newsfilter
← Latest News
Earnings Call Transcripts·

Accelerate AI-Powered Retail Experiences & Enterprise Workflows with IntelliX Prime™ Pre-Built AI Accelerators on IBM Power S1112 & S1122 | IBM Channel Partner Webinar - ILOG SA

  • Financial Performance:

    • The transcript does not disclose IBM's specific quarterly financial results (revenue, earnings, margins) as the focus is a strategic webinar regarding the Fiducia partnership rather than a standard earnings release.
    • Pricing and Promotions: Fiducia's IntelliX Prime software has a list price of $75,000; a 25% discount is available for activations completed by the end of the year, locking in this price for future deployments in 2027.
    • Hardware Configuration: The S1112 (10 cores, 0.5TB RAM, 10TB storage) and S1122 configurations are highlighted as cost-effective entry points for AI workflows, contrasting with the high capital expenditure (capex) of traditional GPU-heavy setups.
    • Revenue Timing: IBM is targeting the activation of these solutions to contribute to Q4 revenue.
  • Guidance and Future Outlook:

    • Roadmap: IBM Power continues its roadmap with Power 11, utilizing chiplet architecture.
    • Spire Accelerator Expansion: Spire (IBM's AI accelerator) is currently focused on specific ML models but is projected to support a broader range of open-weight models by 2027.
    • Scalability Strategy: The immediate strategy emphasizes CPU-based inference for cost-effectiveness, with plans to integrate GPUs for specific large-scale needs soon.
    • Growth Trajectory: The partnership aims to scale AI deployment from the smallest footprint (S1112) to larger enterprise configurations, leveraging Spire for concurrent user scaling and reduced first-token response times.
  • Business Segments and Product Lines:

    • Partnership: Expansion of the IBM Power ecosystem through integration with Fiducia, an AI workflow orchestration partner.
    • Core Technology: Introduction of "Sovereign AI" workflows hosted on-premise or on IBM Cloud/PowerVS, ensuring data control and compliance.
    • Product Suite:
      • IntelliX Prime: The foundational intelligence orchestration layer.
      • KnowledgeX: An accelerator for ingesting, vectorizing, and indexing enterprise data (RAG systems).
      • SpeedShotX: A low-code/no-code UI-driven accelerator for rapid AI agent and workflow creation.
      • VideoX: Listed as a foundational component alongside accelerators.
    • Hardware Focus: Deployment specifically targeted on IBM Power 11 systems, notably the S1112 (2U half-width) and S1122 (higher core/memory capacity) servers.
  • Market and Competitive Landscape:

    • Differentiation: IBM Power positions itself against x86 and GPU-centric competitors by offering "para-virtualization" with <3% overhead and OMI (Open Memory Interface) memory with error correction and redundancy.
    • Cost Efficiency: Challenges the industry myth that massive GPU clusters are required for AI; demonstrates that open-weight models can run efficiently on CPUs for enterprise workflows.
    • Target Sectors: Focus on Retail (e.g., virtual try-on integrations with Shopify), Defense (sensitive data processing), and Industrial sectors.
    • Open Ecosystem: Supports open-weight models (e.g., Qwen, Granite, Mistral) alongside frontier models (e.g., GPT-5.4) and integrates with Watson X.
  • Risks and Challenges:

    • Data Sovereignty: The primary challenge addressed is the risk of sending sensitive data to external frontier models; the solution offers an air-gapped, sovereign alternative.
    • Complexity: Mitigated by "AI workflows" that integrate with existing legacy systems without requiring total infrastructure replacement.
    • Model Limitations: Certain advanced tasks (e.g., deep learning-intensive try-ons) may still require frontier models or GPUs, though CPU-based inference handles the majority (approx. 80% of services) of current AI workloads.
    • Adoption: Overcoming the perceived necessity for expensive GPU investments to achieve "real" AI results.
  • Management Commentary and Tone:

    • Tone: Highly confident, optimistic, and collaborative. Management emphasizes the rapid deployment capabilities ("magic" activation in days) and the maturity of the technology.
    • Key Narrative: "Intelligence translates to accelerators (GPUs), but we don't need them to start." The narrative focuses on CPU-first efficiency, cost savings, and the ability to operationalize AI in days rather than months.
    • Partnership Dynamics: Strong emphasis on the synergy between IBM's hardware (Power, Spire) and Fiducia's software stack (IntelliX), describing the collaboration as "magical" and "enterprise-grade."
  • Analyst Questions and Answers:

    • Q: Value of Spire for existing customers?
      • A: Spire improves first-token response time and significantly increases the capacity for concurrent user requests without increasing hardware footprint. It is a growth option for scaling workloads.
    • Q: Is the solution compatible with Red Hat?
      • A: Yes, tested and supported on Red Hat and Ubuntu; not strictly required but a preferred option for IBM customers.
    • Q: Can this run without GPUs?
      • A: Yes, the demo confirmed that open-weight models run effectively on CPU-only configurations (S1112), with GPUs (or Spire) being a scalable option for higher volume or specific deep-learning needs.
    • Q: Future of GPU offerings?
      • A: IBM is developing a specific GPU offering for customers with massive scale needs that exceed CPU/Spire capabilities, but the current strategic focus is on cost-effective CPU inference.
  • Other Key Points:

    • Case Study: A successful 5-day deployment for luxury brand Kate Spade (via Kathy from IBM) enabled a live virtual try-on feature integrated with Shopify, demonstrated at New York Fashion Week and featured in TechCrunch and WWD.
    • Technical Specifications: The S1112 demo utilized 10 cores, 512GB RAM, and 10TB storage to run models like Qwen 3.5 and Granite with response times of 2–3 seconds for text and ~9.5 seconds for complex image processing (compared to GPT-5.4).
    • Commercial Availability: On-prem software availability for IntelliX Prime is scheduled for November 14th.
    • Trial Program: A zero-cost, 30-day trial is available via IBM Cloud; 25% discount valid until year-end applies to on-prem activations.
    • Industry Trends: Prediction that QR codes will replace traditional barcodes in retail to facilitate AI-driven, voice-activated commerce experiences.
    • Event Cadence: Weekly workshops scheduled every Friday through mid-December to demonstrate use cases and configurations.
Company
ILOG SA
Keywords
ILOG SA