Valve announced Steam Deck 2 would arrive when AMD or Nvidia shipped a chip with a 'generational leap' in performance and efficiency — then refused to say what that meant. Now driver leaks suggest AMD's Gainsborough might be the one. The skepticism from Valve's own ecosystem reveals something more important than the specs: the company has weaponized vagueness to avoid shipping a failure.
This is not new. In 2015, Valve partnered with manufacturers to build Steam Machines — dedicated hardware running Steam OS that would sit under your television and play AAA games at console-tier performance. The machines shipped. Nobody bought them. Valve's public line was that partners had failed to meet an unnamed threshold, one the company had never specified.
What Valve calls a 'generational leap' now is the same unmeasured bar that killed Steam Machines then. Not a defined FPS target at 1080p, not a wattage budget, not a price ceiling. Just a feeling that partners had to exceed, one only Valve would recognize after the fact. AMD built hardware that looked powerful on paper. Valve looked at it, found it insufficient for an unnamed reason, and the whole category died quietly.
Gainsborough may be genuinely transformative. The driver leaks suggest real improvements in architectural efficiency. Better power per frame. Matters for a handheld device running eight-hour stretches on battery. But Valve has already positioned itself to reject it. The company waits not because the hardware doesn't exist, but because rejecting hardware keeps the company risk-free. If Steam Deck 2 launches and underperforms the gaming audience's expectations, Valve has shipped a failure. If it never launches, Valve simply waited for the impossible bar to be cleared.
This is how institutions protect themselves when they're afraid to move. You see it in hiring committees that claim to want 'innovation' while rejecting every unconventional resume, in product teams that demand 'disruptive thinking' while punishing anyone who actually disrupts. The unnamed threshold becomes the alibi. Gainsborough might be good enough. It will never be good enough to measure.