SergeyCYW.substack.com
Capital is flowing toward companies controlling bottlenecks: compute, memory, software, or distribution. YTD performance reflects where leverage sits today. 👇
$NBIS | +198.7% YTD
Capital is flowing toward companies controlling bottlenecks: compute, memory, software, or distribution. YTD performance reflects where leverage sits today. 👇
$NBIS | +198.7% YTD
AI spending started with chips and data centers.
The next leg is software.
AI spending started with chips and data centers.
The next leg is software.
The highest expected revenue growth, above 20%, is concentrated in a small group of platforms.
The highest expected revenue growth, above 20%, is concentrated in a small group of platforms.
Video is where AI impact is most visible.
Global video advertising spend is projected to reach $207.5B in 2025, reinforcing the shift toward AI-optimized visual storytelling.
Video is where AI impact is most visible.
Global video advertising spend is projected to reach $207.5B in 2025, reinforcing the shift toward AI-optimized visual storytelling.
1. $ZETA | Zeta Global
2. $MU | Micron Technology
3. $AMD | AMD
4. $ARM | Arm
5. $RBRK | Rubrik
6. $FIG | Figma
7. $NBIS | Nebius
8. $IREN | IREN Limited
9. $S | SentinelOne
10. $NU | Nu Holdings
I don’t own them, but I’m considering buying.
1. $ZETA | Zeta Global
2. $MU | Micron Technology
3. $AMD | AMD
4. $ARM | Arm
5. $RBRK | Rubrik
6. $FIG | Figma
7. $NBIS | Nebius
8. $IREN | IREN Limited
9. $S | SentinelOne
10. $NU | Nu Holdings
I don’t own them, but I’m considering buying.
Most AI leaders control parts of the stack.
Google controls all of it—from silicon to users—under one balance sheet.
That structural integration is starting to show up in cost curves, scale, and adoption.
Most AI leaders control parts of the stack.
Google controls all of it—from silicon to users—under one balance sheet.
That structural integration is starting to show up in cost curves, scale, and adoption.
Valuations are compressing while fundamentals is improve. A few stocks stand out where execution is accelerating but multiples have not rerated. 👇
Valuations are compressing while fundamentals is improve. A few stocks stand out where execution is accelerating but multiples have not rerated. 👇
Markets rotate. Business models compound.
Long-term winners tend to share three traits: structural tailwinds, pricing power, and capital discipline.
Here are five names where fundamentals still look asymmetric. 👇
1/ $V — Visa
Markets rotate. Business models compound.
Long-term winners tend to share three traits: structural tailwinds, pricing power, and capital discipline.
Here are five names where fundamentals still look asymmetric. 👇
1/ $V — Visa
It’s about scale, automation, and capital concentration. 🧵
The market is already $843B in 2025 and is expected to reach $1.42T by 2029, growing at a 14% CAGR.
Analysts expect the strongest NTM revenue growth from:
It’s about scale, automation, and capital concentration. 🧵
The market is already $843B in 2025 and is expected to reach $1.42T by 2029, growing at a 14% CAGR.
Analysts expect the strongest NTM revenue growth from:
#NVDA - Dominates AI computing with Blackwell GPUs and CUDA software, powering global data centers and deep learning workloads.
#NVDA - Dominates AI computing with Blackwell GPUs and CUDA software, powering global data centers and deep learning workloads.
$META 22.6x | +18.2%
$NVDA 27.1x | +53.6%
$GOOGL 29.6x | +13.4%
$MSFT 29.2x | +18.3%
$AMZN 31.2x | +11.2%
$AAPL 33.1x | +10.8%
$TSLA 241x | +12.0%
$META 22.6x | +18.2%
$NVDA 27.1x | +53.6%
$GOOGL 29.6x | +13.4%
$MSFT 29.2x | +18.3%
$AMZN 31.2x | +11.2%
$AAPL 33.1x | +10.8%
$TSLA 241x | +12.0%
At first glance, Q3 looked uninspiring. Reported revenue growth slowed to +17.7% YoY, down sharply from +27.4% a year ago. Q4 guidance implies similar deceleration. Free cash flow margin compressed.
At first glance, Q3 looked uninspiring. Reported revenue growth slowed to +17.7% YoY, down sharply from +27.4% a year ago. Q4 guidance implies similar deceleration. Free cash flow margin compressed.
Big Data platforms still benefit from recurring revenue, high gross margins, and expanding AI use cases. Data infrastructure remain embedded in enterprise workflows, supporting long-term demand for #DDOG, #ESTC, #PLTR, #CRM, #ORCL, #SAP, #NOW, #CFLT.
As AI scales, the real constraint is no longer GPUs or data. It’s megawatts.
Hyperscalers are racing to secure power at the same pace they’re building compute, and the grid isn’t ready.
As AI scales, the real constraint is no longer GPUs or data. It’s megawatts.
Hyperscalers are racing to secure power at the same pace they’re building compute, and the grid isn’t ready.
#TTD is down -68.3% YTD, while S&P500 up +17.2%.
Should you invest now? Let’s examine the valuation, growth rate, key metrics, profitability, and expenses. 👇
#TTD is down -68.3% YTD, while S&P500 up +17.2%.
Should you invest now? Let’s examine the valuation, growth rate, key metrics, profitability, and expenses. 👇
I’m aligned with them on these names.
Here’s the bull case where fundamentals, positioning, and setup still look asymmetric. 👇
1/ #DDOG | Datadog
Analysts PT $212.3 | +50% upside
I’m aligned with them on these names.
Here’s the bull case where fundamentals, positioning, and setup still look asymmetric. 👇
1/ #DDOG | Datadog
Analysts PT $212.3 | +50% upside
The Rule of 40 helps investors balance growth and profitability, offering deeper insights into company efficiency, unlike EV/S multiples which reflect only revenue expectations.
The Rule of 40 helps investors balance growth and profitability, offering deeper insights into company efficiency, unlike EV/S multiples which reflect only revenue expectations.
Operating leverage, capital discipline, and control over critical infrastructure are becoming the real differentiators.
1/ $PLTR
Operating leverage, capital discipline, and control over critical infrastructure are becoming the real differentiators.
1/ $PLTR
The Big Data and security markets are no longer fighting demand. They are fighting execution, cost, and trust.
The Big Data and security markets are no longer fighting demand. They are fighting execution, cost, and trust.
Every AI model sits on a four-layer stack worth trillions. Each layer compounds the next—manufacturing, chips, networking, and cloud.
Layer 1: Manufacturing chokehold
Every AI model sits on a four-layer stack worth trillions. Each layer compounds the next—manufacturing, chips, networking, and cloud.
Layer 1: Manufacturing chokehold
Hyperscalers: $AMZN $GOOGL
AI Infra: $IREN $NBIS
Agentic AI: $PLTR $NOW $SNOW
Cybersecurity: $NET $RBRK $CRWD
Fintech: $FOUR $SOFI $DLO
E-commerce: $GRAB $MELI $SE $SHOP
Semiconductors: $NVDA $AMD $ALAB $MU
Hyperscalers: $AMZN $GOOGL
AI Infra: $IREN $NBIS
Agentic AI: $PLTR $NOW $SNOW
Cybersecurity: $NET $RBRK $CRWD
Fintech: $FOUR $SOFI $DLO
E-commerce: $GRAB $MELI $SE $SHOP
Semiconductors: $NVDA $AMD $ALAB $MU