Vantix Research

$KRKNF / $PNG.V UK and France are leading a 40-nation mine clearance operation in the Strait of Hormuz. Royal Navy loading RFA Lyme Bay at Gibraltar. Armed Forces Minister Al Carns on the record: insurance underwriters need "absolute certainty" before the strait reopens, and the autonomous sonar capability being loaded is what provides it. The photo shows a Kraken AquaPix MINSAS sensor on a Hydroid REMUS, prepared by the Royal Navy's Mine and Threat Exploitation Group. Peter Hunter founded Hydroid. He sits on Kraken's board. The REMUS-AquaPix integration has been operational for years. The Anglo-French stack now executing in the Gulf runs Sonardyne navigation, Chelsea Technologies sensors, Forcys harbor security, and AquaPix SAS payloads. Kraken acquired the first three through the Covelya transaction announced early March. Kraken is continuing to execute, and this is just the beginning of subsea robotics!🦑 apnews.com/article/iran-u…

Brett's analysis is sharp, and your Set 3 point is 100% right. It'll be discussed all year, but on mechanics: The one-year deferral works for individual obligated parties, not the system. Bank goes negative by YE27 on Set 2 alone. Deferral shifts who retires what when. Doesn't create new RINs. Set 3 timing actually pressures buyers to buy now. Set 3 is due Oct 31, 26, refiners have to make 26-27 compliance decisions before clarity. Uncertainty pulls buying forward, not back. Brett's three Set 3 paths all leave the moat intact. Option 1 preserves D4 with D5 coprocessing that doesn't qualify for 45Z (best case for MRL). Option 2 modestly eases D4, but the §45Z + CI moat holds independently of RFS. Option 3 drops D4 but finalizes IRR, which stacks on §45Z and doubles the import penalty for foreign-fed competitors. Variance is on D4 magnitude, not direction. "Limited 2028 capacity" is exactly the moat. Political risk is definitely there, but EPA had the moment to break the program on March 26.

@VantixResearch Don’t take my word for it

$CLMT The Trump-era Department of Energy publicly names one company on its website as the producer of "about half of all North American SAF and about 12% of all global SAF through 2030." $1.44B in federal financing at 4.884% fixed. 100M gallons of fuel pre-sold at $1-2/gal premium pricing. The lowest production cost in North America. The market values it like a leveraged specialty refiner. Six of its main competitors are gone or idled. The company that pioneered commercial SAF in 2016 just signed a 70-million-gallon supply agreement as the customer. In 2027, the federal RIN bank goes negative for the first time in 20 years. The forward curve hasn't priced it. A contractual force-monetization clock activates inside the subsidiary on August 5, 2027. Stock at ~$36. The specialty business standalone is worth $22-26/share. The SAF asset (the one DOE just called the largest in North America) is essentially free. Deep dive is live, link below! Also a big shout-out to @Zerosumgame33, @GavMcCracken, and @toiletkingcap for the original coverage and introducing me to the play!

Good catch! That line's worded backward. The 26-27 window comes from the §45Z domestic-feedstock requirement (effective post-2025), not the half-RIN deferral. Asian UCO, Indonesian palm, and Brazilian sugarcane don't qualify for the PTC starting now. That's the actual moat. The half-RIN penalty is a separate RFS provision deferred to 2028. Stacks on top once it kicks in.

@Zerosumgame33 @VantixResearch May I clarify something? @VantixResearch: "The half-RIN import penalty, deferred to 2028, creates a 2026 to 2027 window in which MRL’s domestic camelina supply chain captures full credit value, while ... importers do not" Isn't that logic back-to front? Or am I misunderstanding?

