Outlier Capital
@outliercapx
The Founder.
Asymmetric stocks. Generational wealth.
Portfolio target: 5x+ returns in 5 years.
Join 4K+ readers → Full portfolio + Theses, link.
142 Following 8.1K Followers
$LITE $COHR $AAOI $SIVE
$AXTI just confirmed InP demand is converting into real revenue and margins, and the entire optics stack is the beneficiary.
> Q3 revenue $73.4M (+162% YoY, +54% sequential) vs ~$66M guidance.
> Adj. EPS $0.41 vs $0.31 est.
> GAAP GM exploded to 56.0% (+3,370 bps YoY, +11.1 pts sequential).
> Operating margin ~38.7%.
Management said they surpassed the goal of doubling InP capacity and more than quadrupled the 2026 InP revenue opportunity via productivity and higher-value large-diameter substrates.
This is the upstream materials layer of the AI photonics/CPO cycle.
Strong InP demand + capacity expansion + margin leverage means the laser and optical-engine buildout is real and accelerating.
It feeds directly into the names that sell the lasers, modulators, and CPO engines.
Adds to the strength already visible across $LITE $COHR $AAOI $SIVE.
The optics stack is no longer a narrative, it’s showing up in shipments and margins.
Position accordingly. Optics is the next Memory.
Show more
$AMZN guiding >40% AWS growth in both 2027 and 2028.
2027 is inflection point for hyperscalers
2028 is inflection point for neoscalers and the neocloud guppies who survive
$amzn $nbis
$AMZN just printed another clean CONFIRMATION of the AI Infra supercycle.
> AWS +39% YoY to $46.0B ($184B annualized run rate).
> CapEx $77.3B vs ~$57B est.
> Management guiding >40% AWS growth in both 2027 and 2028.
Demand is not the constraint. Capacity is!
They’re spending at a pace that implies the multi-year buildout is locked in, with the cash-flow hit already showing up in TTM FCF of –$42.4B. Classic infrastructure cycle math: front-load CapEx, monetize as capacity comes online.
This is the same signal as the rest of the hyperscaler + SpaceX prints.
The AI Infra trade (power, photonics/CPO, cooling, custom silicon, memory, etc.) just got another multi-year visibility upgrade.
Adds to the prints from $SPCX, $META, $MSFT, $GOOG. All systems go.
Bullish!
Show more
$AMZN just printed another clean CONFIRMATION of the AI Infra supercycle.
> AWS +39% YoY to $46.0B ($184B annualized run rate).
> CapEx $77.3B vs ~$57B est.
> Management guiding >40% AWS growth in both 2027 and 2028.
Demand is not the constraint. Capacity is!
They’re spending at a pace that implies the multi-year buildout is locked in, with the cash-flow hit already showing up in TTM FCF of –$42.4B. Classic infrastructure cycle math: front-load CapEx, monetize as capacity comes online.
This is the same signal as the rest of the hyperscaler + SpaceX prints.
The AI Infra trade (power, photonics/CPO, cooling, custom silicon, memory, etc.) just got another multi-year visibility upgrade.
Adds to the prints from $SPCX, $META, $MSFT, $GOOG. All systems go.
Bullish!
Show more
$MRVL x $LITE
What the sellside thinks:
Stifel: Reiterates $MRVL Buy, $370 PT. The strategic logic is strong:
> Extreme codesign across AI fabrics.
> Scale-up integrated optics.
> Build a scaled U.S. transceiver supplier amid potential Chinese restrictions.
> Lock in laser supply (the real bottleneck).
Bernstein: A deal of this size “likely can’t be funded entirely with cash.”
> Dilution + integration risk is real.
Morgan Stanley: Marvell’s DSP client base includes optical module makers that compete directly with $LITE. This could undermine the supplier-neutral reputation $MRVL relies on.
> Could create a major potential conflict of interest that may put its current DSP revenue at risk.
Market reaction was clear: LITE +9–11% AH, MRVL -8%.
The bottleneck thesis doesn’t change.
Whoever controls the laser controls scale-up.
It is very important to pick the winners now.
