$HLIT × $AAOI - One Network, One Capex Budget, Two Tickers
Everyone's watching $AAOI for the AI optics story. Almost nobody noticed that buried under it is a second business - CATV amplifiers - that just turned AAOI into a real-time indicator for a completely different stock: $HLIT.
Here's the connection ABSOLUTELY NOBODY has stitched together.
In June 2025, Mediacom selected Harmonic's cOS vCMTS and Ripple remote-PHY nodes for its DOCSIS 4.0 upgrade. Last month, the same operator named AAOI the primary vendor for the amplifier side of the same program - 1M homes by end of 2026.
One network, one capex budget, two tickers. Harmonic supplies the brains and the nervous system; AAOI supplies the muscle between node and home. Every upgraded Mediacom home is an AAOI hardware sale and an HLIT capacity expansion at the same time.
Charter is the identical playbook at 30x scale. Harmonic's cOS is expanding across Charter's entire broadband base with unified-D4.0 Pebble-2 RPDs, while AAOI's 1.8GHz amps are Charter-certified and shipping - record cable revenue, +24% QoQ, full-year CATV guidance raised to ~$320M.
The Cox merger footprint adds fresh TAM for both.
And because amp cascades get replaced plant-wide as each segment converts, AAOI's shipment curve is a live odometer for the plant conversion that Harmonic's software monetizes.
When AAOI management cites "broad-based demand" in CATV, that's independent confirmation (from one layer over in the same stack) that cable D4.0 spending is accelerating RIGHT NOW.
The market already senses the pairing: the day the Mediacom amplifier news hit, AAOI ran 24% and HLIT ran 12% on someone else's press release.
Now layer on what Harmonic JUST filed on Monday.
June 8 was the one day either side could walk away from the $145M sale of Harmonic's Video business. HLIT spent it filing an 8-K nobody required - a voluntary, deadline-day confirmation that the deal remains on track to close this quarter.
Read that the way you'd read an insider buy.
Companies facing a dying deal go silent and let the lawyers draft termination language.
Companies that file unprompted reassurance on the contract's scariest date are telling you which version of the future they're operating in.
What closes alongside the deal: the last reason to model HLIT as a melting video vendor. What remains is a pure-play broadband software company growing 43%, sitting on a $582M backlog up 87%, with raised guidance and $145M of fresh cash - still priced with a conglomerate discount the filing just put on a countdown.
The chart agrees about where, if not when. May's breakout cleared a multi-year downtrend on ~92M shares - the heaviest month in the stock's modern history. That's institutional repositioning, not retail.
The fade since has been orderly and thinning: 0.786 retrace of the $12→$18 swing held at $13.67, the $14.30 gap filled and defended, weekly RSI reset from 76 to the mid-40s with structure intact.
Deep-but-held retrace, record-volume base break, dated catalyst.
Small caps rarely line all three up.
Watch AAOI's CATV numbers if you own HLIT. Watch HLIT's close if you own AAOI. Same wave, two boats.
NFA. Long $HLIT.
$AAOI
Every single point is 100% true.
I don't think people understand how insane the demand is for these transceivers. They are literally tracked down by the box.
Not even sure what the bear case is for AAOI at this point.
1) sounds like the capacity buildout is going well, in terms of hitting their 2027/2028 targets, it's definitely not a demand issue... they have way more demand than they can handle
2) sounds like there's a clear path to 35% gross margins in the next few years
3) their biggest
"North America sales halved purely from US defense budget approvals slipping into H2"
--> Sivers' photonics division doesn't have material U.S. defense photonics revenue to lose. Their photonics business is a LiDAR/CPO play - automotive and datacom, not defense radar or military satcom. The CHIPS Act award applied to their wireless/RF division, not photonics. Photonics revenue was DOWN 32% from last year.
--> The opportunity pipeline is meaningless word salad.
--> Speaking of dilution, we happen to know exactly when and where the next round is! On June 15, at their Annual General Meeting, the board will pass a 15% dilution round.
If you absolutely must buy into this amazing "growth" story, probably wait until that passes.
Good luck!
$AIRJ just published a white paper that reframes the entire thesis. This isn't a water story. It's a permitting story.
Here's the problem they're actually solving.
A 100 MW hyperscale data center generates $3–5M in revenue per day once operational. In the U.S. Southwest, Australia, Singapore, and the Middle East, water permits at that scale now take months to years. Every quarter of delay is hundreds of millions in deferred capacity revenue.
The conventional fix is switching to air-cooled chillers. Problem solved, right? Except it isn't.