$CLMT This absolute stud @VantixResearch wrote up $CLMT on substack today...... Incredible $CLMT + $LNG analog. He writes: "The closest historical analog to what is happening with CLMT is Cheniere Energy in 2017. Cheniere was the first US company to ship liquefied natural gas internationally, but for years the public market priced it like a utility, capping its multiple at levels for refining and energy infrastructure because the analytical category was wrong. The private market eventually cleared at infrastructure multiples once the take-or-pay contracts started flowing through earnings. Cheniere went from approximately $50 in 2017 to over $150 by 2022, peaking near $200 in 2023, roughly 4x in five years. The structural setup was identical to what CLMT has today: regulator-driven demand inflection, capacity-constrained physical supply, non-recourse project financing on the asset itself, take-or-pay contracts validating offtake, and a sponsor-structured monetization clock. The renewable diesel and SAF version of that mispricing is unfolding now, and Calumet is the analog the market has not yet identified." I've been pounding the table on this concept every single day of my life on X: Equities EXPLODE when multi-year take-or-pay contracts show up in earnings power. $CLMT is the next $LNG

Appreciate it bro! The punt only works if there's a bank to pull from. Set 2 alone takes it to around -930M by YE27. You can't defer compliance into a year where the RINs you need don't physically exist. SRE was the lobbying lever, and EPA already closed it. March 26 rule reallocated 70% of 23-25 exemptions back into 26-27. Future SREs flow back the same way. Trump EPA had every chance to gut Set 2 and instead locked the highest BBD volumes in the program's 20-year history. That was the tell. Set 3 risk is real. But the trade is on Set 2.

@VantixResearch ok i skimmed this. the major concern is what happens during the leadup to set 3 given how dire this likely will be. you did not mention the ability to defer requirements every other year, so most likely most will punt in 27 after doing so in 25 in the hopes of lobbying reform

$EOS.AX | $EOPSF A $500 drone forces a $2M interceptor off the rail. The math is simple: the attacker runs out of money last. Every NATO procurement office has circled the same conclusion. Kinetic interceptors don't scale against mass drone warfare. One Australian company claims to have inverted the equation entirely. Speed of light. <10¢ per shot. Infinite magazine. Zero US components. That last part is the part nobody's pricing correctly. ITAR doesn't tax exports. It vetoes them. Every American laser weapon needs Washington's permission to sell. EOS's Apollo needs none. One regulation. Two markets. One company that can sell to both. I just published a free deep dive, link below.

Free Deep Dive: vantixresearch.substack.com/p/gabler-group…

The free $XK4 Gabler Group deep dive is live. What I found: The founder wrote one of three textbooks MIT uses to teach submarine design. His company supplies the mast systems those submarines can't function without. Three companies on earth do this. The switching cost isn't contractual. It's architectural. SubCtech batteries power everything from US Navy deep-sea systems to a vehicle built to explore Jupiter's moon Europa. They're tripling production capacity this year (like $KRKNF but more offshore-focused). Develogic sells titanium housings directly to the US Navy under a sole-source, no-competition contract. Three consecutive orders. The most underpriced monopoly in European defense, or am I wrong?

Most people think $XK4 is just a boring submarine company. It's not. It's four companies hiding inside one ticker. Gabler Maschinenbau: The 63-year-old original. Masts on 185 submarines across 25 navies. Navies are locked in for 40 years per boat. Develogic: Underwater acoustic modems, seafloor surveillance landers, autonomous sensors sitting silently at harbor entrances. A larger NATO country has already deployed them and is working with them on automated threat classification. SubCtech: Acquired months before the IPO. Lithium-ion battery systems for offshore and UUV/AUVs. And right now they are in the process of tripling their production facility to meet demand. (a baby $KRKNF) North.io: Maritime AI out of Kiel. Raw ocean data turned into defense intelligence. Google Maps of the ocean floor. Free deep dive out tomorrow. 🚨

Gabler Group ($XK4) In 1962, a former submarine chief engineer named Ulrich Gabler started a company in Lübeck, Germany to build what no one else could. The systems that let submarines see, breathe, and communicate. 63 years later, his technology is inside nearly 1,000 masts on 185 submarines across 25+ allied navies. There are only two other companies on Earth that do this. Once designed into a sub, you're locked in for 40+ years. The aftermarket alone runs ~40% of revenue. Last year revenue hit €62M. Margins: 28%. Backlog: €359M. Market Cap: ~€275.9M. And Europe just kicked off the largest rearmament cycle since the Cold War. They quietly IPO'd in Frankfurt. Almost no one covered it. Deep dive coming soon.

