The new 512GB MacBook Neo has demonstrated impressive performance in heavy database workloads, outperforming certain cloud server instances in key benchmarks. Tested by DuckDB developer Gábor Szárnyas, the entry-level Apple laptop ran large data queries faster than cloud setups with significantly more cores and memory in initial tests, thanks to its fast local storage.
The evaluation compared the MacBook Neo against two cloud instances: an AMD EPYC-powered c6a.4xlarge with 16 vCPU cores and 32GB RAM, and a powerful AWS c8g.metal-48xl with 192 Graviton4 cores and 384GB RAM. Szárnyas ran DuckDB benchmarks ClickBench and TPC-DS, focusing on realistic database operations such as filtering, aggregation, and window functions over datasets measuring up to tens of gigabytes.
In the cold run-when caches start empty-the MacBook Neo completed 43 ClickBench queries in under a minute, up to 2.8 times faster than either cloud machine. This advantage mainly stemmed from its fast local NVMe SSD, contrasting with the network-attached storage used by cloud servers, which incurs higher latency. However, in hot runs when caches were warmed, the Graviton4 cloud instance finished in 4.35 seconds, far ahead of the MacBook’s 54.27 seconds, though the MacBook still performed close to the AMD cloud server’s 47.86 seconds despite its hardware disadvantages.
More notably, the MacBook Neo’s median query runtimes sometimes outpaced the mid-tier c6a.4xlarge cloud instance, showing that for certain high-demand database tasks, a laptop’s local hardware can compete with larger cloud deployments. The Apple chip’s efficiency also kept the MacBook’s total runtime within 13% of the faster cloud server, despite having 10 fewer CPU threads and only one-fourth the RAM.
When running the more complex TPC-DS benchmark, the MacBook Neo performed well up to a scale factor of 100, completing most queries in about 1.6 seconds each. At scale factor 300, memory limits led to heavy disk spilling and an extended total runtime of 79 minutes. Still, processing such large workloads on a laptop highlights the progress of mobile computing power.
These tests confirm that Apple’s integrated hardware and fast local storage can deliver significant performance for data-heavy tasks, rivaling cloud options that are far more costly to operate. As data workloads continue to grow, the balance between local processing power and cloud flexibility becomes increasingly nuanced, with devices like the MacBook Neo challenging traditional views on where intensive data work should be performed.