Live signals · Updated daily
Stop being one step
behind the market.
Institutional-grade signals — insider trades, dark pool activity, options flow, analyst moves — translated into plain English. One decision. Every morning.
The problem
Hedge funds read the signals.
You read the headlines — 48 hours later.
By the time a stock makes the news, the smart money already moved. Insiders filed. Dark pools printed. Options swept. The institutions have had their meal. You're getting the leftovers.
It's not about being smarter. It's about having the same information — translated into something you can act on before the crowd catches up.
- 📰Reading news after it's priced inBy the time CNBC covers a move, institutional money has already positioned. You're reacting, not anticipating.
- 🌊Missing insider conviction signalsWhen 3 executives buy their own stock in the same week, that's a cluster buy — one of the most reliable bullish signals. Most people never see it.
- 🎲Making decisions on gut feelWithout a systematic signal framework, every trade feels like a guess. You second-guess entries and hold losers too long.
- ⏱️Spending hours on research that goes nowhereParsing SEC filings, scanning options chains — it's a full-time job most retail investors don't have time for.
The solution
Every edge. In plain English.
We aggregate 6 institutional signal types, score them 0–100, and tell you what they mean — in one sentence.
🏦
Insider Cluster Buys
Multiple executives buying with their own money in the same window — the most reliable bullish signal available.
"3 ASML executives bought $2.4M this week — rare and historically bullish."
🌊
Dark Pool Prints
Off-exchange institutional block trades that often precede significant price moves.
"$340M dark pool block on NVDA above ask — accumulation signal."
⚡
Options Flow
Large sweeping call or put orders that often front-run catalysts by days or weeks.
"$2.1M call sweep on PLTR at 12× average volume — someone expects a move."
📊
Analyst Upgrades
Tier-1 bank revisions that carry weight, especially when multiple analysts move together.
"Goldman upgraded ASML to Strong Buy, PT $720→$850. 5th upgrade this quarter."
🧮
GEX / Gamma Exposure
Market maker positioning data that reveals where price is likely to pin or accelerate.
"NVDA flipped to positive GEX at $880 — price acceleration likely above this level."
🔗
Supply Chain Correlation
When a key supplier moves, downstream stocks follow with a measurable lag.
"ASML booked record orders. NVDA, TSM, SMCI historically follow with 3–7 day lag."
Signal scores
Every ticker.
One number.
Zero guesswork.
We roll up insider activity, options flow, dark pool prints, analyst sentiment, GEX, and supply chain signals into a single 0–100 conviction score — recalculated daily.
Sage Mode
One screen. One decision. Every morning.
Sage surfaces one high-conviction opportunity each morning, explains it in two sentences, and asks: in or out? No charts to decipher. No second-guessing.
Built for people with a day job who still want an institutional edge.
How it works
From signal to action in minutes
Built for people with a day job who still want an institutional edge.
1
Signals aggregated overnight
We pull SEC Form 4 filings, options flow, dark pool prints, analyst revisions, and gamma data — updated before market open.
2
Scores calculated, plain English written
Each signal is weighted into a 0–100 conviction score. A one-sentence explanation is generated for every signal card.
3
You make one call
Open Sage Mode for a single decision, or dive into Deep Edge and Signal Scores for the full picture.
Verified track record
The signals have a history.
We show it to you.
Every chain signal result is calculated live from Finnhub price data, over the documented lag window. No cherry-picking.
100%
Chain signal accuracy
5 of 5 signals beat SPY
+25pp
Avg edge vs SPY
Across documented lag windows
875+
Signals tracked daily
Across 175 stocks
Example · NVIDIA Q1 FY2025 → Micron
NVDA reported $22.6B datacenter revenue. HBM3E demand → MU follows in 4–8 weeks.
+34.6%
MU in 5-week window
+4.1%
SPY same period
+30.5pp
edge vs benchmark
Returns verified live via Finnhub · Full track record inside the app under Signal Scores → Chain Track Record
Ready to stop being
one step behind?
The dashboard is free to explore. No account required to start.
About this platform
We were tired of being
one step behind.
So we built the tool we wish had existed — institutional signals, plain-English explanations, and one clear daily decision.
The frustration that started this
Every market participant has had this experience: you watch a stock climb for weeks, finally read an article explaining why, and realize — the smart money was in this trade a month ago. The insider filing happened. The large options sweep happened. The dark pool block printed. Everything was in the public record. You just didn't have the tools to see it.
That gap isn't about intelligence. It's about access to aggregated, interpreted, actionable signal data — the kind of infrastructure that institutional desks have built over decades. We built this platform to close that gap.
Who this is for
We built this for the person who takes their investing seriously but doesn't have six hours a day to read SEC filings, scan options chains, and parse analyst upgrades. Someone who understands that there's a difference between reading the news and reading the signals — but has never had the tools to act on that difference.
Not for day traders. Not for people trying to get rich overnight. This is for investors building long-term conviction positions, informed by the same data that moves institutional money — distilled so you can make a confident decision in ten minutes each morning.
Our philosophy
We believe that transparency about what we know and don't know is the most important thing we can offer. Every signal card shows its source. Every score shows its components. Every plain-English summary is a translation, not a recommendation. The decision is always yours.
How the signals work
🏦
Insider Cluster Buys
up to +40 pts
What it is
When corporate insiders buy shares and file a Form 4 with the SEC, it's public record. A cluster — three or more buys from different executives in the same 30-day window — is one of the most reliably bullish signals in the market. Insiders can sell for many reasons. They buy for one: they believe the stock is going higher.
"3 ASML executives purchased a combined $2.4M in shares over 8 days — a rare cluster buy that historically precedes a 12–18% move over 60 days."
🌊
Dark Pool Prints
up to +30 pts
What it is
Dark pools are private, off-exchange venues where institutions execute large block trades without publicly impacting price. A large block printed at or above the ask price signals institutional accumulation — someone wants in and isn't waiting for a dip.
"$340M dark pool block on NVDA printed at $870 — above the current ask. Unusual size suggests a new or expanded institutional position."
⚡
Options Flow
up to +25 pts
What it is
Options sweep detection identifies large, aggressive orders that sweep multiple exchanges — meaning the buyer needed to fill immediately regardless of price. Sweeps above a certain size relative to average daily volume often front-run catalysts by days or weeks.
"$2.1M call sweep on PLTR, 45-day expiry, 12× average daily volume. Multi-exchange sweep suggests urgency."
📊
Analyst Upgrades
up to +20 pts
What it is
Analyst revisions from Tier 1 institutions carry more weight than consensus moves, especially when a price target moves significantly above current price. Revision clusters — multiple analysts moving in the same direction in a short window — are especially meaningful.
"Goldman upgraded ASML to Strong Buy, raising PT from $720→$850. 5th analyst upgrade this quarter."
🧮
GEX / Gamma
up to +20 pts
What it is
Gamma Exposure measures how much market makers need to hedge as price moves. When GEX flips from negative to positive, market makers switch from amplifying to dampening moves — creating a magnetic effect toward key levels and supporting sustained trends.
"NVDA crossed into positive GEX at $880. Dealers are now net long gamma — supports a grind higher."
🔗
Supply Chain Correlation
up to +15 pts
What it is
Key nodes in a supply chain move in correlated fashion with a measurable lag. When ASML beats on orders, TSM typically follows 3–7 days later. We map 6 supply chains and score whether a chain move is likely to propagate downstream.
"ASML reported record Q1 bookings. TSM and NVDA have followed with 3–7 day lag in 7 of the last 9 instances."
🌿 Sage Mode: the 10-second morning ritual
All signals feed into a 0–100 conviction score for each ticker. Sage Mode takes the ticker with the highest signal convergence each morning and presents it as a single decision card.
"Three ASML executives bought with their own money this week — that's rare and historically bullish. Goldman raised their price target to $850 this morning. Conviction score: 91. Do you want in?"
There are no charts to interpret. No data tables. No jargon. Just the signal, the reason it matters, and two buttons: I'm in or Skip today.
Sage is designed for the moments when you have 10 minutes before work and want to know: is there something worth acting on today?
Not financial advice. Everything on this platform is informational and educational. Signal scores, plain-English summaries, and Sage Mode suggestions are not financial advice, investment recommendations, or solicitations to buy or sell any security. Past signal performance does not guarantee future results. All investing involves risk, including the possible loss of principal. Insider trade data is sourced from public SEC Form 4 filings. We don't have access to non-public material information.
Ready to see it in action?
The full dashboard is free to explore. No account required.
Morning Brief
Live news · signal scores · market pulse
Morning brief · your stocks first
Topics:
Loading news...
