Document Processing · IT Operations Automation · Real-time Analytics
HTX needed AI that understood Singapore's public-safety context, could be trained on sensitive data inside a secure environment, and would remain under government control rather than depending on generic off-the-shelf models or externally hosted services.
Reported impact
The partnership accelerated HTX's development of locally controlled AI, enhanced productivity for public-safety officers, delivered top-tier benchmark performance for Phoenix models on HTX infrastructure, and created a reusable platform that is already extending into robotics, video analytics, fire safety, and cybersecurity.
HTX partnered with Mistral to build a sovereign public-safety AI stack instead of buying a hosted AI service. Mistral supplied the training and operational tooling-AI Forge for proprietary-data model training and AI Studio for operationalizing models-while HTX combined that with its NGINE infrastructure, product teams, and in-house engineers. The result is a locally controlled ecosystem that trains and deploys Phoenix family models, including the 123B-parameter Phoenix-VL 1.5 Medium, on HTX-owned infrastructure for Singapore-specific public-safety use cases. Proprietary data and evaluation frameworks stay inside an air-gapped environment with no web search, so domain knowledge has to be embedded into model weights before deployment. Those models are then used for mission workflows such as summarizing incident intelligence, classifying documents, providing chatbot assistance, and supporting cybersecurity response, creating a repeatable pipeline from frontier R&D to operational products.
HTX trains Singapore-context models on proprietary data in AI Forge, operationalizes them through AI Studio, deploys them on NGINE, and then uses those models to ingest incident reports, social media, sensor feeds, and documents for decision support, chatbot assistance, and workflow automation.
"AI Forge for model training on proprietary data, and AI Studio for AI operationalization and foundational models"
"Automated workflows handle document classification and chatbot assistance, cutting administrative load and returning time to frontline staff."
"Intelligent decision support lets officers ingest and summarize real-time intelligence, such as incident reports, social media, and sensor feeds"
"It needed to deploy those models on air gapped infrastructure with no access to web search"
"Phoenix-VL 1.5 Medium, a 123B-parameter natively multimodal and multilingual foundation model for Singapore context"
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Google Cloud Sample Modified · Oct 9, 2026 Anthropic During large-scale training runs, traffic is highly variable—dropping to near zero during model checkpointing before bursting to millions of events per minute. Anthropic needed a serverless messaging queue that scaled… Handles peak throughput of over 1M events per second BigQuery · Bigtable · Cloud Pub/Sub Google Cloud Sample Modified · Jun 26, 2026 Hallam Hallam builds custom no-code agents with Gemini Enterprise, removing development bottlenecks and scaling customer insights. For digital marketing agencies, capturing client attention is harder than ever. Online… Accelerates agent deployment from months to 3-4 hours with Gemini Enterprise BigQuery · Gemini · Gemini Enterprise Google Cloud Sample Archived · Mar 14, 2026 ComplyAdvantage For ComplyAdvantage , a financial crime prevention platform, this transformation presented an opportunity to scale its AI-driven services to get ahead of increasing and evolving threats. 100% improvement in merge requests per developer Gemini · Gemini Code Assist 
