Evolutionary computation (EC) has developed over the past few decades to successfully address problems across a wide range of domains. Its capabilities make EC a prime candidate for the development of anticipatory computation, but its applicability is inhibited by reliance on the formal logic of closed systems. Living organisms have evolved from stimulatory-reaction to intelligent-anticipation as part of an open ecosystem, and such a relationship is prerequisite for the evolutionary development of anticipatory computation. Effective anticipation depends on the accurate correlation of historical multiscalar data, and ultimately it must be realized as a computer, rather than on a computer. This paper presents an early roadmap for the implementation of evolutionary anticipatory machines, through the critical concepts of ecosystemic rationality and hyperscale to the first construction of multiscalar simulators.