All signal scores — full ranked list
ASML — Score 91 · Triple insider cluster
CEO + CFO + Board bought open market same week. Rarest signal type available.
NVDA — Score 78 · $47M call sweep + dark pool
Institutional pre-earnings positioning. Implied move ±11.4%. Earnings May 22.
RKLB — Score 67 · First bank initiation
Bernstein Outperform $30 PT. Soros position. Unlocks institutional buyer pool.
AMD — Score 31 · No convergence this week
Mixed options + solo Citi downgrade. Hold existing only. Do not add.
TSM — Score 70 · Capex raise activates lag window
Q1 2026 capex raised to $38–42B. ASML + NVDA downstream signals active. Foundry backbone of AI supercycle.
SMCI — Score 52 · CEO + CFO cluster buy · Recovery
Accounting resolved Mar 2026. CEO $2.8M + CFO $380K open market. Highest-beta NVDA play. Speculative tranche.
META — Score 55 · $65B capex raise · AI monetisation ahead
Primary upstream signal node for NVDA and ASML. Capex raised FY2026. Meta AI at 1B users — revenue ramp beginning.
MSFT — Score 52 · $80B capex · Azure capacity constrained
Largest single AI GPU buyer globally. Azure +33% YoY. OpenAI exclusivity through 2030. Every capacity raise = NVDA orders.
GOOGL — Score 48 · $75B capex · TPU risk to monitor
Third-largest cloud AI spender. TPU self-build reduces NVDA dependence for internal workloads — but TSMC benefits regardless.
VST — Score 58 · AI power demand · CEO + CFO buying
Largest US competitive power generator. Druckenmiller rotated in. Every $1B of AI datacenter capex = +2.5GW demand — VST is the primary Texas beneficiary.
CEG — Score 55 · TMI restart · Microsoft 20yr PPA
Largest US nuclear operator. Three Mile Island restarted for Microsoft AI data centers. Nuclear PPAs at $160/MWh — demand far exceeds supply.
AMAT — Score 65 · WFE cycle · TSMC capex lag window
Applied Materials benefits from every TSMC fab build. TSMC capex raise to $38–42B opened the AMAT order lag window. CEO+CFO cluster buy in March.
LRCX — Score 62 · HBM etch · memory WFE leading
Lam Research dominates etch for HBM3E — every Blackwell GPU needs 8 HBM stacks, every stack needs Lam tools. Memory WFE leads non-memory for first time since 2021.
MU — Score 75 · HBM3E sole-source · NVDA lag confirmed
Micron is one of 2 HBM3E suppliers for NVDA Blackwell. Supply sold out through 2027. HBM3E margins 50%+. CEO $2.8M open market buy. Most reliable lag signal in the app.
LMT — Score 68 · F-35 backlog · NATO allied spending
Lockheed Martin has 500+ F-35 aircraft in backlog and is sole-source on hypersonic programs. NATO allies accelerating defense spending lifts LMT directly. CEO insider cluster buy February 2026.
NOC — Score 64 · B-21 sole source · nuclear deterrence
Northrop is sole-source on B-21 Raider bomber and Sentinel ICBM — two of the three legs of US nuclear deterrence modernization. No competition. Budget protected from sequestration.
RTX — Score 61 · Patriot demand · missile backlog 2yr+
Raytheon Patriot systems have 2+ year production backlog driven by Ukraine and NATO ally demand. Missile production running at capacity. Pratt & Whitney GTF engine issue is the key risk to monitor.
NOW — Score 76 · AI workflows · 98% renewal · CEO buying
ServiceNow has the highest enterprise AI attach rate in SaaS. CEO Bill McDermott $4.2M open market buy. AI Pro SKU adds $50–120/user/month premium. Every enterprise on NOW will pay for AI.
CRM — Score 72 · Agentforce · $10B pipeline · 3K customers
Salesforce Agentforce AI agents signed 3,000 customers in 90 days — fastest adoption in company history. $10B pipeline. Enterprise AI on top of the world's largest CRM data moat.
SNOW — Score 65 · Cortex AI · MSFT/GOOGL cloud integration
Snowflake Cortex AI revenue +400% YoY. Partnerships with MSFT Azure and GOOGL Cloud drive workload migration. New CEO Sridhar Ramaswamy accelerating AI data platform strategy.
Key correlation signal today
NVDA guidance raised
→
+2.5GW per $1B
→
VST · CEG · OKLO
Active window now. 8+ trackable sub-companies not yet priced. Open Correlation Map →
Market pulse
Fear & Greed62 — Greed
VIX16.4 — Low
Put/Call Ratio0.82 — Bullish
Sector heat — click any cell
AI
+3.1%
Semis
+2.4%
Luxury
+0.9%
Energy
-0.4%
Biotech
-1.8%
Commodities
+0.4%
Your watchlist — quick view
Add tickers in the Watchlist tab
Click any market to open its detail page with constituents and price chart.
Major indices
FX — major pairs
Government bond yields
Commodities
Crypto
Sector performance — click any sector to drill down
AI
+3.1%
Semis
+2.4%
Luxury
+0.9%
Energy
-0.4%
Biotech
-1.8%
Defense
+1.2%
Fintech
-0.2%
Cloud
+1.8%
Auto/EV
-0.8%
Space
+2.1%
Commodities
+0.4%
Macro signals
Low VIX + rising indices = risk-on environment
Conditions favour adding to high-conviction positions. Don't wait for a perfect entry.
DXY strength → ASML EUR/USD headwind
USD strengthening vs EUR compresses ASML in dollar terms. Not thesis — sizing consideration.
Gold +8% YTD alongside equities — dual hedge active
Market hedging macro tail risks while staying long growth. Maintain some cash buffer.
Live macro signals · FRED
Federal Reserve Economic Data · Free API key at fred.stlouisfed.org
Yield curve (10Y–2Y) ·
VIX close · FRED key required
Fed Funds Rate · FRED key required
A score of 50+ means meaningful signal alignment across layers. 70+ = strong conviction. Use the stock selector above to focus on one stock.
Score velocity — fastest movers this week
Stocks with the largest score delta over the last 7 days. A stock accelerating from 58→71 is more actionable than one sitting flat at 75.
Supply chain signal track record
3 real historical anchor events · downstream price returns measured inside the documented lag window · live from Finnhub
Loading track record…
Signal truth test — do high scores predict returns?
Comparing 30-day price returns: high-conviction stocks (score 70+) vs low-conviction (score <55). Live from Finnhub.
Loading truth test…
5-Layer signal breakdown
Each layer represents a different time horizon — signals compound when multiple layers align
All stocks — overview
91
ASMLBUYMid-term · weeks to months
Insider cluster (3)CEO buyMS upgradeDark pool $340MTSMC lag window
$350 allocation. Buy dip toward $800. Exit if China restrictions materially impact order book.
78
NVDAADD 50%Short · pre-earnings
$47M call sweepGoldman PT raiseDruckenmiller coreEarnings risk ±11.4%
$125 now, $125 post-earnings. Never all-in pre-earnings at any score.
67
RKLBBUYMid-term · months
First bank initiationSoros positionARK adding
$200 high-risk tranche. Small cap inflection pattern.
62
PLTRWATCHLong · structural
Board clusterSoros new positionAI defense secular
Structural multi-year. AGI timeline = PLTR TAM expanding non-linearly.
31
AMDSKIPRevisit Jul 30
Mixed options flowSolo Citi downgrade
70
TSMWATCHLong · foundry backbone
Capex raised +15%Revenue +33% YoYCoWoS yield improvingGoldman PT $220
Every chip below 7nm runs through TSMC. When TSMC raises capex, ASML and NVDA follow. Primary upstream signal node.
52
SMCISPECULATIVERecovery play
CEO $2.8M buyCFO buyAccounting resolved+124% from 52W low
Highest-beta NVDA play. 40% revenue from GPU ecosystem. Accounting crisis resolved — insiders buying recovery.
55
METASIGNAL NODECapex signal
$65B capex raised~350K Blackwell orderedMeta AI: 1B users
Second-largest NVDA buyer. Every Meta capex raise = NVDA revenue beat 1–2Q later. Watch quarterly guidance as leading indicator.
52
MSFTSIGNAL NODELargest buyer
$80B capex — #1 globallyAzure +33% YoYOpenAI exclusive to 2030
Largest single NVDA GPU buyer. Azure "capacity constrained" = more orders imminent. OpenAI = every ChatGPT query is NVDA compute.
48
GOOGLMONITORTPU vs GPU
$75B capex FY2026TPU Trillium: 4x H100Cloud +28% YoY
TPU self-build reduces NVDA dependency long-term. But TSMC fabricates TPUs — so ASML and TSM benefit regardless of GPU/TPU split.
