What Cruxml does

CruxML Pty Ltd is an Australian private company headquartered in Eveleigh, New South Wales, founded by Dr Barry Flower (CEO) and Prof. Philip Leong (CTO) — collaborators of nearly 30 years from the University of Sydney — to commercialise real-time machine learning inference on FPGA hardware. The company develops an Application Development Kit (ADK) that abstracts ML deployment across CPU, FPGA, GPU, and ASIC targets, with microsecond latency and reduced size, weight, power, and cost (SWaP-C) versus general-purpose accelerators. Its named proprietary products are CruxVision (real-time computer vision), the CruxML CANE (Cybersecurity Analysis Neural Engine — an FPGA-based line-rate hybrid intrusion detection system), and supporting IP including research into ternary neural networks.

CruxML targets six primary enterprise verticals — finance, cybersecurity, robotics, defence, space, and mobile/embedded — where microsecond latency and tight power budgets preclude off-the-shelf ML platforms. Customer pain points addressed include algorithmic-trading latency, line-rate cyber defence, autonomous navigation SWaP constraints, ruggedised defence computing, radiation-tolerant space inference, and on-device mobile/IoT inference. The technology narrative leans on FPGA-specific advantages over CPUs, GPUs, and DSPs, citing third-party research (Xilinx xDNN) showing up to 20x throughput and 3x power-efficiency gains versus GPUs.

The business model is a blend of professional services (custom ML development, FPGA acceleration, hardware/software integration, and training) and one-time product/IP licensing via the ADK and IP cores. Distribution is direct enterprise sales with a consultative, project-based motion; marketing is primarily content-driven through a technical blog and published whitepapers. Core technology partners include Xilinx/AMD and Groq, with a broader ecosystem (Avnet, Consunet, ACRE X, CEL, Millibeam, RFTEQ, Axiom, solinnov, ebor systems, Hendon). CruxML is founder-controlled, with no disclosed external funding rounds or institutional investors in the available sources, and operates with 1-10 employees.

Name

Cruxml

Legal name

CruxML Pty Ltd

Website

https://cruxml.com.au

Company type

Private

Founded year

2020

Operating status

Operating

Headcount range

1–10 employees

Short description

CruxML is an Australian FPGA-accelerated machine learning company that delivers real-time, microsecond-latency inference solutions to enterprise clients in finance, cybersecurity, robotics, defence, space, and mobile/embedded markets.

Ownership category

akta.pro rank

Product category

Edge AI / FPGA-accelerated Machine Learning Software

NAICS

Custom Computer Programming Services (541511)

SIC

Services-Computer Programming, Data Processing, Etc. (7370)

akta.pro primary industry

End-to-End MLOps & ML Platform Suites (HDAAABAA)

Keywords

• FPGA machine learning
• Real-time ML inference
• Edge AI acceleration
• Hardware ML optimization
• Intrusion detection systems

Headquarters

HQ city

Sydney

HQ country

Australia

HQ region

Oceania

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Supply Chain, Marketing or Sales, Operations

Revenue model

1. Custom ML Solutions and Services: CruxML provides services and solutions to help clients adopt real-time machine learning. The company offers full-stack expertise including data collection, ML model development, off-the-shelf and bespoke FPGA acceleration, hardware/software integration, and training. Revenue appears to come from client-facing custom development and system integration engagements.
2. Integrated Product Sales (ADK and IP Cores): CruxML sells integrated products accessible through an Application Development Kit (ADK) supporting CPU, FPGA, GPU, and ASIC platforms. Products target machine learning in mobile, autonomous, and low-power domains, suggesting product licensing alongside services.

Go-to-market motion1 record

Distribution channels1 record

Marketing channels3 records

Core offering

CruxML provides real-time machine learning solutions accelerated on FPGA hardware, targeting enterprise clients in finance, cybersecurity, robotics, defence, space, and mobile/embedded markets. The company delivers bespoke ML development alongside integrated products (CruxML Platform, CruxVision, CANE intrusion detection engine) accessible through an Application Development Kit supporting CPU, FPGA, GPU, and ASIC platforms. Offerings focus on microsecond-latency inference with order-of-magnitude improvements in size, weight, power, and cost (SWaP-C).

