We show that for all $\varepsilon>0$, for sufficiently large $q\in\mathbb{N}$ power of $2$, for all $δ>0$, it is NP-hard to distinguish whether a given $2$-Prover-$1$-Round projection game with alphabet size $q$ has value at least $1-δ$, or value at most $1/q^{1-\varepsilon}$. This establishes a nearly optimal alphabet-to-soundness tradeoff for $2$-query PCPs with alphabet size $q$, improving upon a result of [Chan, Journal of the ACM 2016]. Our result has the following implications: 1) Near optimal hardness for Quadratic Programming: it is NP-hard to approximate the value of a given Boolean Quadratic Program within factor $(\log n)^{1 - o(1)}$ under quasi-polynomial time reductions. This improves upon a result of [Khot, Safra, ToC 2013] and nearly matches the performance of the best known algorithms due to [Megretski, IWOTA 2000], [Nemirovski, Roos, Terlaky, Mathematical Programming 1999] and [Charikar, Wirth, FOCS 2004] that achieve $O(\log n)$ approximation ratio. 2) Bounded degree $2$-CSPs: under randomized reductions, for sufficiently large $d>0$, it is NP-hard to approximate the value of $2$-CSPs in which each variable appears in at most $d$ constraints within factor $(1-o(1))\frac{d}{2}$, improving upon a result of [Lee, Manurangsi, ITCS 2024]. 3) Improved hardness results for connectivity problems: using results of [Laekhanukit, SODA 2014] and [Manurangsi, Inf. Process. Lett., 2019], we deduce improved hardness results for the Rooted $k$-Connectivity Problem, the Vertex-Connectivity Survivable Network Design Problem and the Vertex-Connectivity $k$-Route Cut Problem.