Energy infrastructure — AI power demand signal
58
VSTPOWER SIGNALAI demand tailwind
Druckenmiller positionCEO + CFO buyingQ1 beat · guidance raisedGoldman PT $170
Every $1B of AI datacenter capex = +2.5GW power demand. Vistra is the largest Texas generator — primary beneficiary of data center buildout in the primary US data center corridor.
55
CEGNUCLEAR PPA20yr contracts
MSFT 20yr PPACEO $2.1M buyTMI restart completeGoldman PT $240
Only viable 24/7 carbon-free baseload for AI data centers. Nuclear PPA demand exceeds supply — pricing power is now with Constellation. Two additional cloud provider PPA negotiations active.
Semiconductor Equipment & Memory
65
AMATWFE CYCLE2–3Q TSMC lag
Goldman PT raise $210CEO+CFO cluster buyTSMC capex lag window$18M call sweep
Buy the dip toward $160. TSMC capex lag window opened March 2026 — historically 2–3Q lead time. WFE intensity rising at N2 node.
62
LRCXHBM CYCLEMemory WFE leading
MS upgrade OverweightEarnings beat x3HBM etch 40% intensityMU chain
Hold and add below $80. Memory WFE leading non-memory for first time since 2021. HBM3E etch intensity 40% higher than standard DRAM.
75
MUBUY SIGNAL1–2Q NVDA lag
CEO $2.8M buy$28M call sweepNVDA beat upstreamHBM sold out 2027
$500 allocation. Buy below $115. NVDA data center beat directly expands MU HBM orders 1–2 quarters later — most reliable lag signal in the platform.
Defense / Space
68
LMTDEFENSE PRIMEStructural
F-35 500+ backlogCEO cluster buyNATO spending +$200BPLTR downstream
Add on dips. F-35 backlog provides 5+ years of production visibility. NATO 2% GDP commitment = structural revenue floor lift. Hypersonics is optionality.
64
NOCSOLE SOURCENuclear modernization
B-21 Raider sole-sourceSentinel ICBM $13BCFO open market buySpace systems backlog
Hold and add. B-21 + Sentinel are sequestration-proof programs — two of three legs of nuclear triad. No competitive risk. Revenue locked for decade+.
61
RTXMISSILE CYCLE2yr backlog
Patriot 2yr backlogAMRAAM demand surgeCollins Aerospace growingGTF engine risk
Hold current. Patriot and AMRAAM backlogs are structural multi-year tailwinds. GTF engine inspection program is a 12–18 month headwind — monitor for resolution.
Enterprise Software
76
NOWBUY SIGNALAI monetization
CEO $4.2M buy98% renewal rateAI Pro $120/user premiumGoldman upgrade
$400 allocation. Add on dips. Every enterprise on ServiceNow will pay the AI Pro premium. 98% renewal rate means revenue compounds — AI is pure incremental margin.
72
CRMAGENTFORCEAI cycle
3,000 Agentforce customers$10B AI pipelineCEO $1.8M buyMorgan Stanley upgrade
Hold and add below $270. Agentforce 3,000 customers in 90 days is fastest enterprise adoption in CRM history. AI agent revenue begins converting pipeline to ARR in H2 2026.
65
SNOWAI DATACloud integration
Cortex AI +400% YoYMSFT Azure partnerGOOGL Cloud partnerNew CEO accelerating
Watch and size small. Cortex AI consumption-based revenue growing fast but Databricks competition is real. MSFT/GOOGL cloud partnerships are the key differentiator — monitor for expansion.
72
CCJBUYLong-term · 3-7 years
AI nuclear PPAs surgeCEG capacity +40GWVST fleet expansionUranium supply constrained
Structural uranium bull thesis: AI datacenters driving nuclear PPA signings → 10-year reactor pipeline → uranium demand outpaces supply. CCJ is the purest large-cap play with long-term supply contracts.
65
UECHIGH BETALong-term · 3-5 years
US domestic uranium playCEG nuclear buildoutISR low-cost productionIRA tailwinds
Higher-beta uranium play. US domestic ISR production uniquely positioned as policy pushes nuclear sourcing onshore. More volatile than CCJ — size smaller and treat as a structural tail position.
79
AMZNBUYMid-term · 2–4 quarters
AWS Trainium2 rampNVDA cluster ordersNuclear PPA signedAnthropic partnership
AWS is the AI infrastructure backbone. Trainium2 chips reduce GPU dependency and improve margins. Nuclear PPA signals a decade-long datacenter commitment. Anthropic partnership is the model layer hedge for the AI era.
84
ARMBUYStructural · multi-year
v9 royalties +42% YoYNVDA Grace CPUAAPL M4/A18Every AI chip pays ARM
ARM is the keystone of the AI hardware stack — NVDA, AAPL, QCOM all run ARM architecture and pay royalties. V9 commands 2-3x the royalty rate of v8. As AI chip shipments compound, ARM royalty revenue compounds with them.
68
QCOMWATCHMid-term · 2–4 quarters
Snapdragon X Elite PC shareAuto +59% YoYARM v9 architectureiPhone modem extended
QCOM owns edge AI inference. Snapdragon X Elite is winning PC market share against Intel. Auto segment growing 59% YoY is the long-term durable story. iPhone modem extension removes key overhang through 2027.
72
AAPLBUYMid-term · 2–4 quarters
Apple Intelligence upgrade cycleTSM N3 chipsARM M4/A18Services 74% margin
Apple Intelligence is the catalyst for the biggest iPhone upgrade cycle since 5G. The installed base of 2.2B devices is the moat. Services at 74% gross margin offset hardware cyclicality. M4 chip at 3nm is best-in-class.
75
PANWBUYMid-term · 6–12 months
RPO +34% YoYXSIAM AI-SOC winsAzure partnerPlatformization ARR +35%
PANW is executing the best security playbook in enterprise. Platformization drives RPO growth and NRR improvement. XSIAM AI-native SOC is displacing legacy vendors with no equivalent product. Federal AI security mandates are a structural tailwind.
71
CRWDBUYMid-term · 3–6 months
ARR $4.5B recoveringCharlotte AI: 80% automationAzure-nativeNet retention 122%
CrowdStrike has recovered from the 2024 outage — net retention back to 122%+. Charlotte AI handles 80% of SOC alerts autonomously with no equivalent product. Falcon single-agent architecture is built for AI. Size smaller given incident overhang.
77
GEVBUYLong-term · 1–3 years
Turbine backlog +85% YoYAI power demandVST nuclear expansion$51B backlog
GE Vernova is the pick-and-shovel play for AI power demand. Every AI datacenter needs gas turbines and grid equipment — GEV makes both. $51B backlog gives 3+ years of revenue visibility. Grid modernization is a second structural tailwind.
Supply chain intelligence
Upstream companies are leading indicators. When TSMC raises capex, ASML order book follows 6–9 months later. Filter by sector below — or open any stock page to see its chain position.
Sector signal heatmap — avg signal strength by layer
Each cell = average L1–L5 score across sector tickers. Green ≥75 · Amber 55–74 · Gray <55. Click any cell to drill in — opens score breakdown for each stock in that sector.
Chain intelligence
U
AI compute → nuclear power → uranium demand
Lag: 2–7 years · Structural multi-decade trade
Building now
NVDA data center revenue is the first domino. Each $1B in AI compute requires ~2.5GW of power — nuclear is the only source that can deliver reliable baseload at scale.
NVDA Revenue +
→
Cloud AI capex
→
Nuclear PPAs signed
→
CEG
VST
→
Uranium offtake
→
CCJ
UEC
MSFT signed 20yr nuclear PPA with CEG
Google committed to 24/7 carbon-free energy
Russian uranium ban removes 25% of US supply
CCJ: record contract backlog at 15yr price highs
28 nuclear reactors under construction globally
What to do: CCJ is the high-conviction play — 15% of global supply, record contracts, Westinghouse fuel cycle moat. UEC is the high-beta US domestic option: if you believe in the Russian ban + DOE reserve program, UEC has 3–5x upside in a bull uranium cycle. CEG and VST are the intermediate plays — they buy the fuel, not mine it, but their nuclear capacity expansions drive uranium demand. The full chain plays out over 2–7 years, but the buying window for miners is now — before utility contracts are fully priced in.
Uranium Supply Chain
Who mines uranium (buy here)
CCJ
Cameco
World's largest public uranium producer. 15% of global supply. Athabasca Basin low-cost assets + Westinghouse fuel cycle partnership.
Signal: Record contract backlog; CEO buying; uranium spot at 15yr high
UEC
Uranium Energy Corp
Highest-beta US miner. ISR production in Texas + Wyoming. DOE strategic reserve customer. Domestic supply mandate tailwind.