Show more
A lot of you have been asking me to weigh in on the $MRVL / $LITE rumor.
IMO, the leak probably killed the deal (if not closed yet). $MRVL has been getting hit while $LITE has ripped higher, so any premium and the dilution just got meaningfully worse. Hard to see that deal closing on attractive terms now.
I think the idea makes sense, securing InP lasers is a pretty smart hedge if you’re worried about the CPO transition hitting your DSP franchise.
Feels like a cleaner and more realistic target at this point could be $AAOI…
Show more
Another question I am getting a lot:
How would the $MRVL / $LITE deal actually affect $COHR $AAOI and $SIVE?
It would be extremely bullish.
Lumentum is sold out into 2029. Once that InP capacity is captive inside Marvell, open-market laser supply gets tighter overnight.
> $COHR becomes Nvidia’s clear preferred independent source.
> $SIVE’s pure-play capacity turns into an even more scarce asset.
> $AAOI’s own laser fab becomes even more valuable.
Fewer lasers on the open market = higher ASPs for whoever still controls capacity.
The deal for the rest of the InP stack, it’s a supply-side win.
Position accordingly.
Show more
A lot of you have been asking me to weigh in on the $MRVL / $LITE rumor.
IMO, the leak probably killed the deal (if not closed yet). $MRVL has been getting hit while $LITE has ripped higher, so any premium and the dilution just got meaningfully worse. Hard to see that deal closing on attractive terms now.
I think the idea makes sense, securing InP lasers is a pretty smart hedge if you’re worried about the CPO transition hitting your DSP franchise.
Feels like a cleaner and more realistic target at this point could be $AAOI…
Show more
A lot of you have been asking me to weigh in on the $MRVL / $LITE rumor.
IMO, the leak probably killed the deal (if not closed yet). $MRVL has been getting hit while $LITE has ripped higher, so any premium and the dilution just got meaningfully worse. Hard to see that deal closing on attractive terms now.
I think the idea makes sense, securing InP lasers is a pretty smart hedge if you’re worried about the CPO transition hitting your DSP franchise.
Feels like a cleaner and more realistic target at this point could be $AAOI…
Show more
+ $GOOG just delivered the Cloud print that pushes the AI infra trade even harder.
• Google Cloud revenue $30.7B (+103% YoY), beat.
• Cloud operating income $11.1B (beat), ~36% margin.
• CapEx $51.4B (~40% of revenue).
Sundar Pichai: Cloud grew 103% powered by huge demand for AI infrastructure and AI solutions, with strong TPU demand. First-party model APIs now processing nearly 32 billion tokens per minute (4x a year ago).
Together with $MSFT’s Azure +48%, the signal is clear: demand is still outrunning supply and the capacity build is intensifying. Direct fuel for power, photonics/CPO, and the rest of the AI infra complex.
We are long and bullish AI Infra!
Show more
$MSFT didn’t just keep the AI trade alive. It just pushed it even harder.
• Revenue $92.3B (+19% YoY, beat).
• Adj. EPS $4.89 (+18% YoY, beat).
• Azure $33.2B +48% YoY (ex-FX +48% vs ~45% est) — accelerating hard.
• Microsoft Cloud $63.9B (+30% YoY).
• CapEx $39.4B (~43% of revenue) — the buildout is running full speed.
• Operating income +21% > revenue growth → operating leverage intact.
Azure accelerating at this scale means demand is still outrunning supply. Every print like this forces more capacity, more power, more photonics/CPO, and more interconnects.
This is the AI infrastructure spend cycle intensifying.
Commercial RPO came in light vs estimates, but the growth engines (Azure + Cloud) more than offset it.
Next up: $GOOG. If Google Cloud and CapEx confirm the same story, the AI infra trade gets even hotter.
Show more
$MSFT didn’t just keep the AI trade alive. It just pushed it even harder.
• Revenue $92.3B (+19% YoY, beat).
• Adj. EPS $4.89 (+18% YoY, beat).
• Azure $33.2B +48% YoY (ex-FX +48% vs ~45% est) — accelerating hard.