Product overview

CruxML offers a real-time machine learning platform centered on FPGA-accelerated solutions. The core platform (CruxML) is complemented by specialized products including CruxVision for computer vision and the Cybersecurity Analysis Neural Engine (CANE) for intrusion detection. All solutions are accessible through the Application Development Kit (ADK) supporting CPU, FPGA, GPU, and ASIC targets. The company serves finance, cybersecurity, robotics, defence, and space industries with full-stack ML expertise including data collection, model development, FPGA acceleration, and hardware/software integration.

Differentiator

Problem solved

Functional benefit

Brands

• CruxVision: A product brand mentioned as 'See the invisible' - appears to be a machine learning vision product for real-time applications.

Products and services

• CruxML Platform

Quantifiable outcome

• Microsecond latency for ML inference on FPGA platforms
• +1 more outcomes

Segments6 records

Ideal customer profiles6 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability4 records

Feature4 records

Partnerships

Two partnerships are on record, tiered core.

• Xilinx AMDcoreTechnology or IntegrationXilinx (now part of AMD) is a key FPGA technology partner. CruxML develops solutions on Xilinx FPGA platforms and references Xilinx research (e.g., xDNN processor architecture). The partnership involves deep technical integration for optimised ML inference on Xilinx/AMD FPGA hardware.
• GroqcoreTechnology or IntegrationGroq appears as a partner on the CruxML Partners page, indicating a technology or integration relationship for ML acceleration hardware.

Recent moves6 records

Expansion highlights5 records

Emerging players

• Syntiant: Syntiant makes ultra-low-power neural decision processors for always-on edge ML in battery-powered devices. It competes for the same 'ML at the extreme edge' use cases CruxML addresses in mobile, IoT, and embedded segments.
• Octasic: Octasic develops low-power DSP/ML silicon for communications, defence, and industrial embedded systems. Its expertise in real-time, low-power signal processing overlaps with CruxML's FPGA-based real-time ML focus.
• Mythic AI: Mythic builds analog compute-in-memory AI inference silicon for power-constrained edge devices. It is an emerging alternative to FPGA-based inference for the same SWaP-sensitive workloads that motivate CruxML's offering.

Others

• Edge Impulse: Edge Impulse (now part of Qualcomm) provides a developer platform for building and deploying ML on edge devices. It is a complementary software-side tool in the same end-to-end edge ML workflow that CruxML's ADK targets with hardware acceleration.

Direct peers

• Groq: Groq builds high-performance, low-latency ML inference silicon and is named as a core CruxML partner. Both companies target microsecond-class inference for enterprise and real-time workloads, putting them on a convergent competitive-and-collaborative axis.
• Hailo: Hailo designs energy-efficient edge AI accelerators for vision and real-time ML in robotics, automotive, and industrial use cases. It competes with CruxML for the same SWaP-constrained, latency-sensitive inference workloads the ADK targets.
• QuickLogic: QuickLogic combines FPGA IP with embedded AI/ML solutions for endpoint and edge devices. It targets similar ultra-low-latency, low-power ML use cases as CruxML, including defence-adjacent and industrial customers.
• BrainChip: BrainChip develops neuromorphic, event-based AI processors for ultra-low-power edge inference. It competes head-to-head with CruxML in delivering real-time on-device ML for defence, IoT, and embedded applications with strict power budgets.

Broad incumbents

• Lattice Semiconductor: Lattice offers low-power FPGAs and an embedded ML/sensor stack (sensAI) targeting edge AI. As a smaller FPGA vendor positioned for power-constrained inference, it overlaps with CruxML's FPGA-first ML acceleration thesis.
• Xilinx (AMD): Xilinx (now part of AMD) supplies the FPGA platforms CruxML optimises for and is also a broader incumbent providing its own ML acceleration IP (e.g., xDNN). It is both a key technology partner and a competitor at the IP/platform layer.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat4 records

Key risks6 records

Key highlights7 records

Customer concentration

• LinkedInlinkedin.com/company/cruxml

Revenue estimate

Valuation estimate

Number of profiles

Profiles3 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Cruxml