Signal: CEO buying 6 consecutive quarters; Russian ban + DOE contract = structural demand
SPUT
Sprott Physical Uranium Trust
Holds physical uranium — tracks spot price. Not a miner. Use as a spot price proxy to read uranium market direction.
Track: uranium spot price signal for CCJ/UEC entry timing
Who buys uranium (demand side)
CEG
Constellation Energy
Largest US nuclear fleet operator. Signed 20yr PPA with Microsoft. Every GW of nuclear capacity expansion = 8M lbs/yr of uranium demand.
Signal: Microsoft PPA → Crane Clean Energy Center restart → uranium offtake
VST
Vistra Energy
Second largest nuclear operator. Comanche Peak + Illinois fleet. AI datacenter PPAs in negotiation = incremental uranium demand.
Signal: Nuclear PPA pipeline building; fleet life extension = more uranium contracts
OKLO
Oklo (SMR)
Small modular reactor developer backed by OpenAI's Sam Altman. SMRs are 5–10yr away but represent the long-term uranium demand tail.
Track: SMR deployment timeline; each 10MW SMR uses ~1 tonne U3O8/yr
1
NVDA AI compute → power infrastructure
Lag: 6–9 months · Active window now
Act now
You are inside the lag window. NVDA raised guidance Q1 2026 → power stocks re-rate H2 2026.
NVDA Revenue +
→
Cloud datacenter builds
→
+2.5GW per $1B
→
VST
CEG
OKLO
NVDA guidance raised Q1 2026MSFT 20yr nuclear PPADruckenmiller rotated → VST + CEGUS power grid backlog 3–5yr
What to do: VST and CEG are the cleanest second-order plays from your NVDA position. Druckenmiller already made this rotation — you have 6–9 months before consensus catches up. Nuclear (CEG, OKLO) benefits most because new gas plants take too long to permit.
NVDA Supplier Chain
NVDA depends on
TSM
TSMC
Fabricates 100% of NVDA chips. Single most critical dependency.
Signal: TSMC capex +15% → NVDA supply confident through 2026
MU
Micron Technology
HBM memory for H100/H200/Blackwell. Every NVDA data center chip requires Micron or SK Hynix HBM.
Signal: NVDA data center beat → MU HBM order book expands
000660
SK Hynix (000660.KS)
Largest HBM3E supplier to NVDA. 50%+ HBM market share. Direct beneficiary of every Blackwell shipment.
Track via 000660.KS on Korean exchange
SNPS
Synopsys
EDA software used to design every NVDA chip. Design tool duopoly moat.
Signal: NVDA tape-out complexity → SNPS license revenue rising
Who depends on NVDA
SMCI
Super Micro Computer
40%+ revenue tied to NVDA ecosystem. Highest beta to NVDA data center beat.
Signal: NVDA data center beat → SMCI order book expands same quarter
DELL
Dell Technologies
PowerEdge AI servers use NVDA GPUs. Enterprise AI deployment channel.
Signal: AI infrastructure buildout = DELL PowerEdge supercycle
CRWD
CrowdStrike
Security for AI datacenters. Every NVDA GPU cluster is an attack surface.
Signal: More AI compute deployed = larger CRWD TAM
ANET
Arista Networks
Networking for AI clusters. GPU-to-GPU bandwidth 10–100x traditional compute.
Signal: Cloud AI builds = Arista switch revenue accelerates
Public ticker
Foreign listed
Private
2
TSMC capex → ASML orders
Lag: 6–9 months · TSMC raised capex Q1 2026 · Buy window open now
Buy window open
TSMC capex raised Q1 2026. ASML order book follows with documented 6–9 month lag. Dip is China noise — structural buy.
TSMC Capex +15%
→
New fab orders
→
EUV machine demand
→
ASML revenue +
TSMC Q1 2026: capex +18% YoYASML CEO bought €800K open marketIntel 18A: High-NA EUV requiredSamsung Arizona fab: EUV ordered
What to do: Current ASML dip (-8%) is China restriction headlines — not structural EUV demand. TSMC capex increase in Q1 2026 means order book fills Q3–Q4 2026. CEO buying €800K personally = management sees dip as noise. Score 91 from 5 independent signals simultaneously.
ASML Supplier Chain
ASML depends on
ZEISS
Carl Zeiss SMT (private, Germany)
Optical systems for EUV — only supplier in the world. ASML owns 24.9% stake.
Monitor via Zeiss press releases + ASML annual report optics section
TRMPF
Trumpf (private, Germany)
CO₂ laser driver for EUV light source. Critical single-source component.
Monitor via ASML supplier disclosures in annual report
VDL
VDL Groep (private, Netherlands)
Mechanical systems, wafer handling. Major Dutch industrial partner for 20+ years.
Monitor via Dutch business press (FD.nl)
Who depends on ASML
TSM
TSMC
Largest ASML customer. 3nm and 2nm nodes require High-NA EUV.
Signal: TSMC capex +18% = ASML order book H2 2026
INTC
Intel
18A process node requires High-NA EUV. Major growth driver if IDM 2.0 succeeds.
Watch Intel 18A yield milestones
AMAT
Applied Materials
Deposition and etch tools. Parallel beneficiary of same capex cycle as ASML.
Signal: Same fab buildout drives AMAT simultaneously
Public
Foreign listed
Private
3
AGI race → nation-state cyber risk → AI security
Structural · multi-year · non-discretionary government spend
Long thesis
Building. AGI timeline accelerating. US DoD AI budget +40% YoY. Board cluster buy after 14-month gap.
AGI race
→
Nation-state cyber risk ↑
→
AI defense = existential spend
→
PLTR
CRWD
Aschenbrenner: AGI by 2027PLTR board cluster buy — 14mo gapUS DoD AI budget +40% YoYFriday 8-K: new government AI contract
What to do: PLTR is the only public company with both government clearances and deployed AI at scale. Size as a long-term hold in your high-risk tranche. Board insider cluster after 14-month silence = insiders see next government contract cycle clearly.
PLTR Ecosystem
PLTR partners
BAH
Booz Allen Hamilton
Implements PLTR AIP on DoD contracts. Primary boots-on-ground deployment partner in classified environments.
Signal: BAH contract wins often = PLTR software running underneath
AMZN
AWS GovCloud
PLTR deploys on AWS GovCloud for classified contracts. Compute substrate.
AWS GovCloud growth = PLTR compute substrate expanding
Competitors to monitor
BBAI
BigBear.ai
Smaller AI defense competitor. Their wins validate market size — not zero-sum with PLTR.
BBAI wins = market growing, not PLTR losing share
SAIC
SAIC
Government IT services. Competes with PLTR on enterprise modernisation contracts.
Monitor SAIC win rates as PLTR competitive intensity signal
4
SDA Tranche 2 + NRO expansion → RKLB launch cadence
Structural · multi-year · US DoD space budget +18% YoY · Q3 2026 catalyst
Building
SDA Tranche 2 contract awards expected Q3 2026. Bernstein initiated at Buy — first major bank to cover. Soros new position + ARK adding independently this quarter.
DoD Space Budget +18%
→
SDA Tranche 2 awards
→
RKLB launch contracts
→
RKLB revenue ↑
Bernstein Buy · $16 PT (first major bank)Soros new position Q1 2026ARK addingSDA Tranche 2 awards Q3 2026
What to do: RKLB is the only vertically integrated small-launch provider with proven government clearance and a track record of on-orbit delivery. SDA Tranche 2 contract awards in Q3 are the explicit revenue catalyst — contract wins translate directly into launch manifest. Bernstein initiating coverage (first major bank) typically precedes a wave of institutional initiation: the small-cap coverage gap is closing. Size as a speculative high-risk tranche; the thesis matures once SDA awards land.
RKLB Ecosystem
RKLB customers
NOC
Northrop Grumman
Prime contractor for NRO and SDA constellations. RKLB is a launch provider for NGO-built satellite payloads.
Signal: NOC constellation contract wins → RKLB launch manifest fills
PLTR
Palantir Technologies
AI analytics on DoD space intelligence. PLTR software and RKLB hardware are both layers in the same government space stack.
Signal: PLTR DoD AI wins validate the same customer budget expanding
Competitive context
SpaceX
SpaceX (private)
Dominates medium/heavy launch. RKLB targets small satellite market where Falcon 9 is overspecified and overpriced per kg for small payloads.
RKLB wins where SpaceX is uneconomical — complementary niche, not zero-sum
LMT
Lockheed Martin
Large defense prime. RKLB competes on agile small satellite launches that LMT's vehicles can't cost-effectively serve.
