Globally over 2.2 billion people are blind or visually impaired (World Health) Organisation, 2022). Roughly half of these visual impairments cannot be prevented or addressed. Games are no longer just for fun, and as virtual environments often referred to as Metaverses and Meta-spaces continue to emerge, cyberspace is becoming increasingly spatial and important to everyday life. These virtual spaces do not afford tactile interaction. This presents an accessibility problem, particularly for visually impaired people, many of whom rely on haptic feedback to navigate the spatial environment of the real world. The games industry has a rich history of innovation in accessible virtual spaces, with games such as Sightlence (Nordvall, 2013), Papa Sangre (Somethin’ Else, 2010) and Blind Legend (Dowino, 2015) being exemplary of non- visual-first approaches. Games are excellent sandboxes for exploring virtual accessibility for people without a visual-first understanding. In this paper, we present an analysis of a workshop that engaged blind and visually impaired people in the first stage of a co-design process to develop a new audio-only game which will focus on spatial-audio (Frauenberger, Noistering, 2003) through generic low-end game hardware. The workshop was facilitated in collaboration with sight loss charities, with the objective of exploring how to create more inclusive virtual spaces from their conceptions (McElligott, Leeuwen, 2004). Through synthesis of these conversations, we bring to light a desire from blind and visually impaired people to include non-visual accessibility as a core consideration in the design of virtual space software and hardware to support the future of non-visual cyberspace access.