Serenity on X
I think almost everything about my $AXTI thesis from 2025 was validated? From AXT's position in the InP substrate bottleneck... Being able to threaten the hyperscaler buildout (per Reuters)... $IQE has been flagging this as a major risk, VPEC and Taiwan epiwafer foundries have been crying over InP substrate bottlenecks. And there's even AXT direct confirmation of partnerships with $LITE and $COHR as the InP substrate supplier. AXT also stated $130m/quarter InP revenue capacity exiting 2027 (~$520m annualized) so that's a massive revenue inflection compared to previously. Now... the biggest question is: How much can AXT price hike? The substrates they make are just such a tiny part of the BOM... but can threaten hyperscaler buildouts with optics. So if they manage to price hike to $SNDK NAND degrees (which they can), then I think AXT could get re-rated a ton from here. But if they end up like Japanese ABF suppliers that are less willing to abuse their position, that's a different story. Will need to do more channel checks to see how that's coming along. But at the very least, I would argue there should be some valuation baseline just due to their critical importance (eg. upstream refinery ownerships, crucible processing, etc) + existing capacity revenue without hikes. TLDR: - Bottleneck thesis was at like ~$500m MC - Upside now comes from price hikes due to duopoly position + ownership of upstream bottlenecks in the InP substrate manufacturing process. (which they can due to low cost of InP substrates in overall costs) - Uncharted territory, we'll see (also a central part of geopolitical tensions so there's always risk there).