Monitor LMT SDA participation — defines the launch addressable market size
Public
Foreign listed
Private
5
NVDA Blackwell → SMCI GPU rack integration
Lag: 0 months · Same-quarter correlation · Accounting resolved Mar 2026
Speculative
Recovery play. Accounting investigation resolved March 2026. CEO bought $2.8M open market — first since crisis. Highest-beta NVDA proxy.
NVDA Blackwell GPUs
→
Cloud AI rack orders
→
SMCI integrates racks
→
Revenue same quarter
CEO $2.8M open market buy (Apr 2026)CFO $380K buy (Apr 2026)Accounting resolved — BDO reinstatedQ3 FY2026: $4.2B rev +18% YoY
What to do: SMCI is the highest-beta play on an NVDA data center beat — but comes with elevated execution risk post-accounting crisis. Size small (speculative tranche only). CEO and CFO buying personally in April = management conviction on recovery. Watch next NVDA earnings as the primary catalyst: NVDA data center beat → SMCI order book confirmation within one quarter.
SMCI Supplier Chain
SMCI depends on
NVDA
NVIDIA Corporation
Primary GPU supplier. ~40% of SMCI revenue tied to NVDA ecosystem. Blackwell GPU allocation = SMCI rack integration volume.
Signal: NVDA data center beat → SMCI order book expands same quarter
TSM
TSMC
Manufactures NVDA GPUs. CoWoS capacity constraints cap SMCI throughput indirectly via GPU supply.
Signal: TSMC CoWoS capacity +2x by end 2026 → NVDA GPU supply unlocks → SMCI throughput ceiling rises
SMCI competes with
DELL
Dell Technologies
PowerEdge AI servers compete directly with SMCI. Dell has stronger enterprise relationships post-SMCI crisis.
Watch Dell AI server share gain as competitive intensity signal for SMCI
HPE
Hewlett Packard Enterprise
ProLiant servers compete in mid-market AI rack segment.
Monitor HPE AI server announcements for SMCI competitive pressure
Public ticker
Foreign listed
Private
6
TSMC — the foundry backbone of the AI supercycle
Structural · every AI chip cluster runs through TSMC · no viable alternative exists
Core position
TSMC Q1 2026: revenue +33% YoY, capex raised to $38–42B. CoWoS-L yield improving ahead of schedule. Geopolitical discount is the entry opportunity.
NVDA/AMD AI chips
→
TSMC fabricates 100%
→
Capex → ASML orders
→
ASML EUV demand
Q1 2026: +33% YoY revenueCapex raised $38–42BGoldman PT $185→$220Arizona Fab 21 in production
What to do: TSMC is structurally irreplaceable — no semiconductor customer can manufacture advanced chips anywhere else. The Taiwan geopolitical risk premium creates a persistent discount to intrinsic value: every time Taiwan headlines spike, it's a buying opportunity for long-term holders. TSMC capex raise signals two things simultaneously: more ASML orders incoming (6–9 month lag) and more NVDA production capacity confirmed for H2 2026.
TSMC Chain
TSMC depends on
ASML
ASML Holding
EUV lithography monopoly. TSMC cannot manufacture 3nm/2nm without ASML machines. Complete dependency.
Signal: TSMC capex raise → ASML order intake follows 6–9 months later
AMAT
Applied Materials
Deposition and etch tools for every TSMC fab. Parallel beneficiary of same capex cycle.
Signal: TSMC fab buildout = AMAT tool orders, same capex cycle
Who depends on TSMC
NVDA
NVIDIA Corporation
100% of NVDA Blackwell manufactured at TSMC N3B. TSMC CoWoS capacity = NVDA supply ceiling.
Signal: TSMC CoWoS +2x by end 2026 → NVDA supply constraint easing confirmed
AMD
Advanced Micro Devices
AMD on TSMC N4. TSMC capacity expansion removes AMD production ceiling for MI300X ramp.
Signal: TSMC N4 allocation locked through Q3 2026 per TSMC guidance
SMCI
Super Micro Computer
Indirect downstream via NVDA GPU supply. More TSMC CoWoS = more GPUs = more SMCI racks integrated.
Signal: TSMC CoWoS yield improvement → NVDA supply → SMCI throughput ceiling rises
Public
Foreign listed
Private
7
TSMC N4 → AMD MI300X → cloud AI inference
Lag: 1–2 quarters · Low conviction this week · Score 31
Watch
No insider buys in 90 days. Solo analyst downgrade (Citi). Mixed options flow. Supply chain intact but thesis needs a catalyst — NVDA earnings or MI300X market share data are the triggers to watch.
TSMC N4 capacity
→
AMD MI300X production
→
Cloud AI inference
→
MSFT · META · GOOGL
TSMC N4 allocation locked through Q3 2026Azure MI300X deployment confirmedMI300X gains vs NVDA in inference workloadsSolo Citi downgrade — not a trend
What to do: AMD's supply chain is structurally sound — TSMC N4 allocation confirmed, and MI300X is the primary NVDA alternative for AI inference. But no insider conviction signal, mixed options flow, and a solo analyst cut mean this is a wait-and-watch. The catalyst gate is NVDA earnings: if NVDA data center beats big, AMD MI300X orders accelerate in the same quarter as cloud providers diversify GPU vendors. Do not add until that signal fires.
AMD Supplier Chain
AMD depends on
TSM
TSMC
Manufactures 100% of AMD CPUs (Genoa/Bergamo) and MI300X AI GPUs on N4/N3 process. No alternative foundry at this node.
Signal: TSMC N4 allocation locked through Q3 2026 → AMD production ceiling removed
AMAT
Applied Materials
Deposition and etch tools used in TSMC fabs where all AMD chips are made. Indirect dependency via TSMC.
Signal: Same TSMC capex cycle that lifts ASML also benefits AMAT
AMD competes with
NVDA
NVIDIA Corporation
Competes directly for cloud AI accelerator budget. MI300X vs H100/Blackwell. NVDA dominates training; AMD gaining in inference.
Watch: NVDA data center beat → cloud providers diversify → AMD MI300X orders follow 1Q lag
INTC
Intel Corporation
x86 server CPU competition (Granite Rapids vs EPYC Genoa/Bergamo). AMD gaining share — Intel competitive pressure declining.
Signal: Intel data center CPU share loss confirmed → AMD EPYC server share growing
Public ticker
Foreign listed
Private
Cloud AI capex cluster
— the demand engine behind NVDA, ASML, and TSM signals
8
META $65B capex → NVDA GPU demand → TSMC CoWoS
Lag: 1–2 quarters · Second-largest cloud AI spender · Primary NVDA demand node
Signal node
META raised FY2026 capex to $64–72B. Every $1B Meta commits flows to NVDA GPUs → TSMC fabrication → ASML EUV machines within 2 quarters. Meta is not a primary recommendation here — it is the single most important demand signal for your NVDA and ASML positions.
META $65B capex
→
Blackwell GPU orders
→
NVDA revenue ↑
→
TSMC CoWoS demand
FY2026 capex raised $64–72B~350K H100 equivalent ordersMeta AI: 1B monthly usersLlama open-source → GPU demand scaling
What to watch: META earnings capex guidance is a leading indicator — if they raise again next quarter, NVDA data center revenue will beat by a wider margin 1–2 quarters later. If Meta cuts capex guidance, it's a negative signal for NVDA. Track Meta quarterly. The $64–72B range already widened upward — the trajectory is bullish for the whole semiconductor cluster.
META Capex Flows To
Primary beneficiaries
NVDA
NVIDIA Corporation
Primary GPU vendor for all Meta AI training. Every Meta capex raise = NVDA Blackwell order.
Signal: $1B META capex → ~$600M NVDA revenue within 2Q
TSM
TSMC
Manufactures all NVDA GPUs that Meta buys. Meta capex → NVDA orders → TSMC CoWoS capacity demand.
Signal: META capex raise → TSMC CoWoS utilisation +2 quarters later
Indirect beneficiaries
ASML
ASML Holding
TSMC CoWoS expansion requires new EUV machines. Meta capex → TSMC expansion → ASML orders (6–9mo lag).
Signal: Meta capex → TSMC capex → ASML order book fills
SMCI
Super Micro Computer
Meta uses SMCI liquid-cooled racks for GPU deployments at scale.
Signal: Meta data center buildout = SMCI rack integration revenue
9
MSFT $80B capex → Azure AI → NVDA GPU demand (largest single buyer)
Lag: 1–2 quarters · Largest cloud AI spender globally · OpenAI exclusivity amplifies demand
Signal node
Microsoft is the single largest NVDA GPU buyer on the planet — $80B capex guided FY2025, OpenAI exclusivity through 2030. Azure is "capacity constrained" per their own earnings call. When MSFT raises capex, NVDA wins.
