J.M.'s comment points you in the direction of why this doesn't work. Iterating $x^7$ 50 times (even if $k=0$) is $(x^7)^{50}$.
(x^7^50)
(* x^1798465042647412146620280340569649349251249 *)
That exceeds the maximum number representable in Mathematica:
In[1]:= $MaxNumber
Out[1]= 5.297557459040040*10^323228467
Even if we considered $x^2 + k$ instead of $x^7 + k$, it will still overflows.
ff[x0_, k_] := NestList[ #1^2 + k &, x0, 50]
In[10]:= ff[1., 0.1]
During evaluation of In[10]:= General::ovfl: Overflow occurred in computation. >>
Out[10]= {1., 1.1, 1.31, 1.8161, 3.39822, 11.6479, 135.773, 18434.5,
3.39832*10^8, 1.15486*10^17, 1.33369*10^34, 1.77873*10^68,
3.16389*10^136, 1.00102*10^273, 1.002046001003433*10^546,
1.004096188126972*10^1092, 1.008209155011116*10^2184,
1.016485700248229*10^4368, 1.033243178809133*10^8736,
1.067591466555603*10^17472, 1.139751539462343*10^34944,
1.299033571706781*10^69888, 1.687488220421276*10^139776,
2.847616494060564*10^279552, 8.108919697245777*10^559104,
6.575457865638055*10^1118209, 4.323664614278137*10^2236419,
1.869407569676091*10^4472839, 3.494684661562269*10^8945678,
1.221282088375859*10^17891357, 1.491529939387699*10^35782714,
2.224661560089872*10^71565428, 4.949119056941505*10^143130856,
2.449377943978157*10^286261713, Overflow[], Overflow[], Overflow[],
Overflow[], Overflow[], Overflow[], Overflow[], Overflow[],
Overflow[], Overflow[], Overflow[], Overflow[], Overflow[],
Overflow[], Overflow[], Overflow[], Overflow[]}
Of course, if you have a starting value $x0<1$, your original function is fine, and converges quickly.
In[12]:= ff7[x0_, k_] := NestList[ #1^7 + k &, x0, 50]
In[13]:= ff7[0.5, 0.1]
Out[13]= {0.5, 0.107813, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, \
0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, \
0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, \
0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1}
Power values below 1 are also fine:
In[16]:= ff2[x0_, k_] := NestList[ #1^0.8 + k &, x0, 50]
In[17]:= ff2[1., 0.8]
Out[17]= {1., 1.8, 2.40036, 2.81475, 3.08851, 3.2649, 3.37689, \
3.44736, 3.49147, 3.51898, 3.53611, 3.54676, 3.55338, 3.55748, \
3.56003, 3.56162, 3.5626, 3.56321, 3.56359, 3.56382, 3.56397, \
3.56406, 3.56411, 3.56415, 3.56417, 3.56418, 3.56419, 3.56419, \
3.5642, 3.5642, 3.5642, 3.5642, 3.5642, 3.5642, 3.5642, 3.5642, \
3.5642, 3.5642, 3.5642, 3.5642, 3.5642, 3.5642, 3.5642, 3.5642, \
3.5642, 3.5642, 3.5642, 3.5642, 3.5642, 3.5642, 3.5642}
But in this case, I question why keeping the last ten iterations makes any sense. The function will have converged by then. It is behaviour of the first ten iterations that is more interesting.