Air-cooled systems consume 1.5–2.5× more electricity than water-cooled. At 100 MW IT, that's $5–13M per year in additional electricity costs. Permanent. Paid every hour for the life of the facility. Operators are trading a permitting problem for an energy problem they can't recover from.
AirJoule is the third option. Keep the water-cooled architecture. Generate the water on-site from waste heat the facility already produces. No permit required. No architectural compromise.
The white paper models this across five climate zones using PUE figures sourced directly to Microsoft, Google, AWS, and Meta sustainability reports. The methodology is rigorous.
The Midland, TX case study is where it gets specific.
Two scenarios:
Case A - 75% permit secured, AirJoule closes the gap. 695 Prime units, ~$105M capex. Adds only +0.06 to PUE vs. the +0.15 permanent penalty of going air-cooled - worth $4.7M/yr in electricity savings for the life of the facility.
Case B - full water independence, no permit at all. 287 units, ~$45M capex, +0.03 PUE impact. No GCD approval, no public hearings.
Here's the number that reframes the investment case.
At $3M/day hyperscale revenue, the entire AirJoule fleet capex is recovered in 15–35 days of permit acceleration.
If AirJoule pulls a schedule forward by even 100 days, captured revenue of $300–500M exceeds the entire investment several times over.
This isn't a sustainability argument. It's an IRR argument. Different conversation. Different decision-makers.
Unit pricing disclosed for the first time: $100,000–$200,000 per Prime unit.
At the $150K midpoint, a 287-unit fleet is ~$43M to the customer. A 695-unit fleet is ~$104M.
Real contract sizes. Real revenue events for AIRJ when they convert. The first large commercial contract would be a material catalyst.
One detail the market hasn't priced: the 18–30 month lead time disclosure.
"We engage with operators and EPC partners 18 to 30 months before commercial operation date."
If the unnamed hyperscaler in active technical engagement places an order in 2026, revenue materializes in 2027–2028 - exactly consistent with management's guidance. The pipeline is further along than it looks.
The white paper is written for technical buyers and EPC engineers, not retail investors. PUE figures anchored to hyperscaler sustainability disclosures. Climate methodology referenced to ASHRAE standards. Construction benchmarks from JLL and Turner & Townsend.
That's a signal about where the commercial conversation actually is.
AIRJ isn't selling water. It's selling permit independence and schedule certainty to operators for whom a single quarter of acceleration is worth $270–450M in captured revenue.
$45–105M contract sizes. Active hyperscaler engagement. Prime built and operational. Runway through 2027. The white paper is the commercial document. The contracts are the catalyst.
NFA. DYOR. Long $AIRJ.
First, IQE's value isn't just MOCVD deposition - it's the proprietary process recipes, NanoImprint Lithography for wafer-scale grating patterning, and decades of yield optimization on compound semiconductors that are notoriously difficult to grow consistently. Buying an Aixtron tool doesn't transfer that IP. You get the oven, not the recipe.
Second, the defence qualification moat is completely separate from the internalization question. Raytheon and BAE aren't going to build internal epi capability for classified radar and missile seeker programs. The qualification process alone takes years. IQE's sole-source positioning in Western defence isn't threatened by MOCVD availability.
Third, the MACOM strategic investment with a long-term InP supply agreement cuts directly against your internalization argument. MACOM (a sophisticated buyer who absolutely could evaluate building internal epi capability), chose instead to take an equity stake and lock in a long-term supply agreement with IQE. That's a revealed preference from the most informed possible counterparty.
@ChairmansLedger@AtlasShrug1 100% agree with Galt here. This is the final innings - the labor market is completely cooked as well, and if you believe the 4.3% unemployment numbers I have a bridge to sell you....
Good framework, a few additions worth considering:
On Joré - this needs more context. People see "2001 bankruptcy" and assume incompetence or worse. The actual story is the opposite. Matt Joré grew up in Ronan, Montana - population 1,800. He and his brother Mike, along with their father Merle, built a power tool accessories manufacturing company from scratch in rural Montana in 1993. No venture capital. No tech hub. No network. Just a factory near Ronan and a product good enough to land on the shelves of Home Depot, Lowe's, Sears, Black & Decker, Stanley, and Makita simultaneously. By 1997 revenues hit $23M. By 1998, $44.9M. They were doubling and tripling revenue year after year, becoming the largest private employer in Lake County. In 1999 the business was so compelling that sophisticated financial institutions lined up to hand them $40M in a public offering overnight.
