For Defold to be useful for mobile webgames, it would need a new runtime made optimized for mobile web. The current one has too much overhead, which is acceptable for desktop browsers, but not for mobile browsers. As mobile browsers/devices get more powerful in the long term it won’t much matter so I don’t know the effort would be worth it. Better to say if people need optimized mobile webgames, such as to take advantage of possible popularity of embedded chat games, they should use a framework meant for mobile web html5 games such as ImpactJS or Phaser. Otherwise promote native version. It might be a good idea to make a cut down version of your game in Phaser meant to be shareable, and use it to upsell full featured native versions. Kind of like what some “game demo” ads already do but more of a complete game.
I personally wouldn’t bet on embedded chat games being huge… I remember back when facebook games were really booming that some dev friends tested in facebook stream games and they were not hugely successful. This situation may be different because of the 1 on 1 competitive nature of personal chat.
Many of the bigger chat apps upsell their own in house games, use their chat apps for auth and cross app rewards. I imagine if this does take off more chat apps will do it, but also be pretty protectionist and limiting, such as disallowing linking to your own app, and wanting the lion’s share of profitability.