$\Lambda$CDM and Holographic Dark Energy Holographic dark energy is usually written $$\rho_{\rm DE} = 3M_p^2\frac{c^2}{L^2}$$ where $c$ is a dimensionless HDE parameter and $L$ is an infrared cutoff. In Li's 2004 model, $$L=L_e \equiv a(t)\int_t^\infty\frac{dt'}{a(t')}$$ is the future event horizon. Li assumed constant $c$, giving $$w_{\rm DE} = -\frac13 \left( 1+\frac{2\sqrt{\Omega_{\rm DE}}}{c} \right).$$ However, with $c=1$ and $\Omega_{{\rm DE},0}=0.6889$, $w_0\simeq-0.887$, rather than $w=-1$. Does not match observations! Also, constant-$c$ observational fits often give $c<1$, leading to phantom evolution, null-energy-condition violation and a turning point in $H(z)$. That conclusion assumes $c'=0$. What about if $c$ is not constant? Then: $$c<1, \qquad c'=c-\sqrt{\Omega_\Lambda}>0, \qquad w=-1.$$ Therefore $\dot H = -4\pi G(\rho_m+\frac43\rho_r) \le0$, so there is no finite-redshift turning point and no phantom phase. Thus ...
$\Lambda$CDM and Holographic Dark Energy Holographic dark energy is usually written $$\rho_{\rm DE} = 3M_p^2\frac{c^2}{L^2}$$ where $c$ is a dimensionless HDE parameter and $L$ is an infrared cutoff. In Li's 2004 model, $$L=L_e \equiv a(t)\int_t^\infty\frac{dt'}{a(t')}$$ is the future event horizon. Li assumed constant $c$, giving $$w_{\rm DE} = -\frac13 \left( 1+\frac{2\sqrt{\Omega_{\rm DE}}}{c} \right).$$ However, with $c=1$ and $\Omega_{{\rm DE},0}=0.6889$, $w_0\simeq-0.887$, rather than $w=-1$. Does not match observations! Also, constant-$c$ observational fits often give $c<1$, leading to phantom evolution, null-energy-condition violation and a turning point in $H(z)$. That conclusion assumes $c'=0$. What about if $c$ is not constant? Then: $$c<1, \qquad c'=c-\sqrt{\Omega_\Lambda}>0, \qquad w=-1.$$ Therefore $\dot H = -4\pi G(\rho_m+\frac43\rho_r) \le0$, so there is no finite-redshift turning point and no phantom phase. Thus ...