AI Semiconductors
Blackwell
Blackwell is an NVIDIA computing architecture designed for advanced artificial intelligence and accelerated computing workloads. Blackwell-driven data center revenue reached $62.31 billion in NVIDIA's fiscal Q4 2026 (up 75% year-over-year), pushing full fiscal-year revenue to $215.9 billion, while NVIDIA's total supply purchase commitments nearly doubled to $95.2 billion to secure capacity through 2027.
Sources
Every claim traced to a primary source — evidence, recent activity, and insider filing behavior, all in one place.
Evidence
2 sources · 85% avg. source confidence
- NVIDIA Q4 and full fiscal-year 2026 earnings release — Blackwell drives NVIDIA data center revenue to $62.31B in Q4 FY2026, full-year revenue up 65% to $215.9B90% source confidence
- Futurum Group / NVIDIA Q3 FY2026 earnings coverage — NVIDIA's supply purchase commitments nearly double to $95.2B to meet Blackwell/Rubin demand through 202780% source confidence
Dependency Map
What Blackwell depends on below, and who depends on Blackwell above — click any node to make it the new center, 2 levels deep.
Connection type
Click any node to make it the new center. Scroll to zoom, drag to pan.
The Story So Far (last 6 months)
Auto-generated from this entity's dated milestones, relationship updates, and sourced evidence — not AI-written, just sorted.
Emerging Signals
Next-Generation AI Infrastructure Deployment
CriticalDeployment of new AI computing architectures increases demand across semiconductor manufacturing, memory and data-center infrastructure.
Blackwell's Timeline
A sourced, dated history of Blackwell's key moments — founding to present.
Jan 2024 · Blackwell launched
NVIDIA launched Blackwell as its new flagship AI GPU architecture, targeting frontier-scale training and inference clusters.
Jan 2026 · Fiscal Q4 2026 revenue peak
Blackwell-driven data center revenue reached $62.31 billion in NVIDIA's fiscal Q4 2026, up 75% year-over-year, pushing full fiscal-year revenue to $215.9 billion.
2026–2027 · Succeeded by Rubin
NVIDIA disclosed Rubin as Blackwell's planned successor platform, with TSMC advanced-packaging capacity already being reserved through 2027 for the transition.
How These Connections Evolved
How a relationship changed over time, not just its current state — sourced, dated, and traceable.
Blackwell ↔ NVIDIA
Feb 2026
Blackwell-driven data center revenue reached $62.31 billion in NVIDIA's fiscal Q4 2026 (up 75% year-over-year), pushing full fiscal-year revenue to $215.9 billion.
Now
Blackwell is developed by NVIDIA as part of its advanced accelerated computing platform.
Blackwell ↔ Taiwan Semiconductor Manufacturing Company
2024–2025
Blackwell (GB200/GB300) production faced widely reported ramp delays tied to TSMC's CoWoS packaging and yield constraints before scaling through 2025 as TSMC's advanced packaging capacity expanded.
Now
Production of advanced Blackwell processors depends on sophisticated semiconductor manufacturing capabilities.
Blackwell ↔ High Bandwidth Memory
2024–2025
Blackwell systems standardized on HBM3E memory sourced from SK hynix, Micron, and Samsung, with HBM supply constraints contributing to early production bottlenecks before capacity scaled.
Now
Blackwell AI systems depend on high-bandwidth memory for high-performance data processing.
Blackwell ↔ Advanced Semiconductor Packaging
2024–2025
Blackwell's reliance on CoWoS advanced packaging to integrate compute and HBM components made TSMC's packaging capacity ramp (30-40k to 75-80k wafers/month) a direct gating factor on Blackwell shipment volume.
Now
Blackwell products depend on advanced semiconductor packaging to integrate processing and memory components.
Blackwell ↔ AI Data Centers
Feb 2026
Blackwell deployment across hyperscaler AI data centers drove data-center revenue to $62.31 billion in NVIDIA's fiscal Q4 2026, with TrendForce projecting Blackwell (GB300/B300) would account for over 70% of NVIDIA's high-end GPU shipments in 2026.
Now
Blackwell systems are designed for deployment in advanced artificial intelligence data-center infrastructure.
Blackwell ↔ NVLink
2024–2025
Blackwell's GB200/GB300 NVL72 rack-scale systems relied on fifth-generation NVLink to connect up to 72 GPUs as a single coherent fabric, a key architectural differentiator NVIDIA marketed for large-scale model training.
Now
Blackwell GPUs depend on fifth-generation NVLink to operate as a coherent multi-GPU fabric for large-scale model training.