ASIC Development21 Jan 2025 09:54
Foggy, after lots of talk of making his own ASIC, eventually discovered talk is cheap and ASICs expensive. Why it took him 2 years to uncover that fact is just another QBT mystery. What is baffling is why other avenues have not been explored. I assume he’s using something like a Xilinx 7 series FPGA, after all Xilinx are the world’s biggest player. He Indicates the FPGA is very efficient and the Oracle would take up a small amount of real estate on a chip. Why has he not explored the likes of Multi Wafer Projects MWP to verify the Oracle in ASIC form.
The semiconductor world has always been very fast changing and today is no exception. Google Euro Practise IC Service for example, this gives easy access to down to 12nm CMOS processing. The Xilinx 7 series is I believe 28 nm, the Versal is fast, I think 5nm, but costs thousands for a single chip. The TMSC 16 nm Process could be used at very reasonable cost of the order of a few tens of thousands. This would produce the speed needed and verify the Oracle’s logic without breaking the bank.
In the bigger picture the big miners may reject ‘game changing’ technology which upsets the applecart. Perhaps Bitcoin may change the rules. Fast changing technology has a habit of fooling the poor investor.