• Microsoft Cloud $63.9B (+30% YoY).
• CapEx $39.4B (~43% of revenue) — the buildout is running full speed.
• Operating income +21% > revenue growth → operating leverage intact.
Azure accelerating at this scale means demand is still outrunning supply. Every print like this forces more capacity, more power, more photonics/CPO, and more interconnects.
This is the AI infrastructure spend cycle intensifying.
Commercial RPO came in light vs estimates, but the growth engines (Azure + Cloud) more than offset it.
Next up: $GOOG. If Google Cloud and CapEx confirm the same story, the AI infra trade gets even hotter.
Show more
$ORCL $CRWV $NBIS $IREN
$ORCL just gave the market a clear exit from being a massive AI debt issuer. VERY important for the rest of the stack.
~$54B FCF turnaround in 3 years. Positive in H2 FY28. ≥$30B in FY29. Capex peaks FY27–FY28 then declines. Self-funded from FY29.
Second-order: Oracle has been one of the largest incremental borrowers in the entire AI infrastructure complex.
When that supply leaves the market, credit conditions ease for the pure-play names still funding capacity.
That’s the relief valve for $NBIS $IREN $CRWV.
The funding overhang that has been weighing on the group just got lighter.
Bullish NeoClouds!
Show more
Oracle $ORCL targets $225B+ in annual revenue by FY2030.
Key financial targets:
• FY27 revenue: $90B+
• FY30 revenue: $225B+
• Revenue growth: 35%+ CAGR through FY30
• FY30 non-GAAP operating margin: 40%+
• FY30 non-GAAP EPS: $21+
• Positive free cash flow: 2H FY28
• FY29 free cash flow: $30B+
• CapEx to peak in FY27-FY28
• Growth to become self-funded in FY29
• FY29 net debt/EBITDA: Below 3.5x
Oracle also plans to limit additional capital raises through FY30 to $20B:
• $10B in equity
• $10B in debt
This excludes debt refinancing and is separate from the $20B equity raise already completed in Q1 FY27.
Show more
Bullish — $LITE $COHR $AAOI $SIVE $CRDO
Evercore ISI just called optics a critical bottleneck to the AI buildout. Reasons:
FCC is set to boot Chinese suppliers from the U.S. 3.2T market by Q1’27.
Yields (for Asian players) are running below what the models need. Lane speeds just went from a 5-year cycle to near-annual.
Chinese capacity can’t keep up at the new speeds.
Domestic leaders get the scarcity premium.
The rack is becoming a network.
The bottleneck is still optics.
Position before it’s too late.
We are long many stocks mentioned here. NFA.
Show more
Told you. $VRT $MOD
Heat is a HUGE new constraint. Cooling TAM going $25B → $63B by 2030 (~25% CAGR, faster than the capacity build itself).
Liquid cooling penetration on new capacity 70% → 85%. CDUs + cold plates are the high-growth pieces.
$VRT just confirmed it in the most violent way possible:
• Revenue $4.31B (+61% YoY, 55% organic), crushed ~$3.75B est.
• Adj EPS $2.21 vs $1.82 est (+21% beat, +78% YoY).
• Organic orders >100% YoY.
• Adj Op. Margin 25.2% (+290 bps).
• Raised full-year 2026 guide again.
• Customers already prepaid $4.1B.
Premarket ripping +18% to ~$246. This is the AI power & cooling bottleneck printing in real time.
$VRT and $MOD sit right in the middle of the highest-intensity part of the stack.
We flagged the setup weeks ago.
Still long $MOD (the most asymmetric one). Heat is not optional.
NFA.
Show more
$MOD $VRT
Bernstein: US data center cooling TAM goes from ~$25B (2026E) to ~$63B (2030E) — ~25% CAGR.
That’s faster than the capacity build itself. Liquid cooling penetration on new capacity rises from ~70% to ~85%.
CDUs and cold plates are the high-growth pieces
Heat is the new constraint.
We are long $MOD. NFA.
Show more
You know what’s about to happen to AI Infrastructure stocks today, right?
