Delta and NVIDIA collaborate on autonomous driving systems

Key Takeaways

• Strategic partnership: Delta and NVIDIA are joining forces to fast‑track Level‑4 autonomous‑vehicle (AV) development.
• Core technology: NVIDIA Hyperion pairs two DRIVE AGX Thor modules (≈ 254 TOPS each) with the Halos safety suite, delivering redundant compute and real‑time safety monitoring.
• Delta’s edge: Expertise in xEV power‑train management, high‑performance in‑vehicle computing (HPC), and end‑to‑end system integration.
• Market impact: Combined ecosystems aim to shorten time‑to‑market for robotaxi fleets and premium ADAS solutions, opening new revenue streams in the smart‑mobility sector.

Overview of the Collaboration

Delta Electric Vehicle Systems (Delta) announced a joint effort with NVIDIA to accelerate the creation of next‑generation autonomous‑driving platforms. By merging Delta’s deep knowledge of automotive energy management, electric‑power‑train architecture, and in‑vehicle HPC integration with NVIDIA’s Hyperion reference platform and Halos safety framework, the partnership targets a rapid, production‑ready path to Level‑4 autonomy.

Why the Alliance Matters

• Technology synergy: Delta supplies the power‑train and energy‑budget optimization required for high‑performance compute, while NVIDIA provides the compute‑heavy sensor‑fusion and AI inference stack.
• Ecosystem leverage: NVIDIA’s global developer community (over 1 M registered users) and Delta’s relationships with Tier‑1 OEMs (e.g., Volkswagen, BYD) create a broad go‑to‑market channel.
• Safety focus: Halos delivers functional‑safety certification (ISO 26262 ASIL‑D) and continuous health‑monitoring of the compute pipeline, a critical requirement for robotaxi operators.

NVIDIA Hyperion Platform – Technical Snapshot

Source: NVIDIA product brief, 2024.

Delta’s Automotive Strengths

Data compiled from Delta’s 2023 technical whitepaper.

How the Joint Solution Works

1. Power Delivery – Delta’s energy‑management unit distributes a stable 400‑600 V rail to the Hyperion compute stack, maintaining < 5 % voltage ripple under peak loads.
2. Sensor Fusion – Dual Thor modules ingest data from up to 12 lidar, 8 radar, and 16 camera units (totaling > 2 TB/s raw bandwidth).
3. Safety Loop – Halos continuously monitors compute health, invoking a safe‑stop maneuver within 20 ms if a fault is detected.
4. AI Inference – Pre‑trained perception models (e.g., YOLO‑v7, PointPillars) run at 30 fps, delivering object detection with < 5 cm positional error.
5. System Integration – Delta’s software stack wraps NVIDIA’s DRIVE OS, exposing a unified API to OEMs for rapid feature rollout.

Market Outlook

The global Level‑4 autonomous‑vehicle market is projected to reach $45 B by 2032, growing at a CAGR of 38 % (MarketsandMarkets, 2024). By combining Delta’s power‑train efficiencies (up to 10 % range gain for a 300 kW vehicle) with NVIDIA’s proven compute platform, the partnership aims to cut development cycles from 24 months to under 12 months for robotaxi fleets.

Bottom Line

Delta and NVIDIA’s collaboration creates a tightly integrated, safety‑first autonomous‑driving stack that leverages Delta’s power‑train expertise and NVIDIA’s Hyperion compute + Halos safety system. The joint solution promises to deliver Level‑4 capabilities faster and more cost‑effectively, positioning both companies to capture a sizable share of the burgeoning smart‑mobility market.