"๐ฐ๏ธ $ASTS launch cadence takes a hit: Blue Origin's New Glenn exploded during static fire at LC-36 (May 28), following the NG-3 upper-stage failure that lost BlueBird 7 in April. New Glenn is now out for 2026. ASTS falls back on Falcon 9 (3 sats/launch vs 8), putting the 45-60 sat year-end target in serious doubt. Slip to 2027 likely. $ASTS"
๐ช๐บ European Photonics & Semis โ Ranked by Risk/Reward
1. $XFAB โ ~โฌ9
The new thesis. Specialty foundry at the crossroads of silicon photonics, SiC/GaN and MEMS. Building InP-on-SOI platform for AI optical interconnects. Auto drag = entry window. Production ramp 2027.
Buy zone: โฌ8.50โ9.50
2. $SOI โ ~โฌ150
Near-monopoly on SOI substrates. Every silicon photonics chip needs Soitec wafers. Structural moat is unmatched. Has run hard โ wait for it to come to you.
Buy zone: โฌ120โ140
3. $RPI โ ~700p
Record 2025 revenue, 47% chip sales surge. Agentic AI edge compute demand is the secular tailwind. Most liquid name in the group.
Buy zone: 580โ630p
4. $LPK โ ~โฌ24
Glass core substrates near-monopoly. Advanced packaging demand driven by AI chiplet architecture. Pulled back from highs โ thesis intact.
Buy zone: โฌ20โ22
5. $IQE โ ~50p
Epiwafer supplier to the photonics stack. Thesis validated by $MTSI investment. Big move already done โ add on weakness only.
Buy zone: 40โ45p
6. $ALRIB โ ~โฌ15
MBE systems duopoly. Quantum + photonics equipment demand driver. Already up massively from thesis entry โ patience needed.
Buy zone: โฌ12โ13
7. $SIVE โ ~SEK 70
CPO laser chokepoint. Thesis fully played ($4โ$71). Pipeline up 64% YoY. Hard to chase โ hold existing, trim into strength.
Buy zone: SEK 55โ60
Not financial advice. Small/micro-cap European names โ size positions accordingly
๐ $SPCX S-1 JUST DROPPED โ Tonight. History.
The Deal
๐ฐ Raising up to $75B | Val: $1.75Tโ$2T+
๐ Nasdaq | Ticker: $SPCX | Jun 12 listing
๐ฆ 21 banks | 30% retail allocation (3x normal)
Financials (First Ever Public Look)
๐ Q1 rev: $4.69B (+15% YoY)
๐ 2025 rev: ~$16B | Starlink: $11.4B
๐ Starlink op profit: $4.4B ๐
๐ 2025 net loss: $4.94B (xAI capex)
๐ 2025 capex: $20.7B (5x prior year)
๐ 2026 EBITDA target: $14B
What You're Actually Buying
๐ธ World's #1 rocket company
๐ก 10M Starlink subscribers (growing to 16.8M by EOY)
๐ค xAI + Grok + Orbital AI Data Centers
๐ก๏ธ NASA + DoD contracts
Control
๐ Musk controls 85% of votes
๐ณ๏ธ Public gets Class A โ economic stake, zero governance
Timeline
May 20 โ S-1 public โ
Jun 4โ8 โ Roadshow
Jun 11 โ Pricing
Jun 12 โ First trade
The Risk
110x revenue multiple. Profitability not yet proven at scale. You're betting on 2030+, not 2026.
*Not financial advice. Read the S-1.*
\#SPCX #SpaceX #IPO #Starlink #xAI #SPCXIPO
$NVDA delivered ๐
โข Revenue $81.6B vs Est. $79.2B
โข EPS $1.85 vs Est. $1.78
โข Data Center $75.2B vs Est. $73.5B
โข Gross Margin 75% vs. Est. 75%
Q2 Guidance
โข Revenue $90.6B vs Est. $87.2B
โข Gross Margin 75% vs. Est. 75%
RPO & Backlog = the only metric that cannot be faked ๐
Contracted. Non-cancellable. Already locked in.
The league table:
๐ฅ $MSFT โ $627B RPO, 2.2 years of revenue locked
๐ฅ $ORCL โ $523B RPO, +438% YoY, $300B OpenAI deal
๐ฅ $GEV โ $163B backlog, +71% organic, 4+ years locked
$CRM โ $72B RPO, Agentforce accelerating
$CAT โ $62.7B backlog, +79% YoY, decoupled from cycle
$ETN โ 11 YEARS of data centre backlog ๐คฏ
$NOW โ $27.7B RPO, AI ACV +130%
$FIX โ $11.9B backlog, doubled in 12 months
$PLTR โ FCF doubling, US commercial +137%
$PWR โ $2.4T addressable, 5-10yr grid contracts
The insight:
When backlog grows faster than revenue
Future growth is ACCELERATING not slowing
That gap = the most bullish signal in investing ๐ฅ
These aren't estimates.
They're obligations. ๐"
The CPO laser chokepoint is the most underrated bottleneck in AI infrastructure ๐ฌ
$NVDA locked up $COHR + $LITE.
Everyone else scrambled upstream.
