haYN Capital

SK hynix trading at 11x forward P/E now Just a reminder you can buy it for 5.5x forward P/E through owning SK Square $402340.KS With management actively targeting to reduce the discount to 30%

@insane_analyst SK square a bit better with discount to nav widening to ~50%

@insane_analyst SK the leader in HBM which is closest thing to a non-commodity in memory + Gives optionality with their collab with $SNDK in the case HBF ends up being useful + ADR listing + Korean market mania

You have $100K and must invest it in one of these options. You have access to Korean equities thru IBKR. Bonus points for intelligent comment justifying your vote.

Are you a memory bag holder? Are you looking for a hedge against china building out memory production? Boy do I have the trade for you. $025560.KQ Mirae Corp With Mirae corp you make $ off the chinese memory fab buildout Mirae corp provides back-end equipment in the form of test handlers for memory fabs. Their biggest customers are CXMT & YMTC Even Sk Hynix But the best thing yet is its on sale for 6x annualized ebit (q1 26 ebit annualized) Get it while the deal last theres only ~2.5m shares in free float

“Samsung adviser warns the AI-driven memory super-cycle may lose momentum by 2028 as Chinese chipmakers aggressively expand DRAM and NAND production. “ scmp.com/tech/tech-tren…

Just some Macro Perspective 2026 AI Capex from Hyperscalers is $725B = to ~38% of the US fiscal deficit in 2026 currently $1.9tr Together this equals to ~$2.6tr of spend into the economy = ~8% of GDP We have ~8% of GDP being pushed into the economy in 2026, pro-cyclically I get that Inflation is a problem and bonds are repricing and the FED may hike earlier than the forward curve is pricing right now. But... I think it's gonna take a lot more than 25bps to stop this train.

They also just announced a 8b won purchase order from Unimos electronics on the same day. 100b won market cap btw One purchase order = revenues of 8% of market cap

Mirae corp 025650.KS limit up 30% today No new POs or fundamental news officially filed by them. Just a news article "CEO Lee Changjae has received the Deputy Prime Minister and Minister of Science and ICT Commendation as part of the 2026 Science Day Science and Technology Promotion Merit Awards." Maybe this article put the stock on the radar of South Korean retail traders that can cause some heavy movements with a ~2.3m float asiae.co.kr/en/article/sto…

s/o to @BryzonX for putting it on my radar he has been crushing it!

$Q looks great here. One of those companies in the AI era that wins no matter what. Being a leader in supplying all the chemicals like CMP, photoresists, copper RDL; Etc that continue to be used in higher volumes as vertical stacking of high-end chips continues in memory and logic. Combine this with management putting out conservative guidance and recent capacity expansions in Delaware and Taiwan there's a long runway here for consistent beats on revenue growth and margin growth. TSMC Intel and memory makers expand Fab utilization + capacity + vertical stacking/etchijng steps -> Strong Revenue growth for Qnity

Some bits from $IREN Q3 that may be overlooked $3.4bn AI Cloud contract with NVIDIA 5-year contract for air-cooled Blackwell GPUs Deploying within 60MW of existing data centers at Childress. $3.4b / 5 years = $680m arr $680m arr / 60mw = $11.33m per MW ARR 2027 Expansion to 1,200 MW at $11.33m/MW = potential for ~$13.6B ARR 2027 5 GW = ~$55B ARR "When you look at the GPU financing, which is the lion's share of that CapEx, the Microsoft contract is a great template. We financed 95% of that CapEx at an average interest rate of about 3% through prepayments and GPU financing." 3% is an insanely good wacc "you don't need a sales team in this market, particularly when you've got NVIDIA. They see the whole ecosystem, the introductions, the referrals, putting us in touch with anyone that needs capacity." qualitative benefits of the Nvidia partnership "It certainly plays a role in a number of those conversations and, you know, we are still seeing prepayments being on the table in a large number of instances. That obviously factors in as part of the overall equation. It's not, you know, the single factor that you're looking at. Everything has to go together with a combination of term length, prepayment, credit worthiness, price. Prepayments are certainly very much on the table in the current environment." Prepayments still being on the table gives them potential to finance further Capex at that insanely good 3% level seen in the MSFT deal "Revenues from Microsoft contract and additional 50k GPUs procured during the quarter expected to begin ramping in Q3 CY26 (Q1 FY27)" For anyone wondering why revenues numbers don't look good now they will start showing up in Q1 27 and beyond

Governance Risk Nexturn N Roll Korea: The direct parent of Mirae, owns 40.4%. They make precision CNC machine tools and construction equipment parts. Created in October 2025 by merging Nexturn Bio Sciences (which originally bought Mirae in 2023) with Roll Korea (the heavy machinery affiliate). It's listed itself. Roa & Co Holdings: Owns the other 7.4%. This is a related-party financing entity within the same group — same ultimate controller, separate legal vehicle. Both ultimately roll up to Roa Holdings Company, the family holdco wholly owned by chairman Onn Sung-jun's wife. Now this is where the key risk for the investment here is. Onn Sung-Jun has a lot of influence over the board with 47% ownership and has a shady legal and corporate governance background. I personally am sizing this small just in case of a governance tail risk but I am willing to accept this governance risk for the upside potential in this stock.

