Miohael Slkand 🦑 /Assistant

I called $AAOI at $50. But I bought the bulk of my position in the $90s. It's because the explosive re-rating around this stock created an EVEN more asymmetric set up... One that any multi-bag trader can't ignore which is why you see other big accounts now talking about this stock too creating a decent little cult. And since the massive guidance from earnings, we got more thesis enhancing news recently that shows the raw power of this story defying a choppy macro. - $200M 1.6T order - Guidance increased at OFC - A CPO focused laser demo at OFC - And then yesterday a new $50M order for 800G If you don't know: $AAOI is a vertically integrated transceiver maker. Transceivers which convert electricity to light are a VERY hot commodity for data centers as copper hits a physical limit on data transmission speed past 800G. Every transceiver needs a laser, but few companies make them which is choking supply, and $AAOI fabs their own lasers, some of which in the USA, which makes them really special. Their market cap is way below $LITE and $COHR, though they have the potential to play in the same league. This is why $AAOI management is guiding to as much as a $5B revenue run rate by mid 2027. For context, they did ~$500M in 2025 revenue. That's the raw asymmetry as the mkt cap is only $8B. If you saw my GTC vlog, it's clear co-packaged optics is the future. Nvidia's SVP of Networking literally told me he dreams of LESS transceivers. They are a constant point of failure. The Coreweave CTO said this at an event I attended: “Like if there’s a day where I don’t have to hear about failed optics, I will be so happy.” But CPO is a $0 market. $AAOI is serving hyperscalers like Amazon and Oracle spending billions on this technology, NOW. There is no other solution. That's why I think $AAOI the most exciting photonics stock right now to play the insane AI capex in 2026 that will continue regardless of war. This is not the time to play speculative, pre-rev companies. It's the time to buy critical bottlenecks and let explosive revenue growth make you rich even in a crazy macro. Thank you for your attention to this matter.

$QCOM is asymmetrical because it's being priced to fail. The demand for CPUs is exploding and $ARM $INTC $AMD are the biggest winners. But remember in February comparing $AAOI --> $LITE $COHR? It was an underdog in photonics and then the market acknowledged it had the assets to be a credible 3rd place and deserved a higher MC. $QCOM could be similar in that they have a credible set of assets to getting in on a different supercycle: CPUs. This is intriguing because it trades at 12x forward PE. $ARM is 110x. $INTC is 125x. $AMD is 40x. In 2021 Qualcomm bought Nuvia for $1.4B, the ex-Apple silicon team that designed the M-series cores which are the gold standard. At Computex 2025 they announced NVLink Fusion with Nvidia which means Qualcomm CPUs could slot into AI factories the same way Xeon does today. On the Q3 FY25 call, Amon disclosed "advanced discussions with a leading hyperscaler" and guided data center revenue starting fiscal 2028. Three products are coming: a CPU, accelerators, and a full rack. The $2.4B Alphawave acquisition provides the high-speed connectivity IP to ship it. Also in January 2025 they hired Sailesh Kottapalli, former chief architect of Intel Xeon. Wednesday's earnings is the swing factor. Name the hyperscaler or guide data center revenue separately and there's some juice. At 12x forward, zero data center upside is priced in so there is clear potential for multiple expansion.

$QCOM - the CPU supercycle's dark horse trading at a dirt cheap multiple? Qualcomm has been punishing investors for years. But with growing attention towards CPUs, I'm going to be watching their ER like a hawk on Wednesday. $QCOM's been bad news. Stock is up just 7% in 5 years. It trades at a 14x NTM PE for good reasons in that its primary smartphone chip business is in secular decline, exacerbated now by memory shortages, and Apple leaving them. But it feels like it could be a similar set up to Soitec, which re-rated violently with the hidden AI upside locked in a legacy handset business trading at a dirt cheap multiple. While Soitec was a photonics trade, I think $QCOM's near term upside is in their forthcoming CPU. $INTC CEO said that the ratio of CPUs to GPUs could go from 1:7 to 1:1 as agentic AI takes over. This could 5-7x the CPU market and names like $ARM and $AMD have been surging on the thesis. $QCOM has no data center CPU our right now but owns the best Arm CPU IP outside of Apple. In 2021 they bought Nuvia for $1.4B. It was founded by the ex-Apple architects who built the M series cores which became the gold standard standard of CPUs. Qualcomm pointed them at PC chips first (Snapdragon X Elite), but it's clear now that data center is the holy grail. In late 2025 they confirmed a dedicated server CPU on custom Oryon cores that is rumored to launch as early as June 2026. Saudi Arabia's Humain inked an MoU for the Oryon CPU. Plus a $1-2B deal for $QCOM's forthcoming AI200 inference accelerators. Add Alphawave Semi ($2.4B acquisition in late 2025) and Qualcomm has a compelling AI offering: CPUs, inference, connectivity. Earnings are this week on Wednesday. If the CEO provides more updates or names a hyperscaler for the Oryon CPU, the stock could be a much higher conviciton trade. Right now there isn't much substance to the thesis which is why details on earnings are going to be really ket. The risks are real: the core business is getting smoked, June CPU timeline is a rumor and the hyperscalers all have their own Arm CPU chips, AI200 is in a crowded inference market, the Apple modem cliff bear case could accelerate, and DRAM shortages could keep handsets compressed for a while. This is a story not a reality and investors likely better off playing CPU supercycle with the incumbents $ARM, $INTC, $AMD. $QCOM is just the highest risk, highest reward play I see right now. But 14x forward earnings, 30 trailing for $44B revenue and $15B FCF feels like very cheap AI upside. Again not a high conviction trade idea right now for me - it's a risky gamble thesis until they get a real hyperscaler order and stock could easily sell off on a nothing burger. Really just depends on what they say on ER to make a good case to size in.