An exaflop is a unit of computing speed equal to one quintillion (1,000,000,000,000,000,000) floating point operations per second. It is the highest order of measurement used for supercomputers. Exaflop speeds are currently only achievable with large clusters of powerful processors and considerable energy consumption. Currently they are well beyond the capabilities of most home computers or even server systems. Exaflop speeds are typically employed in high-level research tasks that require tremendous processing capabilities such as artificial intelligence or deep learning applications.  

 The future holds a great potential for exaflops as research progresses and technology advances more efficiently than before. With continued developments in hardware design and new algorithms, exaflops may become more achievable and accessible for everyday use. It is estimated that the first exaflop capable computer will be released sometime in the 2020s with many more to come after that. This could revolutionize the way we think about computing power and bring us closer to solving some of the world’s most complex problems.  

Regardless of its future potential, exaflops remain a powerful measurement of computing speed today, one that is likely to continue dominating supercomputing research well into the future. With its ability to process unimaginable amounts of data at lightning speeds, it has opened up new possibilities for high-level research projects across all disciplines. 

 Cerebras Systems offers a 13.5 million core AI supercomputer, Andromeda, now available and being used for commercial and academic work. Built with a cluster of 16 Cerebras CS-2 systems and leveraging Cerebras MemoryX and SwarmX technologies, Andromeda delivers more than 1 Exaflop of AI compute and 120 Petaflops of dense compute at 16-bit half precision. It is the only AI supercomputer to ever demonstrate near-perfect linear scaling on large language model workloads relying on simple data parallelism alone.