The stack:
โก Lasers: $SIVE + Win (independent Tier 1)
๐ฎ Architecture: $MRVL (Celestial AI CPO)
๐ญ Foundry: $GFS + $TSMC + $TSEM
๐ Overflow: $AAOI $POET
Warning shot: $MRVL cancelled $POET on NDA breach.
Hyperscalers WILL vertically integrate.
Own the independent chokepoints they CAN'T absorb.
A $130M Swedish company $SIVE sits at the center. ๐ธ๐ช
The Complete AI Infrastructure Stack ๐๏ธ
Most investors only own Layer 1
The real alpha is Layers 2 + 3 ๐
Layer 1 โ GPU Compute โ Crowded, priced in
$NVDA $AVGO $MRVL
The factory workers doing heavy lifting
Layer 2A โ CPU Orchestration โก Re-rating NOW
$INTC $AMD $ARM $QCOM
GPU:CPU ratio shifting 1:8 โ 1:1 in agentic AI
= 8x more CPUs needed per rack ๐ฅ
Layer 2B โ HBM Memory ๐พ 18-month backlog
$MU โ only US HBM manufacturer
70% of ALL memory chips now going to data centres
SK Hynix + Samsung = the other two
Layer 2C โ Storage ๐ Evercore PT $1,200
$SNDK $WDC
Every AI inference call reads/writes memory
Agentic AI multiplies cycles exponentially
Layer 2D โ Power + Cooling โก 10-100x power vs traditional servers
$VRT $ETN $PWR $SMCI
Every MW of AI compute needs this layer
Layer 3 โ Optical Interconnect ๐ฌ Almost nobody owns this
$GLW $COHR $LITE $CRDO $FN $AXTI $ONTO
The nervous system connecting everything
Moving photons = 5-20x less power than electrons
$GLW alone just confirmed $18B in contracted deals TODAY
The alpha rule:
Layer 1 = $2.7T market cap, priced for perfection ๐ฌ
Layer 2 = still being discovered ๐
Layer 3 = almost completely ignored ๐จ
The further down the stack you go
The less crowded the trade
The more compressed the valuation
The bigger the upside ๐
Not financial advice. DYOR ๐ค
The most important earnings night of 2026 ๐จ April 29 prediction:
$META $GOOGL $MSFT $AMZN all report after close TODAY
Beat probabilities:
๐ฅ $AMZN โ 94% beat (Polymarket), AWS 26-27% guided, $20B custom chip run rate
๐ฅ $META โ 95% beat, ad revenue 31%+, Advantage+ at $60B run rate, 0 Sell ratings
๐ฅ $GOOGL โ 90% beat, Cloud accelerating 48%โ50%+, stock doubled in a year
4๏ธโฃ $MSFT โ 85% beat, Azure 35-37%, Copilot seat count = the wildcard number
Expected moves:
๐ $GOOGL +8-15% โ biggest gap up potential
๐ $META +6-12% โ zero sellers, beats every quarter
๐ $MSFT +5-15% โ if Copilot revenue explicitly disclosed
๐ $AMZN +5-10% โ tempered by FCF compression
But earnings numbers matter LESS than this:
๐ฅ $600-645B combined AI capex guidance
๐ฅ Any raise = NVDA INTC CRDO MRVL AVGO all gap up Thursday
๐ฅ Any softening = rotate to defensives immediately
The real trade:
Own the earnings fear TODAY ๐
Sell into the capex raise Thursday morning ๐
$NVDA $INTC $MRVL $CRDO $SMH $SOXX all follow if capex holds ๐ฅ
Not financial advice. DYOR ๐ค
Monte Carlo Capital thesis VALIDATED by earnings ๐ฅ
The market is waking up to 3 mandatory AI layers:
Layer 1 โ GPU Compute (already priced in)
$NVDA $AVGO $MRVL โ everyone owns this
Layer 2 โ CPU Orchestration
(the RE-RATING wave NOW)
๐ฅ $INTC โ +120% YTD, 25yr record, EPS beat by 2,800%, supply constrained, Google + NVDA partnerships confirmed
๐ฅ $AMD โ PT $220โ$375, CPU+GPU dual cycle, reports May 5
๐ฅ $ARM โ royalty on EVERY AI chip shipped globally
4๏ธโฃ $QCOM โ sleeper pivot, 6% FCF yield, edge AI CPU undervalued
Layer 3 โ Optical Interconnect (the "plumbing" becoming MANDATORY)
5๏ธโฃ $CRDO โ 272% revenue growth, $750M DustPhotonics acquisition, PT $220
6๏ธโฃ $AAOI โ 438% annual return, $1B revenue guided, demand 10x by mid-2027
Plus the hyperscaler pair:
7๏ธโฃ $AMZN โ $25B Anthropic deal, AWS fastest growth in 13 quarters, reports Wed
8๏ธโฃ $META โ 0 Sell ratings, $80B cash, reports Wed
The alpha insight:
Most investors own Layer 1 heavily ๐
The money is being made in Layers 2 + 3 RIGHT NOW ๐
CPU ratio shifting from 1:8 โ 1:1 with GPUs in agentic AI
That's 8x more CPUs per rack needed ๐
Not financial advice. DYOR ๐ค