Stock-picker Serenity kept a low profile today, surfacing with a single post that nonetheless underscored two of his highest-conviction ideas after one of them spilled onto the wider retail stage.

Serenity noted that a member of the Wall Street Bets community had posted about Reddit (RDDT) while sitting on a roughly $1 million position in Applied Optoelectronics (AAOI) — two names central to his own book. He appeared amused by the ensuing comment thread, especially the mentions of his own name, and pushed back on the prevailing skepticism he saw there.

The crux of that skepticism, as he framed it, was the claim that there is "no moat" in lasers and that AAOI makes no money. Serenity countered that both bets may still be early, pointing readers toward what he expects from the company by the second quarter of 2027.

For readers unfamiliar with the names, both sit at the heart of his thinking. Reddit is the social-discussion platform he regards as a misunderstood "money printer" — a business with gross margins near 91% and net income running at roughly a third of revenue, which he argues has been unfairly dumped into the bucket of software names threatened by AI. His thesis rests on an undisruptable network effect and large untapped advertising headroom relative to peers like Meta; it is a core long for him despite being, by his own account, his biggest loser at an average cost around $148.

Applied Optoelectronics is the piece that ties into his dominant theme: AI photonics, or the optical-interconnect buildout. As data centers scale, optical transceivers — devices that convert electrical signals into light to move data between chips and switches — become a chokepoint, meaning a critical upstream supplier the whole buildout depends on. Serenity's framework is to buy what sits upstream of Nvidia's roadmap, and he views AAOI as the only pure "Made-in-America," vertically integrated transceiver play, running laser fabrication, design and assembly in-house in Texas. The relevant products here are the 800G and 1.6T links that hyperscalers are reportedly buying in all the capacity the company can make. His bull case leans on management's guidance toward a steep revenue ramp — on the order of ten times current levels — by the first half of 2027, which is why he keeps pointing to Q2 2027 as the moment the thesis gets tested.

His retort to the "no moat" and "makes no money" critiques is consistent with the risks he has previously acknowledged on the name: execution on building out that capacity, and a history of dilutive at-the-market share sales. The implicit argument today was that the market is pricing the company on what it earns now rather than on the ramp he expects — a familiar pattern for a value-oriented investor who likes to buy small, overlooked chokepoint suppliers before the numbers arrive.

There was little else of substance from Serenity today; the post was more a defense of standing positions than a new call. For now, he seems content to let the skeptics doubt and to wait for the 2027 figures to make his case.

*This is derived commentary for informational purposes only and is not investment advice.*