MSFT $80B capex
→
Azure AI + OpenAI compute
→
NVDA largest order
→
TSMC CoWoS
$80B FY2025 capex — largest cloud AI spenderOpenAI exclusive compute to 2030Azure +33% YoY — capacity constrainedCopilot: 14M enterprise seats
What to watch: MSFT Azure growth rate is the most reliable forward indicator for NVDA data center demand. "Capacity constrained" in the earnings call = more GPU orders imminent. OpenAI exclusivity means every ChatGPT user is indirectly a NVDA GPU consumer. A MSFT capex cut would be a significant negative signal for NVDA — it hasn't happened and shows no sign of happening.
MSFT Capex Flows To
Primary beneficiaries
NVDA
NVIDIA Corporation
Largest single NVDA GPU buyer globally. $1B MSFT capex raise → ~$700M NVDA revenue within 2Q.
Signal: Azure "capacity constrained" → more GPU orders coming
TSM
TSMC
All NVDA GPUs Microsoft buys are fabricated at TSMC. MSFT capex → NVDA → TSMC CoWoS utilisation.
Signal: MSFT capex raise → TSMC CoWoS backlog extends
Indirect beneficiaries
ASML
ASML Holding
TSMC capacity expansion requires EUV machines. MSFT → NVDA → TSMC → ASML order book fills.
Signal: MSFT capex acceleration → ASML order window extends
PLTR
Palantir Technologies
Azure GovCloud is Palantir's primary compute substrate for classified contracts. MSFT Azure = PLTR infra.
Signal: Azure GovCloud expansion = PLTR compute capacity growing
10
GOOGL $75B capex → Google Cloud AI → NVDA + TPU dual signal
Lag: 1–2 quarters · TPU self-build partially offsets NVDA demand · Net positive for TSMC
Monitor
Google's $75B capex is split between NVDA GPU purchases (for third-party Cloud customers) and TPU self-build (for internal Gemini workloads). The TPU build is a long-term NVDA risk for Google-internal workloads, but TSMC manufactures the TPUs — so ASML and TSM benefit regardless. Google Cloud capex growth is still net positive for the semiconductor cluster.
GOOGL $75B capex
→
NVDA (Cloud GPUs)
TSMC (TPU fab)
→
ASML EUV demand
FY2026 capex $75B — raised from $60BTPU Trillium v5: 4x H100 efficiencyTSMC fabricates all TPU Trillium chipsCloud +28% YoY — capacity demand rising
What to watch: Google's TPU self-build is a long-term structural risk for NVDA's Google revenue share — but it is not a near-term threat. Today, Google Cloud still sells NVDA GPU instances to enterprise customers and buys NVDA for third-party model training. The nuance: TSMC manufactures both the TPUs and the NVDA GPUs Google buys — so ASML and TSM benefit from Google capex regardless of the GPU vs. TPU split. Monitor TPU efficiency claims vs. actual NVDA GPU procurement data in quarterly filings.
GOOGL Capex Flows To
Benefits regardless
TSM
TSMC
Manufactures BOTH Google's TPUs and the NVDA GPUs Google buys for Cloud customers. GOOGL capex = TSMC revenue regardless of GPU/TPU mix.
Signal: GOOGL capex raise → TSMC CoWoS + TPU wafer demand
ASML
ASML Holding
TSMC needs EUV machines to manufacture both TPUs and NVDA GPUs. GOOGL capex → TSMC capacity → ASML orders.
Signal: GOOGL infrastructure growth = TSMC fab demand = ASML order book
Conditional on GPU vs TPU
NVDA
NVIDIA Corporation
Google buys NVDA GPUs for Cloud enterprise customers and third-party model training. TPU self-build reduces but doesn't eliminate NVDA dependency.
Watch: NVDA GPU procurement in GOOGL annual filings vs. TPU efficiency claims
AMD
Advanced Micro Devices
Google uses AMD MI300X for some Cloud AI inference workloads as NVDA alternative. Smaller allocation vs NVDA.
Monitor AMD share of Google Cloud GPU capacity disclosures
Energy infrastructure — AI power demand
11
Cloud AI capex → +2.5GW per $1B → VST Texas power fleet
Lag: 6–18 months · Druckenmiller position · Active window
Watch
AI data centers consume power at a ratio of ~2.5GW per $1B of cloud AI capex. Meta, Microsoft, and Google have collectively committed $220B+ in FY2026 capex — that's a structural demand curve the grid wasn't built for. Vistra's Texas ERCOT fleet (nuclear + natural gas) is geographically positioned in the primary US data center build corridor. CEO and CFO bought open market in March 2026. Druckenmiller disclosed 2.1M share position in Q1 2026 13F.
META $65B capex
MSFT $80B capex
GOOGL $75B capex
→
+550GW demand curve
→
VST Texas grid
MSFT $80B capex FY2026 — #1 globallyMETA $65B capex raised from $40BAI data center = 2.5GW per $1B capex spendDruckenmiller: 2.1M VST shares (Q1 2026)
What to watch: Track cloud giant quarterly capex guidance. Every raise is a direct leading indicator for power demand. The signal lag is 6–18 months — data centers announced today come online next year. Vistra Q1 2026 already beat estimates and raised guidance. Two undisclosed data center PPAs reportedly in negotiation. Texas ERCOT summer peak pricing will be the short-term catalyst.
AI Power Demand Beneficiaries
Primary Texas beneficiary
VST
Vistra Energy
Largest competitive power generator in Texas ERCOT. Nuclear + gas fleet positioned at data center build-out epicenter. CEO + CFO buying.
Signal: Cloud AI capex raise → VST power demand → ERCOT pricing
Nuclear clean power PPAs
CEG
Constellation Energy
Largest US nuclear operator. Three Mile Island restarted for Microsoft. Two additional cloud provider PPAs in active negotiation. 24/7 carbon-free baseload = premium pricing.
Signal: AI carbon-free demand → nuclear PPA scarcity → CEG pricing power
12
AI carbon-free mandate → nuclear PPA scarcity → CEG pricing power
Lag: Contract signing to earnings · 20-year PPA lock-in · Active now
Watch
Major tech companies have committed to 24/7 carbon-free energy for their data centers. The problem: the only available 24/7 carbon-free baseload is nuclear power. Solar and wind are intermittent. Constellation holds the largest US nuclear fleet (21 reactors, 21GW) and just restarted Three Mile Island Unit 1 specifically for a 20-year Microsoft PPA at $160/MWh — a significant premium to spot prices. CEO Joe Dominguez disclosed two additional cloud provider PPA negotiations on the Q1 2026 call. Nuclear PPA demand materially exceeds available supply.
MSFT net-zero mandate
→
24/7 carbon-free required
→
CEG nuclear fleet
→
20yr PPA $160/MWh
MSFT TMI PPA: 835MW · 20 years · $160/MWhNuclear = only viable 24/7 carbon-free baseloadCEO Dominguez: $2.1M open market buy2 additional cloud provider PPAs in negotiation
What to watch: Each new nuclear PPA announcement is a re-rating catalyst for CEG. Goldman PT is $240 vs current ~$191 — the upside case is more PPAs signed at premium prices. The key risk: new nuclear capacity (SMRs, new builds) is 10+ years away — CEG is supply-constrained in the best possible way. Monitor Google, Amazon, and Meta carbon-free energy commitments for additional PPA pipeline signal.
Nuclear PPA Demand Chain
PPA signed
MSFT
Microsoft Corporation
20-year PPA for Three Mile Island Unit 1 restart. 835MW dedicated to Microsoft Azure data centers. First nuclear restart in US history for a single corporate buyer.
Signal: TMI PPA = template. MSFT capex raise → more nuclear PPAs
Pipeline candidates
GOOGL
Alphabet (Google)
Google has committed to 24/7 carbon-free energy for all data centers by 2030. Currently sourcing nuclear in Europe — US nuclear PPA likely.
Watch: GOOGL 24/7 CFE procurement disclosures in sustainability filings
META
Meta Platforms
Meta committed to net-zero by 2030. $65B capex = massive power footprint. Nuclear PPA would be transformational for both Meta and CEG.
