New DRAM that performs computations could revolutionize database performance, but requires new programming approaches.
Create profiling tools and libraries that help developers optimize data structures and algorithms for PIM architectures.
Target database engineers and HPC researchers willing to pay for performance edge tools.
Start with open-source benchmarks demonstrating PIM advantages, then build commercial analysis tools.
The risk is hardware standardization taking longer than expected.