Durability of Revenue Growth TLDR: Test handler industry is fragmented and Mirae is not a market leader but they have carved out a niche growth opportunity in Chinese memory fabs. Likely due to export controls blocking other test handlers (Advantest/Teradyne) from selling to china and mirae being cheaper than competitors. I personally believe the scale of China’s memory expansion is a rising tide lifts all boats situation and I expect revenue growth to continue to be strong which the current valuation does not reflect. CXMT alone accounts for roughly KRW 19.9 billion of disclosed direct contracts in calendar 2025, plus an additional KRW 22 billion via Yiling Trading and KRW 10.2 billion via Unimos Microelectronics, both of which are Chinese trading-house intermediaries that the Korean financial press has linked to Chinese memory and OSAT capacity ramps. We also see purchase orders from SK Hynix as well which establishes a relationship and path to potential further orders from SK Hynix.

Just entered a position in Mirae Corp 025560.KS 🧵 TLDR; Strong theme + tight float + cheap valuation + strong korean market They make “test handlers” robotic machines that take the packaged chips (mostly memory), put them into the tester and then sort them after. Memory fab expansions have lead to increased demand for their products. $35m in revenues FY 25 ~87% revenue growth from $18.6m in FY 24 FY25 ebit of $6.3m ~18% ebit margin. Market cap $55m. ~1.5x sales/9x ebit. I have literally looked 50+ AI related companies in Korea since IBKR opened access and I have not seen a company with valuation metrics this cheap. Float is also tight with ~4.5m s/o and estimated free float at ~2.3m shares (53% of shares).

$NVDA and $IREN announced a strategic partnership to deploy up to 5 GW of AI infrastructure across IREN’s global data center pipeline. Sweetwater in Texas is planned as the flagship NVIDIA DSX deployment. NVIDIA also gets the right to invest up to $2.1B in IREN

$PENG Exited for a quick 20% gain in one day. I bought a small trade size because it looked really interesting, but I personally decided to pass after doing more digging. The main thing that caught my eye was their KV cache server fixing the "memory wall" inference runs into alongside memory guidance moving up to 65-75% yoy growth. Their "Photonic Memory Appliance" angle looked interesting as well. In my personal view, the sexy thesis of solving the "memory wall" and "Photonics + Memory" caused this to rocket up fast. In-reality after research: 1. Their memory wall solution captures a smaller slice of data centers than I initially thought and there are a lot of alternatives/competitive solutions to the memory wall than their KV-cache server. 2. YoY Memory growth of 65-75% sounds great up front. "I would say, for the increase that we're seeing in the second half, that's majority pricing," Q2 26 Transcript Higher DRAM prices not fundamental volume growth is the cause of that sharp revenue growth. They are paying higher prices for DRAM too so that 65-75% rev growth is not accretive to bottom line. 3. PMA to benefit Marvell a lot more and don't see much revenue flowing to $PENG. $PENG is just the module maker that takes Celestial photonics fabric silicon + switch and packages into the end product. Real value lies with Marvell. What the PENG thesis actually is: They have proven themselves to be effective at maximizing compute through rack/server level configurations + software. This is highly useful for end customers who don't have the engineering capabilities of Hyperscalers but need on-prem AI factories. This segment can offer a lot of potential revenue growth for $PENG, and there may be further upside, but a 20% gain in a day needs a better growth opportunity imo.

LPKF $LPK.DE Q1 26 report Apr 30 lpkf.com/fileadmin/medi… “ In the Advanced Semicon Packaging area, we are making good progress and are in concrete discussions with several customers regarding orders for initial production equipment. In addition, order intake in the Development and Electronics segments has increased significantly. This underscores the relevance of our systems to our customers.” “Given the current sharp increase in uncertainty in the economic and geopolitical environment, including risks of war and fragile supply chains, and assuming a significantly lower revenue volume in the Solar segment, we expect consolidated revenue of EUR 105 to 120 million and an adjusted EBIT margin between -3.0% and 4.5% for the 2026 financial year. Potential volume orders from the Advanced Semicon Packaging area are not included in this forecast, as the timing of the market-expected ramp-up is determined not by the LIDE process provided by LPKF, but by the qualification of downstream process steps.”

Took profits on half of my $LPKF Still below my valuation based PT and personally believe still has room to run. For me my thesis was a valuation based PT 24-30 EUR within 3-6 months But… 1. It’s only been 1- 2 weeks 2. its within 25-33% from my PT 3. Its 6x ATR from 50D MA. I always cut half of the trade if it moves up fast and within a certain range of my PT Because the marginal upside is smaller, harder to achieve (likely need fundamental news flow that proves the thesis), and on a longer timeframe More than happy to take advantage of path dependency and the trade arriving close to PT in a short time frame

I think an interesting topic to cover would be applying todays episode but inversely to profit taking/knowing when to sell. How do you incorporate rising vol + personal price targets + volume + factor exposure limits + technical indicators to measure stock extension etc... to make selling decisions

@killapabkai Let me know if there are specific things you want me to cover. I’m going in the ARC I think is a 101… feedback helpful

Risk Management 101 | Episode 6: Path Dependency — Don’t Become Liquidity Path dependency, forced selling, and why lower price does not mean lower risk 05:32 PRICE IS NOT RISK, AND THE MATH PROVES IT 13:46 HAS THE SELLING ACTUALLY STOPPED? 17:15 STAND DOWN, PROBE, REBUILD 21:48 LOSERS AVERAGE LOSERS 26:10 SURVIVING THE PATH IS THE EDGE