Data. Accelerated.
Reconfigurable computing for analytical workloads—from a personal accelerator beside your laptop to arrays built for real-time, data-center-scale performance.
One platform, every scale
dau is building a hardware and software platform that transforms analytical workloads into reconfigurable dataflow. Flash the architecture your work needs, run it through familiar data tools, and scale the same model from your desk to the data center.
Personal computing
A consumer device designed for plug-and-play. Variable configurations for variable workflows from the comfort of your laptop or workstation. A pocket-sized, reconfigureable accelerator.
Portable · Flashable · Developer-friendly
High-performance systems
Four, eight, or more accelerator cards working as one system. An enterprise platform envisioned for massive analytical throughput, low-latency pipelines, and real-time computing at data-center scale.
Parallel · Composable · Scalable
Compose filtering, transformation, aggregation, partitioning, and other analytical operations into streaming hardware pipelines.
Connect reconfigurable compute to dataframe and analytical interfaces instead of rewriting workflows around low-level hardware APIs.
Carry workload configurations across device classes, from accessible personal hardware to arrays designed for extreme throughput.
From proof to platform
dau began with a 5 GB NYSE TAQ dataset and a concrete analytical task: calculate OHLCV bars. On data staged in our initial dpv1 platform, repeated queries ran in 0.92 seconds versus 1.5 seconds on CPU—63% faster, with every result matched against the software golden. Now we are scaling that architecture into a larger, faster platform.
dpv1 · Proven
Initial platform
134K LUTs · 1 GB DDR3
Proved resident analytical queries can outperform CPU execution on accessible hardware.
dpv2 · Scaling now
Next platform
204K LUTs · 8 GB DDR3 · PCIe x8
Expands resident capacity and host bandwidth for larger datasets and more parallel acceleration.
dpv3 · Future
Concept platform
1M+ LUTs · 128 GB DDR5 · PCIe x16 · Networked
Envisions networked accelerator arrays for massive throughput and real-time computing.
Build the next scale with us
We are working with early users and design partners to shape dau hardware, software, and deployment systems. Tell us what you compute, where it runs, and what performance would change for you.