Ferros Capital

SK hynix conference call – View on memory pricing “We do not believe memory pricing is being driven by a temporary supply-demand imbalance, and we expect the trend to look different from past cycles. First, as AI technology advances, the importance of memory has become greater than ever. IT companies are competitively increasing purchases to secure memory supply. On the other hand, supply in the memory industry declined after investment was reduced following the downturn, and meaningful production expansion is constrained by limited available space. To secure room for capacity expansion, suppliers have resumed building new fabs and investing in infrastructure, while continuing to increase the scale of their investments. However, it will still take more time before meaningful new cleanroom space and production capacity come online, so the supply-demand imbalance is persisting. IT companies now view securing volume as more important than price. In addition, the importance of memory in the AI computing market is continuing to grow. We expect the favorable pricing environment to remain in place for the time being.”

SK hynix conference call – View on memory pricing “We do not believe memory pricing is being driven by a temporary supply-demand imbalance, and we expect the trend to look different from past cycles. First, as AI technology advances, the importance of memory has become greater than ever. IT companies are competitively increasing purchases to secure memory supply. On the other hand, supply in the memory industry declined after investment was reduced following the downturn, and meaningful production expansion is constrained by limited available space. To secure room for capacity expansion, suppliers have resumed building new fabs and investing in infrastructure, while continuing to increase the scale of their investments. However, it will still take more time before meaningful new cleanroom space and production capacity come online, so the supply-demand imbalance is persisting. IT companies now view securing volume as more important than price. In addition, the importance of memory in the AI computing market is continuing to grow. We expect the favorable pricing environment to remain in place for the time being.”

Here is SK hynix’s 1Q26 earnings release update: Revenue: KRW 52.5763 trillion (vs. consensus: KRW 51.9346 trillion) Operating profit: KRW 37.6103 trillion (vs. consensus: KRW 36.3955 trillion) Net profit attributable to controlling interests: KRW 40.3302 trillion (vs. consensus: KRW 30.8070 trillion)

$FICO This company has unparalleled pricing power. In fact, increasing prices by 725% for their FICO scores in a couple of years put them in the crosshair of the FHFA (Federal housing finance agency). However, they just found a way to increase their price by another ~700% in a single year! The best part is that they operate such a systemic moat, that even when their competitors offer their product _FOR FREE_, no institution switches over from FICO. This is the kind of company that comes to mind when people say things like "compounding" machine. Ever wondered what that looks like? Well, here, I'll show you. 26% compounded EPS growth despite the past years being the bottom of the cycle (high interest rates, housing crisis). The best part about FICOs score segment is that it is so unbelievable capital light, that their MASSIVE 88% operating margins are almost all converting into FCFF. Now imagine another ~700% price increase; what that would do for revenue- and earnings compounding. To understand how their pricing power is so immense, we first need to understand how it works when a FICO score is pulled. It is quite simple: 1. Consumer inquires about a loan to a lender 2. Lender pulls a FICO score from the 3 credit bureaus (equifax, transunion, experian). 3. The credit bureaus have the consumer data, but need to get the mathematical scoring model from FICO. 4. FICO provides the scoring model and collects a royalty from each of the 3 bureaus. The royalty was fixed at $4.95 in 2025, up from $0.60 in 2018. That's where the 725% increase comes from. However, what is $4.95 (or ~$15 on a tri-merge basis) in the grand scheme of overall mortgage closing costs? Pennies, unnoticable. FICO could increase the royalties much further, and it would still be a drop in the bucket compared to hundreds of thousands, if not millions of dollars borrowed. The FHFA targeted FICO for their pricing power, blaming them for causing the housing crisis due to their price hikes, saying it deters from getting a mortgage. The truth of the matter is that the FICO score cost is immaterial, and only represents a tiny fraction of overall mortgage costs. - In a $500,000 home sale example, a FICO score needs to ben pulled 2000 times to equal a realtor commission at 6%. - If the mortgage broker charges 1.8% in fees, FICO score would need to be pulled 600 times to equal the broker fee. If the FHFA wants to combat expensive housing markets, FICO is not the target as it only represents 0.003% of the closing costs in the $500,000 example. So what did FICO just do to allow them to increase their prices by another ~700%? They combat the FHFA by launching a "direct license model". This offers two transparent options for a lender: - A flat $4.95 fee, with a $33 funded loan fee if the mortgage closes (670% increase in price) - Flat $10 per score, no funded loan fee (100% increase in price) The reason this works is because this is transparent in the invoice. It is clearly outlined what the cost for the FICO score is, as opposed to a bundled cost from the credit bureaus where they apply their own margin, often 5-10x the cost of the FICO score itself. Resellers are used for the direct license model, and the 3 credit bureaus still receive a fee for providing their data, but they aren't able to hike the prices massively with their own margin. Scores revenue grew 27% in 2025 after exiting the bottom of the cycle in the years prior. With the new pricing model, I expect we may see an extreme acceleration in FICO scores and FCFF. For my full equity research report that covers FICOs whole business (including their software segment), with intrinsic value models and more, check the comment section below 👇