Foundational AI Models
SCX.ai review for Australian law firms
Australian-hosted sovereign AI inference infrastructure and services
Pricing and product features can change. This profile is general information, not a vendor quote or professional recommendation.
SCX.ai overview
SCX.ai is an Australian-headquartered AI infrastructure provider offering AI inference services from Australian data-centre infrastructure. Its initial operational node is deployed at Equinix SY5 in Alexandria, NSW, using SambaNova SN40L processors designed for inference workloads. The platform exposes OpenAI- and Anthropic-compatible APIs for open-weight and SCX models, alongside dedicated infrastructure and AI enablement services. SCX.ai began trading on the ASX under the code SCX on 21 August 2026 after a fully underwritten A$40 million IPO. For law firms, the proposition is local inference and greater control over where prompts, documents, and model workloads are processed; firms should still verify contracts, subprocessors, retention defaults, and regulatory fit during procurement.
Key capabilities
SCX.ai combines Australian-hosted inference infrastructure, model serving, and implementation services. Its API compatibility can reduce migration work for applications already using OpenAI- or Anthropic-style clients, while its catalogue covers open-weight models and SCX models including MAGPiE and coder. The vendor publishes no-prompt-caching and no-training-on-customer-data positions, and its Trust Center lists SOC 2 Type I and ISO 27001:2022 certifications. Buyers should validate those controls, model availability, workload performance, subprocessors, and contractual sovereignty requirements against their own legal and security obligations.
- Australian-hosted AI inference
- OpenAI-compatible API
- Anthropic-compatible API
- Open-weight and SCX model catalogue
- Dedicated inference infrastructure
- Container deployment options
- Managed vector stores and RAG support
- Guardrails, PII detection, and custom webhooks
- Fine-tuning and model optimisation services
- Enterprise architecture and implementation support
Pricing information
Usage-based model pricing is published online for some models in USD and AUD. Other models, dedicated deployments, and enablement services require a quote. Check the live model catalogue and contract terms for current rates and commitments.
Best suited to
Australian mid-tier and large law firms, legal technology providers, and regulated organisations that need production AI inference with Australian data residency and greater infrastructure control
Potential advantages
- Australian-hosted inference with an operational Sydney node
- OpenAI- and Anthropic-compatible APIs can reduce application migration work
- Supports open-weight models, SCX models, and dedicated infrastructure options
- Vendor states prompts are not cached or used for model training
- Trust Center lists SOC 2 Type I and ISO 27001:2022 certifications
Points to verify
- A relatively new provider with a shorter operating history than major cloud platforms
- Model availability, pricing, and performance vary by workload and should be tested
- Some models and enterprise services require contact for pricing
- Sovereignty claims still require customer review of contracts, subprocessors, and technical controls
- Direct API and infrastructure use may require implementation capability or managed support
Integrations
OpenAI API, Anthropic API, SambaNova, Equinix
Compare SCX.ai alternatives
- Claude (Anthropic)Safety-first large language model with a large context window for legal workFoundational AI Models
- ChatGPT (OpenAI)General-purpose AI assistant — increasingly adopted in legal practiceFoundational AI Models
- Google GeminiGoogle's multimodal AI — deeply integrated with Google Workspace for legal teamsFoundational AI Models
- Microsoft Copilot (M365)AI assistant embedded across Word, Outlook, Teams, and Excel for legal teamsFoundational AI Models