As per Skill and Chance's suggestion in the congressional redistricting thread, I've taken a look at what a Democratic map of the Texas State House might look like, using the 2018 population estimates. It shouldn't be taken too seriously, particularly since some of the county groupings probably won't hold up after another two years of population growth, but it does give some sense of the range of manouevre.
Map here:
https://davesredistricting.org/join/8d69f2df-f111-437e-8e55-35b2550a9a70I managed to 84 districts that voted for Clinton, as opposed to 66 that went to Trump. What's more, all but a handful of those districts gave Clinton more than 53% of the vote, which given the relatively high third-party shares in 2016 generally equates to a margin of victory above 10 points. Assuming nothing particularly odd happens with vote swings in Texas this year, that's probably a decent marker of what a safe seat might look like if Texas does shift into proper swing state status.
I went for a fairly soft gerrymander - no thin tendrils, but a willingness to crack strongly Republican areas between multiple districts. I did pay some heed to trying to increase minority representation, but working out what Hispanic percentage makes a district perform in which bits of Texas didn't seem worth it, given the hypothetical nature of this map. If I haven't drawn sufficient performing districts, a few more districts might need to be conceded, although in other cases the VRA could still be satisfied with slightly uglier lines.
Distribution of Clinton districts:
Along the border: 16 (out of 16; Clinton's lowest score here was 54.9% and that could easily be bumped up with uglier lines; all are likely to be won by Hispanic candidates as they're all above 70% Hispanic by total population and mostly above 80%)
Nueces County: 1 (56.7% Clinton, 76.7% Hispanic by total 2018 population)
Bexar County: 9 (out of 10; one is only 51.9% Clinton but the other 8 are all north of 54%; all at least plurality; 7 are Hispanic majority by total population and the other 2 are strong pluralities, but may not quite be a plurality in the Democratic primary)
Hays County: 1 (but only 49% Clinton)
Travis/Bastrop: 7 (out of 7; weakest is 55.1% Clinton; two are Hispanic majority by total population but I'm not certain any are by CVAP)
Williamson County: 1 (but only 48.2% Clinton)
Bell County: 1 (56.6% Clinton; a fairly compact Killeen district; would probably be represented by a black Democrat)
Denton County: 2 (out of four and a bit; 47.2% and 48.7% Clinton respectively but probably trending leftwards reasonably securely)
Tarrant County: 6 (out of 11; weakest is 53.7% Clinton but only one is above 60%; all six are majority-minority but wouldn't like to speculate about which would perform for which group)
Dallas County: 14 (out of 14; weakest is 51.1% Clinton but the others are all above 53%; two black-majority seats and one black plurality; three Hispanic-majority seats and three Hispanic plurality - though some of the latter group might be more likely to return black than Hispanic Democrats; one Asian opportunity seats in the NW)
Collin County: 1 (out of 5; only 48.5% Clinton but there's at least one more Democrats would strongly contests from 2022 and four might be competitive by 2030)
Harris County: 21 (out of 25; weakest is 51.8% Clinton and a few others are sub-53%, but all are growing rapidly; 9 Hispanic plurality and 7 Hispanic-majority districts but far fewer than that would perform; one black-majority and one black-plurality seat but probably at least four which would reliably elect black candidates)
Fort Bend County: 3 (out of 4; only is only 50.8% Clinton and another 52.1%; one Hispanic plurality, one Asian plurality, one black plurality)
Jefferson County: 1 (Beaumont-Port Arthur district, black plurality)
Aside from the Clinton districts, there's not much else left on the table. There's another competitive seat in Williamson; one more in Collin (and two more which might be by 2030 if they trend rapidly), a seat in Brazoria and one more in Galveston, and that's about it.
I've tried to keep cities like Waco and College Station whole so they'd probably be the next targets hoving in to view, but honestly if those flip then the map doesn't matter much because Texas will be a securely Democratic state anyway.