Best part is: 80% of this move was caused by the idea that my grandma would vibecode her own language app.
DAU is accelerating. The $DUOL mass abandonment didn’t happen. Turns out there’s more to a company than just their app interface.
Easiest reversal in history of humanity
People argued AI was coming for Duolingo.
In 2024, Duolingo shipped 1800 course units per quarter.
In the first quarter of 2026 they shipped 20500. An 11 fold increase.
They were also able to launch 150 new courses in 2025 and they extended courses teaching its nine most
Also a thought that just occured to me is how $CBRS could be an enabler of more naturally flowing and real-time language tutoring. Would certainly revolutionize $DUOL ‘s video-call feature.
Just one example of what fast inference enables.
Expecting a lot of money to flow back into companies like $DUOL if this plays out. At least that’s one of the ways I’m hedging this short term hardware uncertainty.
I also believe $DUOL eventually becomes a multi bagger from here.
Wolfspeed's edge scales with voltage. The industry’s power architectures are moving up in voltage because current and copper become the constraint. That does two things for the company: it increases SiC content per system in the 1200 V volume market, and it opens a less crowded,