Global cybercrime is expected to reach 10.5 trillion USD by 2025 with an annual increase of 15 percent from 3 trillion USD in 2015. As communication technology evolves and the world gets more connected, the need for securing network communications is ever increasing. Cryptographic Algorithms, such as AES and RSA, are widely used to secure two-way communications between clients and servers on the internet and are also ubiquitous in high security applications such as credit cards, military communications etc. AES is still considered secure even when it is published, as long as the private key is kept secure. However, if an attacker obtains the private key, the security is compromised and may stay compromised for a long time until the user detects this breach and then changes the key. We present a novel approach to cryptography through which the modules of an algorithm such as Confusion, Diffusion, and KeyMixing etc. are evolved to produce cryptographically secure Customized Encryptors. These evolutionary encryption methods maintain security even after the encryption key has been compromised. At least one parameter or a transformation (Confusion Box/ Diffusion Box/ Key Mixing/ Round Logic) may be modified periodically in an encryption algorithm to temporally evolve its behavior by creating a genetically modified variant of the original algorithm. This reduces an attacker's awareness of the structure and operation of encryption algorithm by adding security in the behavior dimension independent of the key. Modifications may be event based, for instance, at the start of a new communication session, periodic, every few milliseconds, or based on a counter of few bytes of data passed. By evolving AES 128, our solution provides a cryptographic strength of 2908 bits while running at a fraction of the processing power of AES 256 that only provides 256 bits security.