Monitor Meta energy procurement announcements — nuclear PPA would move CEG
TSM
→
Capex raise
$38–42B FY2026
$38–42B FY2026
→
AMAT
8–12 week lag
TSMC capex raise → WFE tool orders → AMAT + LRCX order book
Lag: 8–12 weeks · Window opened: March 2026 · Status: Active
TSMC capex $38–42B raised Jan 2026N2 node requires 30% more deposition stepsCEO Dickerson: demand signals strongest since 2021Goldman PT $185→$210
What to watch: Every TSMC capex raise directly flows to AMAT and LRCX order books 2–3 quarters later. The Jan 2026 raise opened the window in March 2026 — we are inside it now. Monitor TSMC Q2 2026 earnings (Jul 17) for any capex revision. The key risk: China export restriction escalation could delay fab buildout and push out the lag window.
WFE Equipment Chain
Primary beneficiaries
AMAT
Applied Materials
Deposition and etch tools. Required for every advanced fab TSMC builds at N3/N2. Largest WFE company by revenue.
Signal: TSMC capex raise → AMAT tool orders 2–3Q lag
LRCX
Lam Research
Etch equipment leader. TSMC N2 node uses Lam etch tools. Also dominant in HBM3E memory fab etch — double exposure to AI capex.
Signal: Same TSMC capex wave + HBM3E etch intensity 40% higher than DRAM
Parallel trigger
ASML
ASML Holding
EUV lithography. Same TSMC capex cycle drives ASML order book — AMAT and LRCX are the non-lithography complement.
Signal: TSMC builds one ASML EUV machine and multiple AMAT/LRCX tools per fab line
NVDA
→
Datacenter beat
$35.6B Q1 2026
$35.6B Q1 2026
→
MU
1–2Q lag
NVDA datacenter beat → HBM order expansion → Micron revenue confirmation
Lag: 1–2 quarters · Most reliable lag in platform · Active now
NVDA Q1 2026 datacenter $35.6B beatHBM supply sold out through 2027CEO Mehrotra $2.8M open market buy$28M call sweep May 6
What to watch: This is the most reliable upstream lag signal in the platform. NVDA data center beat → MU HBM contract expansion has fired consistently for 6 consecutive quarters. Current lag window opened after NVDA Q1 2026 (May 22). Micron Q3 FY2026 earnings Jun 25 is the confirmation catalyst. The key risk: SK Hynix gaining incremental HBM share from MU — monitor quarterly allocation splits.
HBM Memory Chain
HBM suppliers (duopoly)
MU
Micron Technology
HBM3E supplier to NVIDIA. 6-quarter order book locked. Every Blackwell GPU = 8 HBM stacks. Margins transforming to 50%+.
Signal: NVDA data center beat → MU HBM expansion 1–2Q lag
Etch equipment beneficiary
LRCX
Lam Research
Every HBM3E stack manufactured requires Lam etch tools — 40% higher etch intensity than standard DRAM. HBM production ramp = Lam tool orders.
Signal: MU + SK Hynix HBM capacity expansions both require Lam etch
DoD Budget
→
+6% real growth
FY2027 request
FY2027 request
→
LMT
0–12 month lag
Defense budget elevation → prime contractor revenue → PLTR AI platform embedding
Lag: 0–12 months for prime awards · PLTR follows with 1–3Q lag · Active now
DoD FY2027 +6% real growthNATO 2% GDP commitment: +$200BLMT CEO+CFO $2.5M buyPLTR embedded in all three primes
What to watch: Defense budget increases flow to prime contractors (LMT, NOC, RTX) within 6–12 months via contract awards. PLTR is embedded as the AI/data platform across all three — PLTR US Government growth follows defense prime program growth with 1–3 quarter lag. Monitor quarterly PLTR US Government revenue for confirmation. The key risk: Continuing resolution (CR) budget environment delays new program starts — watch Congress appropriations calendar.
Defense Prime Chain
Prime contractors
LMT
Lockheed Martin
F-35 + hypersonics. NATO allied spending creates structural demand floor for decade. Sole-source on critical programs.
Signal: NATO 2% commitment → F-35 allied orders → LMT revenue floor lift
NOC
Northrop Grumman
B-21 Raider + Sentinel ICBM — two of three nuclear triad legs. Budget-protected, no competitive risk.
Signal: Nuclear triad modernization is sequestration-proof — NOC revenue locked for decade
RTX
RTX Corp
Patriot + AMRAAM + SM-3. Ukraine/NATO demand created 2yr+ production backlog across all missile lines.
Signal: NATO restocking + Ukraine replenishment → RTX missile backlog 2yr+
Downstream AI beneficiary
PLTR
Palantir Technologies
AI/data platform embedded across LMT F-35 logistics, NOC B-21 mission planning, RTX targeting systems. US Government revenue (42% of total) grows with defense prime program growth.
Signal: Defense prime program growth → PLTR AI embedding → US Gov revenue 1–3Q lag
MSFT
→
Azure +33%
AI compute capacity
AI compute capacity
→
NOW
2–6 month lag
Cloud AI capex → enterprise compute availability → CRM/NOW/SNOW AI adoption
Lag: 2–6 months · Active now · MSFT + GOOGL both expanding
MSFT Azure +33% · $80B capex FY2026GOOGL Cloud +28% · $75B capexNOW CEO+CFO $5M cluster buyCRM Agentforce 3,000 customers in 90 days
What to watch: Enterprise AI adoption is 2–3 years behind consumer AI. Cloud capacity (MSFT Azure, GOOGL Cloud) expanding creates the compute infrastructure that enterprise software runs on. ServiceNow AI Pro, Salesforce Agentforce, and Snowflake Cortex all require cloud AI compute — when MSFT/GOOGL expand capacity, enterprise AI adoption accelerates 2–6 months later. The key risk: Enterprise AI ROI scepticism — if enterprises delay AI Pro upgrades, the pipeline-to-ARR conversion slows.
Enterprise AI Software Chain
Enterprise AI beneficiaries
NOW
ServiceNow
Highest enterprise AI attach rate in SaaS. AI Pro adds $120/user on top of 98% renewal base. Azure compute availability drives adoption.
Signal: MSFT Azure capacity → ServiceNow AI Pro inference → attach rate revenue
CRM
Salesforce
Agentforce AI agents run on Azure compute. CRM data moat (150K customers, decades of sales data) is the differentiator — AI is only as good as the data it runs on.
Signal: Azure AI capacity → Agentforce inference → pipeline-to-ARR conversion
SNOW
Snowflake
Cortex AI runs AI inference on enterprise data where it already lives. MSFT Fabric + GOOGL BigQuery partnerships mean cloud expansion brings Snowflake customers.
Signal: MSFT+GOOGL cloud expansion → Cortex AI compute available → consumption revenue
Upstream triggers
MSFT
Microsoft
$80B FY2026 capex. Azure is the primary cloud for ServiceNow and Salesforce. Azure +33% YoY = more compute for enterprise AI workloads.
Signal: MSFT capex → Azure AI capacity → NOW/CRM AI feature adoption
GOOGL
Alphabet (Google)
$75B capex. Google Cloud +28% YoY. Snowflake + BigQuery partnership makes GOOGL capacity directly available to Snowflake Cortex AI workloads.
Signal: GOOGL Cloud growth → BigQuery/Snowflake adoption → SNOW Cortex consumption
Research implications
— stocks your chains imply you should investigate
From: NVDA HBM memory chain
MU — Micron Technology
Every NVDA Blackwell requires HBM. Micron revenue follows NVDA data center beat 1–2Q lag.
RESEARCH 1–2Q lag
From: NVDA EDA design chain
SNPS — Synopsys
Every NVDA chip designed on Synopsys tools. Design duopoly moat.
RESEARCH Structural
From: ASML fab buildout chain
AMAT — Applied Materials
Same fab capex driving ASML also drives AMAT. Parallel beneficiary.
RESEARCH 6–9mo lag
From: NVDA server build chain
SMCI — Super Micro Computer
40%+ revenue from NVDA ecosystem. High beta to data center beat.
RESEARCH Earnings lag
From: PLTR government ecosystem
BAH — Booz Allen Hamilton
PLTR's primary implementation partner. BAH wins often run PLTR underneath.
RESEARCH Structural
From: NVDA power infrastructure chain
VST — Vistra Energy
Druckenmiller's largest energy position. Active lag window — research entry now.
RESEARCH 6–9mo entry
Signal layers ranked by time horizon — L1 is most immediate. Use the stock selector above to focus on one ticker.
L1
Market structure · 0–5 day signals
Live signals · refreshes automatically
Dark pool clustering — institutional accumulation
Scanning volume anomalies…
Silent accumulation = elevated volume + price flat = institutions building a cost basis. Breakout typically follows within 30 days, 60–70% of the time.
GEX — gamma exposure price gravity
Computing price structure…
Positive GEX = dealers long gamma = buy dips, sell rips = low-vol drift higher. RSI + MA position is the free proxy: above MA20 with RSI 45–65 = positive GEX regime.
