HP Labs working on new computer architecture with new OS and memristor-based memory
With ‘The Machine,’ HP May Have Invented a New Kind of Computer — If Hewlett-Packard (HPQ) founders Bill Hewlett and Dave Packard are spinning in their graves, they may be due for a break.
Context & Ripple Effects
This is not a product launch so much as the surfacing of a years-long research bet. HP had already signaled its memristor ambitions back in October 2011, when it said an 18-month timeline for a NAND flash replacement was underway; 'The Machine' is that materials-level work expanded into a full-system claim — new architecture, memristor-based memory, and a purpose-built operating system from HP Labs rather than any existing stack.
The timing matters inside HP's own arc: Meg Whitman's tenure has been defined by cost discipline — the board raised her salary from $1 to $1.5 million in December 2013 explicitly for efficiency efforts — amid the Autonomy shareholder class action and rumored departures of enterprise veterans Bradley and Donatelli earlier in 2014. A ground-up computer project gives the company a growth narrative that restructuring alone cannot supply, which helps explain why eight outlets picked up the Labs story within days.
First-order effects
- HP Labs' mandate shifts from component research to owning an entire computing stack — silicon-adjacent memory plus OS — putting HP in direct intellectual competition with the Intel-and-Microsoft ecosystem its server business currently sells into.
- If memristor-based memory performs as described, HP's own server line becomes the first customer, collapsing the DRAM-versus-storage distinction inside its hardware roadmap.
Second-order effects
- Server rivals and memory suppliers face a framing problem: a credible single-memory architecture reframes their products as legacy tiers, forcing responses in storage-class memory roadmaps rather than incremental density gains.
- A proprietary OS for a new architecture pressures the software ecosystem — ISVs must decide whether to port for a platform whose installed base starts at zero, giving HP an early chicken-and-egg problem to solve before enterprise buyers commit.
Third-order effects
- If the pattern holds, computing moves toward memory-centric design — where one persistent, fast memory tier replaces the cache/DRAM/disk hierarchy — shifting value from processor and storage vendors toward whoever controls the memory technology and its system software.
- Large incumbent IT vendors funding clean-slate architecture research (as opposed to acquiring it) would mark a return to Bell Labs-style corporate R&D as a competitive weapon, with HP Labs as the test case for whether that model still pays.
The trend: Memory is being repositioned from a peripheral component to the organizing principle of system design, with HP's memristor bet as the most aggressive corporate attempt to rebuild the computer around it.