Every AI silicon photonics chip shipping today is built on a wafer from one company. Not Nvidia. Not TSMC. A €2B French company called Soitec (SLOIF/SOI.PA). $SOI AI data centers are hitting a physical wall. Copper wires can't move data fast enough between GPUs without burning too much power. The industry is replacing copper with light. Silicon photonics chips convert electrical signals into photons and send them through tiny channels carved into silicon. Those chips need a specific layered wafer to work. Without it, light leaks, the signal dies, and the chip is useless. Soitec makes that wafer using a proprietary process called Smart Cut, protected by 4,300+ patents built up over 30 years. No one has replicated it. They're the sole qualified supplier to Tower Semiconductor, GlobalFoundries, and TSMC for photonics-grade wafers. Tower is 5x-ing its silicon photonics capacity this year. Every wafer is a Soitec order. TSMC's new COUPE silicon photonics packaging enters mass production H2 2026. Nvidia spent $4B on Lumentum and Coherent to lock down optical supply downstream. The upstream wafer monopoly trades at a fraction of those valuations. But photonics is only part of the story. Soitec's wafers are in the 5G radio chips in every modern smartphone (70%+ market share). They're in the mmWave chips in premium iPhones. They're in the processor running Starlink satellites. They're in the RF filters Skyworks just signed a multi-year deal for. 13 customers in production on their filter platform with 12 more testing. So why is the stock down 70% from 2021 highs? Smartphones. The industry over-ordered during COVID and now there's a massive inventory correction dragging revenue down. The market sees a cyclical smartphone parts company at trough earnings. I see the monopoly substrate supplier beneath the AI optical transition, with €1.1B in factory capacity already built, utilization below 60%, and the highest-margin product line growing 34% organically while the rest of the company shrank 29%. The capex cycle is done. The fabs are built. Now it's about filling them. Free deep dive below covering the physics, supply chain, moat, financials, and more: vantixresearch.substack.com/p/soitec-deep-…

@KawzInvests The market's sleeping on this one!

The Anduril supply chain is worth paying attention to. $OPTX - Syntec Optics just announced participation in what they describe as "the largest effort of its kind to equip every U.S. soldier with superhuman perception and decision-making capabilities." That exact phrase appears in Anduril's EagleEye announcement. Anduril took the $22 Billion IVAS contract from Microsoft. Now building AR helmets fusing night vision, AI, and battlefield data for every soldier. $OPTX makes the polymer optics. Lighter than glass. Lower cost. High performance. They claim to be the largest independent manufacturer in the U.S. with full vertical integration. The numbers: $148M market cap ~$28M annualized revenue 5.3x P/S 15.3% EBITDA margin H1 2025 82% insider ownership ~$20M free float NDAA now mandates domestic sourcing for defense optics. China currently dominates polymer optics production globally. This is not a pre-revenue bet. It is a micro-cap component supplier with real earnings and potential exposure to a generational defense program.

$PGY | Pagaya Technologies 42% of US consumers are declined for credit they can actually afford to repay. Pagaya's AI has evaluated $3.2 trillion in applications to identify these creditworthy borrowers that traditional FICO scores miss. The result: a capital-light infrastructure platform embedded in 31 lending partners with 100% retention since inception. KBRA has upgraded 19 tranches of the 2023 vintages. Zero downgrades. Three consecutive GAAP profitable quarters. Core OpEx down to 34% of FRLPC (record low). The classification problem is the opportunity. Every 50bps of FRLPC margin expansion on $12B volume adds $60M EBITDA. That's $840M of enterprise value at 14x, or $6-7 per share. February 9th earnings will spark the reversal.🚀 Free Deep Dive!👇 vantixresearch.substack.com/p/pagaya-techn…

@NFTLunatic @michaelsikand @KrakTrack We can definitely see revenue 10x+ in a base case before 2030! Very compelling play!

@michaelsikand @KrakTrack @VantixResearch How much growth do you think they will have over the next few years?

One of the top authorities on $KRKNF research @KrakTrack @VantixResearch has released a 50 page deep dive into the company. I highly recommend every investor reads to deepen their conviction. If you're still sitting on the sidelines, it could push you over the edge. vantixresearch.substack.com/p/kraken-robot…