$SKHY just printed one of the most insane quarters in memory history:
• Revenue $78.9B (+328% YoY).
• Operating profit $62.8B (+632% YoY).
• 80% operating margin (record).
• HBM4 already overtaking HBM3E in Q4.
• 5-year LTAs with major CSPs locking >50% of 2027-28 capacity.
Memory demand is set to outpace supply through 2030.
This isn’t a memory cycle.
This is a STRUCTURAL physical bottleneck of the AI supercycle.
Show more
$SKHY just printed one of the most insane quarters in memory history:
• Revenue $78.9B (+328% YoY).
• Operating profit $62.8B (+632% YoY).
• 80% operating margin (record).
• HBM4 already overtaking HBM3E in Q4.
• 5-year LTAs with major CSPs locking >50% of 2027-28 capacity.
Memory demand is set to outpace supply through 2030.
This isn’t a memory cycle.
This is a STRUCTURAL physical bottleneck of the AI supercycle.
Show more
SK HYNIX $SKHY Q3’26 EARNINGS HIGHLIGHTS
🔹 Revenue: $78.9B (Est. $74.8B) 🟢; +328% YoY
🔹 Oper Profit: $62.8B (Est. $58.5B) 🟢; +632% YoY
🔹 Oper Margin: 80%; +3,300 bps YoY; record high
🔹 QoQ: Revenue +32% / Oper Profit +38%
Q4 Guide:
🔹 DRAM Bit Growth: Mid-20% QoQ
🔹 NAND Bit Growth: Mid-single% QoQ incl. Solidigm
Q3 Pricing & Volume:
🔹 DRAM ASP: High-20% QoQ; bits ~+10% QoQ
🔹 NAND ASP: Mid-20% QoQ; bits slightly lower QoQ
Profitability:
🔹 Gross Margin: 85.7% (Est. 85.2%) 🟢
🔹 EBITDA: $66.1B; 84% margin; +486% YoY
🔹 EPS: $41.07; +205% YoY
Financials:
🔹 Cash: $86.3B; +311% YoY
🔹 Interest-Bearing Debt: $11.9B; -34% YoY
🔹 Net Debt-to-Equity: -32%; -2,800 bps YoY
🔹 Operating Cash Flow: $41.3B; +281% YoY
🔹 Total Assets: $295.5B
2027 Market Outlook:
🔹 DRAM Demand Bit Growth: +mid-20% YoY
🔹 NAND Demand Bit Growth: +high-teens% YoY
🔹 2027 CapEx: To rise significantly vs 2026
🔸 Memory demand expected to outpace supply through 2030
Product & Technology:
🔹 HBM4: Expected to surpass HBM3E in HBM shipment and revenue share in Q4
🔹 2027 HBM: Negotiations finalized; profitability gap with conventional DRAM expected to narrow meaningfully
🔹 M15X: Ramp-up accelerating to full capacity by end-2026
🔹 321L NAND: Over 50% of domestic capacity in Q4
Business Highlights:
🔹 LTAs: Signed with 14 key customers incl. major global CSPs; standard 5-year term; over 50% of 2027-2028 capacity secured
🔹 Q3 Dividend: ~₩20,000/share (~$15.08), incl. a ₩10,000 special dividend from realized Kioxia gains
🔹 Share Acquisition: ₩20T (~$15.1B) for the quarter, incl. ~₩10T for executive & employee compensation
🔸 Quarterly revenue topped ₩100T for the first time
Note:
🔹 Net Profit: $30.0B; +216% YoY
🔸 Includes a $24.6B non-operating loss: a $17.3B investment-asset loss and an $8.8B FX loss as the won strengthened
FX used for USD conversion: $1 = ₩1,326 (mid-market; Oct 27, 2026)
Show more
We called $CRDO a gift at these levels and bought more 4 days ago:
+14.3% already.
Everyone was staring at $SPCX and the $MXL noise. We were watching the purple cables.
ZeroFlap AECs still own the installed base and the relationships, a 93-99% monopoly (even the competitor admitted it). EVERY new GW of AI compute needs thousands of them.
