Johnny Alerts evidence

__$QCOM · Qualcomm__ Price: $205.44 (+4.1%) Entry: Looking for entries at $185 - $200. First buy: $200 (stock) Second buy: $190 (Leaps half position) Last buy: $185 (200 ema) Full leap position Stop loss: we don't use those here. We're playing a down trend hoping for 1 more tap on support for us to enter then get a breakout to $260+ Position: Stocks for safety, leaps at very greedy entries like $185. Timeline: 6–18 months — multi-quarter re-rate The market still prices Qualcomm like a phone-chip company you trade around the handset cycle. It's not. Underneath is a toll booth — the QTL licensing arm pulls ~$1.4B a quarter at ~72% pre-tax margins on basically every 5G phone sold plus a record auto business ($1.3B last quarter, +38%). The new angle is the data center: on Investor Day they doubled the FY29 targets, guided data center from ~$0 to >$15B by FY29, and put up >$18 EPS. This is an inference play, not a training play — they're not trying to dethrone NVIDIA, they're going after share of the inference/TCO market the big fish leaves on the table. Their edge: power efficiency and LPDDR memory instead of supply-constrained HBM (Dragonfly C1000 CPU + AI200/AI250 accelerators, claimed 8x tokens-per-watt). And they bought Modular (~$3.9B, Chris Lattner's Mojo/MAX team) — the tell that they know the real moat isn't FLOPS, it's NVIDIA's CUDA software stack, and they're buying their way at it. Best part: it's cheap. ~17–18x forward earnings, normal for QCOM and dirt cheap for anything "AI," and it pays you a dividend to wait. Research:(https://www.nefarious.trading/blog/qcom-amd-killer-inference-play) Take Profits 1. $260 — last ath (if you're monthly swinging it and not leaps) No take profits otherwise, manage it yourself. Catalyst: data-center gross-margin disclosure (the one number they wouldn't show), more US hyperscaler wins, Apple 2027 license renewal, auto ramp to $10B Risk Level: Medium — cheap valuation + dividend cushion, but the data-center pivot is unproven and margin is undisclosed Notes: $QCOM is solving one of the biggest problems, ai computing cost. In ai inference (act of running the trained model, not training it) As models get bigger or smarter, their entire model weighs more due to holding all the info, to hold more info it needs memory. Hence, the memory crisis, smarter models = more memory demand. So how does $QCOM solve this? They invented two ai chips designed to break the memory bottleneck. They invented chips that could store more memory, 5x more by regular standards. Allowing massive models to sit in the chip rather than on DRAM. It moves the processor closer to the memory rather than vice versa. This resulted in 10x more efficiency in effective memory banwidth. 🔥 Although they aren't a monopoly which is my preferred investments, they do everything in-house. They bought an ai software company to brdige the gap of allowing QCOMs hardware to fit into all other companies. Basically making companies like $NVDA, $AMD more efficient and taking their market share so companies can now upgrade rather than buy more. My favorite part is they're doing everything in-house. They bought the software, invented new chips, and have years of experience in chip designs. Only issue is they suck at software, so that is their greatest hurdle.