Revenue $10.3B vs Est. $10.1B • EPS $8.43 vs Est. $7.83 • Net Income $3.3B vs. Est. $3.0B • Gross margins 53% vs. Est. 52% Q2 Guidance • Revenue $10.3B vs Est. $10.7B • Gross margins 52% vs. Est. 53% CEO says chip demand continues to outpace supply.

Everyone still talks about ASML like it’s a tool vendor. I think that framing is getting stale. The real shift is that ASML is no longer just selling machines, it’s increasingly monetizing output, throughput, and an installed base that looks more infrastructure-like than cyclical. That changes how the business should be valued. My latest note breaks down why the old demand debate is over, why the installed base matters more now, and why the biggest remaining variable is policy, not end demand. nikhs.substack.com/p/asml-1q26-ea…

$ASML is the AI kingmaker because $TSM cannot make $NVDA most advanced GPUs without its EUV systems. It just raised full-year guidance to $43B from $41B showing AI chip demand is still running well ahead of supply.

Everyone still talks about ASML like it’s a tool vendor. I think that framing is getting stale. The real shift is that ASML is no longer just selling machines, it’s increasingly monetizing output, throughput, and an installed base that looks more infrastructure-like than cyclical. That changes how the business should be valued. My latest note breaks down why the old demand debate is over, why the installed base matters more now, and why the biggest remaining variable is policy, not end demand. nikhs.substack.com/p/asml-1q26-ea…

ASML exited 4Q25 with €38.8B in backlog, covering 6+ quarters at trailing 4Q average system revenue. A meaningful step down from the 8-14Q coverage seen during the 2022 shortage peak. The current backlog is healthier though and not driven by stockpiling like '22 peak. TSMC's trailing 4 quarter capex at $41B and growing while ASML book-to-bill just crossed back above 1.0x. If the historical pattern repeats, ASML b2b should continue rising for 2-3 more quarters. Unlike the 4Q23 spike (driven by China), the 4Q25 jump looks more durable because it's EUV-dominated and memory-driven, which implies multi-year strength.

Everyone still talks about ASML like it’s a tool vendor. I think that framing is getting stale. The real shift is that ASML is no longer just selling machines, it’s increasingly monetizing output, throughput, and an installed base that looks more infrastructure-like than cyclical. That changes how the business should be valued. My latest note breaks down why the old demand debate is over, why the installed base matters more now, and why the biggest remaining variable is policy, not end demand. nikhs.substack.com/p/asml-1q26-ea…

To those shorting ASML, you might want to be careful. TSMC earnings are tomorrow, right? We already know what Uncle C.C. Wei is going to say there. $ASML