What killed it wasn't the business - it was what happened after the IPO. Post-listing over-expansion, outside management layers, and critically, off-balance-sheet debt disguised as equipment leases that added nearly $40M in hidden liabilities. When the 2000-2001 retail hardware downturn hit, the true debt load became visible and the company couldn't survive it. Joré was removed from the CEO seat before the filing - a restructuring specialist was brought in. The personal bankruptcy was a consequence of the company collapse, not a character event.
A founder who watched a genuinely great business get destroyed by hidden off-balance-sheet debt and post-IPO capital mismanagement has learned the most expensive lesson in business. That specific scar is directly relevant to what AIRJ needs to navigate now.
Missing: The Rice family - your framework credits Stuart Porter as the aligned billionaire holder but omits the Rice family entirely. Their long-duration capital and network compound directly with Baucus's government access - patient family capital that stays through dilutive raises is a categorically different signal than institutional money with a quarterly mandate. I would say this nudges management to 13-14/15.
A founder who built something real from nothing in rural Montana, took it public, lost control of the capital structure in a systemic crash, paid the personal price, and came back with GE Vernova, BASF, CATL, and Carrier around the table deserves more than a cautionary asterisk, I would think.
$HLIT
Quietly executing with a catalyst next week.
The Video sale closes June 8
$145M all-cash from MediaKind transforms the balance sheet overnight:
→ Wipes $111M in debt
→ Leaves $140M+ net cash
→ Creates a pure-play broadband company
The customer win cadence is relentless.
Since Q4 2025 alone:
🇺🇸 Vyve Broadband - 16 states, cOS deployment
🇲🇽 izzi Mexico - large fiber rollout, multi-year
🇻🇪 Inter Venezuela - largest private ISP, nationwide cOS
🇩🇪 Vodafone Germany - DOCSIS 4.0 trial → scaled commercial
🇫🇮 DNA Finland - SeaStar optical nodes, MDU
🇨🇭 Canal Alpha Switzerland - XOS AI Media Processor
🇹🇼 KBRO Taiwan - fiber on-demand
🇸🇪 Alcom Nordic - white-label
Vodafone Germany, the largest cable operator in Europe, 25M homes passed - has run on Harmonic's CableOS since 2020.
DOCSIS 3.1 → now upgrading to DOCSIS 4.0 on the same platform.
That's the upgrade cycle compounding on the installed base. Operators don't rip and replace when they score the vendor an NPS of 85.
Speaking of NPS
Customer Net Promoter Score hit 85 in Q1 2026, up from 82 at year-end 2025.
For context: Apple typically scores ~72. Amazon ~73. An NPS of 85 in enterprise infrastructure is virtually unheard of.
Operators deploying billions in capex don't leave vendors they love this much.
The AI intelligence layer is live and working
Beacon, Pathfinder, and Amply aren't vaporware:
→ Beacon reduced customer service calls by 30%+ at early adopters
→ Every cOS deployment feeds proprietary network telemetry data back into the AI layer
→ The moat deepens with every new operator that goes live
This is the software multiple expansion catalyst the market hasn't priced yet.
$HLIT is quietly building a global broadband OS:
→ North America: Charter, Cox, Vyve + Tier-2/3 operators
→ Latin America: izzi Mexico, Inter Venezuela
→ Northern Europe: Vodafone Germany, DNA Finland
→ Central Europe: Canal Alpha Switzerland
→ Asia: KBRO Taiwan
Rest-of-world revenue hit $138M in 2025 vs $95M in 2024. A 45% jump in one year.
The switching cost moat is underappreciated.
Once an operator deploys cOS as their vCMTS it is embedded in the network architecture - not a hardware box you swap out.
It's the operating system of the network.
Vodafone went DOCSIS 3.1 → 4.0 on Harmonic. DNA Finland chose Harmonic for the hardest MDU problem in their network.
These operators are not shopping around.
The numbers:
→ Full year 2026 broadband revenue guidance: $475M–$495M
→ That's the third consecutive guidance raise
→ $582M backlog, up ~87% YoY
→ Video sale closes June 8: $145M cash, debt wiped
→ Pure-play broadband re-rating ahead
The Moat is IQE is the primary pure-play foundry capable of handling advanced InP epitaxial growth at global scale. For any tech giant wanting to challenge Lumentum or Coherent without spending $600+ million on a factory, they go to IQE.