And now the second engine is turning: the 1.6T ZeroFlap Active Light Cable (micro-emitter optics). Same reliability, ~30% less power, 50m reach, 10x lighter. Optics targeting $600M+ next year.
Massive upside left. The market will wake up.
Long and bullish $CRDO. NFA.
Show more
$CRDO is a gift at these levels. We just bought more.
The $MXL noise means nothing. They sell chips, not full cables. First revenue 2027. $CRDO already owns the relationships, the reliability, and the installed base.
Everyone’s staring at $SPCX’s numbers and ignoring the purple cables in the data center photo.
Those are Credo’s ZeroFlap AECs.
They have the monopoly (80-90%+ of the AEC market). Every new GW of AI compute needs thousands of them.
While the market panics over a late entrant, $SPCX just confirmed demand is still exploding.
Start watching the purple cables instead of just the GPUs.
Bullish $CRDO!
Show more
$AAOI is one of the cleanest R/R in the optics stack.
My 2031 base case is unchanged: $75B Market Cap.
From ~$6.7B today that’s ~11x (+1,020%).
Already +14.5% off the recent lows, I was calling this exact setup while it was getting sold hard.
If they execute on vertical lasers + even a mid-single-digit share of 1.6T/CPO (the path their own mid-2027 run-rate implies), this doesn’t re-rate gradually. It explodes.
$SIVE is even more explosive and higher-beta from a much smaller base, but it’s not for everyone.
Positioned in both. NFA.
Show more
Hey @grok, if they execute according to their estimates, what should be the share price for $AAOI and $SIVE by EoY 2027? Give me a short sharp answer.
Hey @grok, if they execute according to their estimates, what should be the share price for $AAOI and $SIVE by EoY 2027? Give me a short sharp answer.
Bullish — $SPCX $NBIS $IREN $CRWV
Morgan Stanley: “We believe this is the most likely outcome” — a residual 38 GW US data-center power shortfall in 2026–28 (33% of 116 GW demand) even after BTM generation, fuel cells and other workarounds.
Chips are still being built faster than the grid can feed them. Time-to-power is the constraint.
That is bullish for the few that can actually offer and execute compute capacity now: $SPCX $NBIS $IREN $CRWV.
Special mention to the power and cooling layer that makes it possible: $BE $MOD $VRT $HPS.A.
Spot the winners and position accordingly.
Show more
$CRDO is ripping +5.4% After Hours for a very specific reason. At $150 it feels like a gift.
The company that actually builds the AI racks and switches just told the market demand is exploding.
$CLS (+16.8% AH) :
• Revenue +83%.
• Networking +102%.
• Guiding 2027 to nearly double ($40B).
• Explicitly named 800G & 1.6T.
• 2028 demand forecasts even higher.
Inside those racks, copper still wins and $CRDO owns ~95% of that market.
The stock just dropped 20% on “1.6T is slowing” fears.
Turns out the demand pipe is accelerating into 2028.
Most people still haven’t connected it.
That’s the move.
We are long and bullish $CRDO. NFA.
Show more
This is not the optics rally.
This is the appetizer.
$LITE, $COHR, $AAOI, $CIEN, $AXT, $CRDO… the whole stack moved today.
Most people still don’t understand how tight the optical bottleneck is getting.
$CRDO is doing something special after hours. Will drop an explanation later.
Thank me in Q1, not yet.
Show more
This is not the optics rally.
This is the appetizer.
$LITE, $COHR, $AAOI, $CIEN, $AXT, $CRDO… the whole stack moved today.
Most people still don’t understand how tight the optical bottleneck is getting.
$CRDO is doing something special after hours. Will drop an explanation later.
Thank me in Q1, not yet.
Show more
Hey @grok, how many bullish posts regarding the Optics/Networking stack ($LITE $COHR $AAOI $SIVE $CRDO and others) have I (@outliercapx) posted in the last month?
Hey @grok, how many bullish posts regarding the Optics/Networking stack ($LITE $COHR $AAOI $SIVE $CRDO and others) have I (@outliercapx) posted in the last month?
