Multi-patented AI Acceleration
Accelerating Neural Network Models to Theoretical Limits

The Service
Gigantor provides massive acceleration for neural networks.
Our customers hand us a well-trained model, and we return a hyper-optimized circuit ready for FPGA or ASIC microchips. If the input is a camera, we need to know the frame rate (FPS) and the image resolution or frame size. Our customers provide the model form but do not need to include the weights, which is the secret sauce.
The final product has the original model precision and runs at the input speed without slowing down or dropping frames.
GigaMACS™
A multi-patented technology providing massive acceleration for AI.
GigaMACS™ converts models into a new system for AI. The new foundation does not use frame grabbers, memory, or shared resources. Data is received directly from the source, cascades flawlessly through the pipeline, and delivers the output in real-time. With dedicated resources at every channel, bottlenecks are eliminated, and latency is reduced to near-zero.
- The Parallel Pipelined Operation Patent no. 11,099,854
- Looking for a deeper understanding, ask for our Whitepaper.


GigaMACS™ Unbounded
Large model, high resolution, and high-speed combined.
GigaMACS™ provides the platform for AI to operate smoothly; the Unbounded technology enables massive workloads from complex models, high-resolution images, and extremely fast frame rates. With GigaMACS™ Unbounded, AI can evolve into larger, more complex systems without sacrificing performance or accuracy.
- Unbounded Parallel Pipelined Patent no. 11,256,981
- Would you like a deep-dive; ask for a Whitepaper.
GigaMACS™ Unlimited
Complete object detection at all ranges.
AI models powered by GigaMACS™ Unlimited identify all the objects in an image in real-time. Only one model is needed to process and reproduce the frames at various sizes; this is a synthetic scaler. We simply tap into the unused power to create a truly unlimited object detection platform.
With GigaMACS™ Unlimited, all objects at all ranges are identifiable in real-time.
- You can learn more about the Unlimited feature; contact us.


GigaMACS™ 3D
The same amazing real-time high-speed power for 3D CNNs
AI models processing 3D data sets are typically found in medical devices like MRI and CT scanning systems. With the 3D technology, AI models can process the data and identify any areas of concern in real-time while the patient is visiting.
- Applied to 3D data-sets Patent no. 11,354,571
- Learn more about accelerating 3D neural networks.