L2
Insider / regulatory · 5–30 day signals
Recent insider transactions — live
Loading insider data…
Insider clusters — buy signals detected
Loading cluster analysis…
L3
Corporate intel · 14–60 day signals
Analyst moves — last 14 days
Loading analyst intelligence…
Options flow — institutional pre-positioning
Loading pre-earnings windows…
High convergence score + earnings within 21 days = classic pre-positioning window. Institutional options sweeps concentrate here — rising score momentum is the free proxy.
L4
Supply chain / contracts · 30–120 day signals
Live chain propagation — anchor moves today
Loading supply chain signals…
Structural chain context — always relevant
NVDA AI compute supply chain
ASML EUV ↗
→
TSMC CoWoS ↗
→
NVDA Blackwell ↗
→
Cloud Giants ↗
TSMC CoWoS capacity constraint easing (Q3 signal)
CoWoS advanced packaging was the Blackwell bottleneck. TSMC adding 30% CoWoS capacity in Q3. Directly unlocks NVDA delivery timelines — 6–9 month revenue pull-forward.
TSMC Q1 capex +15% → ASML orders in 6–9 months
Historical lag: TSMC capex raise precedes ASML EUV orders by 6–9 months. Q1 raise = orders arriving now. ASML order book confirmation expected Q3 earnings.
ASML upstream supply chain
Zeiss Optics
→
Trumpf Laser
→
ASML EUV ↗
→
TSM ↗ / INTC ↗
No ASML EUV alternative exists below 7nm
Zeiss produces the only optical mirrors capable of EUV precision. Trumpf makes the only viable EUV light source. Both are sole-source suppliers to ASML — creating a 3-layer supply chain moat no competitor can replicate in under 10 years.
China restriction = demand redistribution, not loss
DUV restrictions to China are headwind. But demand doesn't disappear — it redistributes to TSMC, Intel, Samsung who absorb with their own capex plans. Net ASML demand is structurally intact.
TSMC + SMCI — new cluster signals
TSM capex raised to $38–42B — ASML and NVDA lag window active
TSMC Q1 2026 capex raise is the strongest leading indicator in the semiconductor cluster. History: TSMC capex raise in Jan 2025 → ASML order book expanded 2Q later. The same window is now open again. Every dollar TSMC commits to capex flows through ASML machines, NVDA production, and AMD capacity within 1–3 quarters.
SMCI — CEO + CFO open market buy cluster (Apr 2026)
CEO Charles Liang bought $2.8M and CFO David Weigand bought $380K on the open market in April — their first purchases since the accounting crisis began. Two senior executives independently buying with personal cash after a crisis = management conviction the worst is behind them. Accounting resolved March 2026. Same pattern as ASML insiders before the Jan 2025 rally.
SMCI — highest beta to NVDA data center beat
~40% of SMCI revenue ties to NVDA GPU ecosystem. If NVDA beats on data center next earnings, SMCI order book expands the same quarter — same-period correlation, not lagged. Speculative position only: size small in your high-risk tranche, with NVDA earnings as the explicit catalyst gate.
Government contracts · SAM.gov
PLTR · RKLB — contract awards as leading revenue indicator
PLTR — Palantir Technologies
Click Fetch to load SAM.gov contract awards
RKLB — Rocket Lab USA
Click Fetch to load SAM.gov contract awards
SAM.gov requires a free API key —
L5
Macro / structural · 60–365 day signals · see Global Markets for live FRED data
Structural tailwinds — multi-year duration
AI capex supercycle — $300B+ FY25 guided
MSFT $80B + GOOGL $75B + META $65B + AMZN $75B. Each dollar flows through NVDA → TSMC → ASML. Duration: 3–5 years minimum. Every cloud giant capex raise is an ASML and NVDA order.
Sovereign AI wave — $100B+ national AI budgets
UAE, Saudi Arabia, Japan, India each announced $100B+ AI infrastructure. New demand vector independent of US cloud giants — NVDA hardware + PLTR software are the two infrastructure layers every sovereign program needs.
Space + defense spending acceleration (SDA, NRO)
Space Development Agency Tranche 2 awards Q3. NRO constellation expansion. RKLB and PLTR are the two small-cap pure-plays. Both structurally positioned before institutional coverage broadens.
Macro risks — position sizing factors
Yield curve — monitor 10Y-2Y spread
Extended inversion historically precedes tech multiple compression by 6–9 months. Re-steepening = positive catalyst. Use as position sizing input — not sell signal unless >12 months inverted.
EUR/USD at 1.08 — ASML USD revenue headwind
~4% revenue haircut on USD reporting. Use currency weakness as entry opportunity, not thesis change. European operations = natural hedge. Strong USD is temporary FX noise.
US defense budget trajectory — secular growth
NATO 2% GDP minimum target being met or exceeded. US defense budget up 3.2% FY25. AI-enabled warfare (PLTR) and small satellite constellations (RKLB) are the two highest-growth pockets inside the defense budget.
L1 0–5d · Market structure
L2 5–30d · Insider / regulatory
L3 14–60d · Corporate intel
L4 30–120d · Supply chain / contracts
L5 60–365d · Macro / structural
Earnings calendar · Next 90 days
Live from Finnhub · EPS estimates + beat/miss history
📅
Loading earnings calendar…
Fetching upcoming reports from Finnhub
Before earnings — what to watch
Options sweep 1–3 weeks before
Large call sweep ($10M+) before earnings = institutional conviction. Most reliable pre-earnings directional signal available.
Analyst estimate acceleration
3+ raises in 21 days from multiple banks = positive channel checks. Analysts are getting confirming signals from the supply chain.
Insider selling = ignore (usually)
Pre-scheduled 10b5-1 plans are noise. Only flag open-market sells (rare) or same-week buys from C-suite.
After earnings — reaction framework
Never hold through if implied move >15%
At ±15%+ implied move, options premium is too rich. Better to build position post-event on any dip than to gamble on direction.
Guidance > beat on the quarter
A beat with flat guidance = sell the pop. A miss with raised guidance = buy the dip. Forward guidance is the only number that matters.
Supply chain implications
After an anchor stock beats, check CHAIN_GRAPH for which downstream stocks enter their lag window. That's your next 4–8 week setup.
NVDA · May 22 after close
17 days away · Beat rate 78% over last 8 quarters
±11.4%
Implied move
±8.2%
Historical avg
$24.5B
Revenue est.
The one number that matters
Data Center revenue est $21B+. Miss here = -15%+ regardless of other metrics.
Watch the Blackwell supply language
"Constraints easing" = neutral. "Constraints persisting" = -5%, even on revenue beat.
Q3 guidance is the real catalyst
Above $26B = +10%+. In-line = flat. Below consensus = -8%.
Pre-earnings signal stack + strategy
Options + dark pool — Bullish
$47M call sweep at $600 strike May 23 + $180M dark pool above VWAP. Two independent institutional signals.
Estimate acceleration — Bullish
6 raises in 21 days, avg +8.4%. Analysts getting positive channel checks from supply chain.
Strategy — 50% now, 50% post-earnings
$125 now. Reserve $125 for post-result. Data center beats AND Q3 guidance >$26B → deploy. Miss → wait 48h.
Live insider activity · SEC EDGAR Form 4
No API key required · Public data · 6-month lookback
Pull live Form 4 filings directly from SEC EDGAR — no key required
Notable fund managers tracked
Who holds this stock — live 13F data
Institutional ownership · SEC 13F filings · Updated quarterly
Select a ticker above to load institutional holders
My watchlist
Community — most watched
Signal scores for your watchlist
$1,000
Weekly budget
$0
Deployed
$1,000
Remaining
0
Trades this week
Everything is free. Pro is coming.
Real-time signals · Email alerts · Portfolio tracking
⚡ Real-time signals
🔔 Email alerts
📊 P&L tracking
🔌 API access
Watchlist
My watchlist
Community picks
Signal scores
Positions
—
Portfolio value
—
Unrealised P&L
—
Invested
$0
Realised P&L
Open positions — click ticker · Sell to close
No positions yet — add below
Add position manually
Trade history
No trades yet
Portfolio rules check
Max 12 positions—
Top 3 at 40–50%—
Budget
$1,000
Weekly budget
$800
Deployed this week
$200
Remaining
3
Trades this week
This week's allocation
AS
ASML$350
NV
NVDA$250
RK
RKLB$200
Set budget
Suggested split — based on signal scores
Signal score weighting: higher conviction = larger allocation. Rebalanced weekly.
ASML
Score 9135% · $350
NVDA
Score 7825% · $250
RKLB
Score 6720% · $200
PLTR
Score 6215% · $150
Cash
Reserve5%