AI
Oct 8, 2026
From enterprise information retrieval to Physical AI infrastructure, the UAE-founded technology group is developing systems designed to bring AI into business operations and the physical world.
[For more news, click here]
AI models are only one part of the enterprise AI equation. Businesses also need systems that can retrieve reliable information from internal records, operate within their infrastructure requirements and support decisions in real-world environments. AHOY is developing technology across these areas, from enterprise AI retrieval through AHOY Labs, aka Trouvé, to its Machine Studio platform for Physical AI. Its expansion into Canada also points to a broader effort to connect research, infrastructure and real-world deployment.
Testing AI against enterprise data
Enterprise knowledge is spread across emails, meeting transcripts, customer records and internal documents. Retrieval-augmented generation (RAG) connects AI models to this information, helping them draw on relevant company data rather than relying solely on what they learned during training. The quality of the answer, however, depends on whether a system can find the right documents and use them accurately.
On October 6, AHOY Labs reported a score of 74.95 on EnterpriseRAG-Bench, saying it ranked eighth among 26 systems in a leaderboard snapshot. The benchmark uses more than 500,000 simulated company records and 500 questions, including tasks that require information to be retrieved across multiple documents. According to the company, its system answered 79.6% of questions correctly, achieved 82.32% document recall and returned an average of 1.03 irrelevant documents per answer. AHOY Labs also reported a score 18.67 points above the benchmark average and 24.71 points above the median.
From enterprise AI to physical operations
Retrieving information from business records is one challenge. Applying AI in environments such as logistics, aviation, traffic management and utilities introduces additionaldemands, including how systems interact with their surroundings and how their deployment is governed. AHOY Machine Studio is designed to address these demands through a platform for building, governing and deploying what the company calls Real-World AI.
The platform is organized around three components: AMS Core, which handles identity, policies, signed releases and auditing; AMS Forge, for building, testing and approving models, agents and applications; and AMS Anvil, for deployment and operations. AHOY says the platform is designed to operate across cloud, on-premises and edge environments, including offline and air-gapped settings. Its Perception Suite combines vision, audio and data intelligence, extending the platform's focus beyond digital information to inputs from physical environments.
Sovereign AI and the infrastructure behind it
AI sovereignty is about more than where data is stored. It also involves where AI systems run, who can access information, how deployments are governed and how much an organization depends on external infrastructure. These questions are increasingly relevant to governments and businesses deploying AI in sensitive or operationally critical environments.
At World Summit AI in Amsterdam on October 7–8, AHOY showcased Machine Studio at Booth B4, highlighting its approach to edge deployment and on-premises infrastructure. The company also promoted founder and CEO Jamil Shinawi for a session titled “From Chips to Models: Owning the Full Stack,” focused on sovereignty and data ownership. Running AI closer to the data, including on local infrastructure or at the edge, can give organizations greater control over their systems and help support operations where connectivity or access to centralized infrastructure is limited.
Canada expansion combines research and growth
Back in May, AHOY announced Montréal as its North American headquarters and primary research and development hub, alongside a partnership with Mila, the Québec AI Institute. The company also announced plans to fund a Physical AI Lab in Canada, expand its workforce and introduce its Perception Suite. The expansion brings research collaboration, product development and recruitment into a single North American strategy.
AHOY also announced a nine-figure agreement to acquire Wrk, a canadian automation company. The company named Hood Khizer as chief scientific officer and Benoit Julien as chief technologist as part of its wider expansion. Together, these developments connect AHOY's work on enterprise automation and AI infrastructure with its plans to build out research and Physical AI capabilities in Canada.
The expansion comes amid significant reported growth. AHOY stated its annual recurring revenue had exceeded $100 million, representing 330-fold growth over five years, and reported positive operating margins. The company also disclosed advanced discussions with tier-one investment banks about growth-stage funding. These company-reported figures provide financial context for its plans to expand its research, product development and Physical AI capabilities.
The next phase of AI will depend on more than increasingly capable models. Access to reliable information, control over infrastructure and the ability to operate in real-world environments will shape how effectively businesses put AI to work. AHOY's work across enterprise retrieval and Physical AI reflects this shift, with a focus on the systems needed to move AI from development into day-to-day operations.
Related articles
AI Is Turning Data Sovereignty Into an Enterprise Infrastructure Problem
How Confluent Is Fixing the Data Layer That Is Breaking Most AI Projects
Related Articles