The Moat is IQE’s patented NanoImprint Lithography (NIL) stamps an entire array of multi-wavelength gratings onto a wafer simultaneously in a single step. For customers requiring dense, multi-laser arrays for co-packaged optics, IQE offers a massive, unmatchable cost-and-speed advantage.
The Moat is IQE allows hyperscalers and tier-1 system vendors to bring their proprietary laser designs to a neutral, third-party foundry. This protects the hyperscaler from being locked into Lumentum's or Coherent's standard product catalogs and proprietary margins, securing IQE's position as a critical infrastructure partner.
The Moat is... you get the idea.
@unfadv@mkfilko so Lumentum does outsource to IQE.
who is MACOM using? IQE. for? CW DFB.
who are the defense primes using? IQE.
who is sivers using? IQE.
who just announced multiple Tier 1 design wins
for their InP platform? IQE.
etc...
1K Followers 5K FollowingRetail and E-commerce Health Consultant Nutritional Advisor Practices Meditation Spirituality and Healing Therapies Cancer Survivor
3 Followers 973 FollowingWith great risk comes great reward, always hunting for the next 10x - $BRUN, $MRLN, $IQE, $AMPG, https://t.co/radzinzjqN, $SHT, $AIRJ, $INFQ, @AionAnalytics
26 Followers 73 FollowingStocks & Options trader. My posts are not financial advice, just my opinions. Please do your own research! Heavy in $LAES, $CEPT, $IOVA $BCAR $AMBA $VPG $BB $QS
11 Followers 698 FollowingData-driven trader. Trading large trends, key themes, & smaller opportunities in between. The future is NEVER "just like" the past.
17K Followers 51 FollowingSome flow. Some dark pool. Maybe a little charts sprinkled in. And a whole lot of dry humor to keep it from feeling like C-SPAN with candlesticks
9K Followers 265 Following23 y/o Tax guy | Covering the AI infrastructure buildout focused on energy and data centers | My Youtube https://t.co/h7S9gEPpiY
16K Followers 3K FollowingUsing Twitter as unusual flow journal. Also personal acct
Politics, Psychology, Economics, Markets Event Driven Short Squeeze Trader RT are likely bookmarks
42K Followers 658 FollowingSoftware Engineer managing a $5M+ portfolio to build generational wealth for my 2 babies.
Covering AI infrastructure, Semis and Datacenters
9K Followers 362 FollowingDad. Management Consultant. Private Investor.
"Yesterday I was clever, so I wanted to change the world; today I am wise, so I am changing myself . . . " -Rumi
2.0M Followers 624 FollowingOfficial X account for The Kobeissi Letter, an industry leading commentary on the global capital markets. Email us: [email protected]
38K Followers 213 FollowingTechnical Stock Analyst | Long Term Investor & Swing Trader | Navigating markets with strategy and discipline.
Not Financial Advice.
34K Followers 79 Following| enjoy analyzing data | GAMMA , delta, vanna charm analysis | Not Financial Advice | Thank you for your support~ NEVER CHART AGAIN :). 📈 👑
120K Followers 330 FollowingContent is for info/entertainment purposes only and not for financial or investment advice. Investing involves risks.
*NFA = Not Financial Advice.
Hebrews 13:5
2K Followers 3K FollowingPassionate about stocks and options. 10+ yrs of experience in the market. My tweets are just my opinions, not an advice. Best picks: $ROKU $TSLA $AAPL $AMD
38K Followers 654 FollowingDeciphering options flow & dark pool data since 19'
Market education and technical analysis
Fibonacci 🐚
I'm live on voice every morning - link below
41K Followers 4K FollowingEnergy, tech, AI guru. Traveling the world. Houston/Tulum based. Messing around with ideas.
https://t.co/ESDPgiT3Lg
https://t.co/s4aVYs6ShK
200K Followers 668 FollowingMarket Wizards | Contrarian stoic trader and long term investor | since 2013 | Tweets are purely for entertainment purposes and not investment advice
64K Followers 3K Followingfinding asymmetric opportunities in any and every mkt+NOT financial advice -- positions will change - do your own dd - do not rely on mine
13K Followers 167 FollowingEx Goldman Sachs Quant - Worked 30 yrs in Wall Street
📈 Data based simple option strategies and analytics
$SPY $QQQ $IWM $ES_F
306 Followers 514 FollowingI am a private investor
investment consultant and financial analyst.
They are not investment advice.
Audentes fortuna iuvat.
14K Followers 475 FollowingProfessional trader. Wall Street veteran for 20 yrs in NYC working w/ the largest HF's in the world. Founder CSC Macro Report & Trading Services 👇👇👇