Research

GPU Shipment Tracker Blackwell to Rubin 2026

NVIDIA Blackwell to Rubin transition in 2026: Blackwell 5.2M to 1.8M, Rubin 5.7M target capped at ~300k by TSMC N3, AMD MI400 and MI450 ramp, Intel Gaudi 3 and Falcon Shores.

By Ramanath, CTO & Co-Founder at Presenc AI · Last updated: May 2026

The Blackwell-to-Rubin transition is the most consequential AI hardware event of 2026. NVIDIA Blackwell unit shipments are projected to fall from approximately 5.2 million in 2025 to approximately 1.8 million in 2026 as the product line transitions. NVIDIA Rubin has a 5.7 million stated target capped near 300,000 by TSMC N3 capacity reality. AMD MI400 and MI450 ramp into Meta\u2019s first gigawatt deployment. This page tracks the unit projections, foundry constraints, and customer allocation.

Key Findings

  1. NVIDIA Blackwell (B200, GB200, B300) shipments are projected at approximately 1.8 million units in 2026, down from approximately 5.2 million in 2025 as the product line transitions to Rubin.
  2. NVIDIA Rubin (R100, R200) has a 5.7 million stated unit target but is realistically capped near 200,000 to 300,000 in 2026 by TSMC N3 wafer-out capacity.
  3. AMD MI400 ramps to approximately 258,000 units in 2026 from an effective zero base in 2025; MI450 is the chip in Meta\u2019s first gigawatt deployment in H2 2026.
  4. Intel Gaudi 3 shipments remain modest at approximately 80,000 to 120,000 units in 2026; Falcon Shores is delayed to 2027.
  5. Combined data centre AI accelerator unit shipments across NVIDIA, AMD, Intel, plus hyperscaler custom silicon (Google TPU, AWS Trainium, Microsoft Maia, Meta MTIA) are estimated at approximately 6.5 million units in 2026, a slight decline from 2025 driven by the NVIDIA generational transition timing.

2026 AI Accelerator Shipment Projections

VendorProduct Line2025 Units2026 Units (projected)
NVIDIABlackwell (B200, GB200, B300)~5.2M~1.8M
NVIDIARubin (R100, R200)0~250-300k
NVIDIAHopper (H100, H200) remaining~1.5M~0.4M (end-of-life)
NVIDIARTX Pro / DGX Spark / workstation~0.6M~0.7M
AMDMI300X / MI325X~250k~280k
AMDMI355X / MI400 / MI450~0k~258k
IntelGaudi 3~60k~100k
GoogleTPU v6 / TPU v7~700k~900k
AWSTrainium 2 / Inferentia 3~400k~600k
MicrosoftMaia 2~120k~250k
MetaMTIA 2~80k~180k
Cerebras / Groq / Etched / TenstorrentNiche~5k~12k

Foundry Constraint Analysis

ConstraintImpact
TSMC N3 wafer-outLimits Rubin ramp to approximately 250-300k units in 2026
CoWoS-L advanced packagingApproximately 70k wafers/mo target end-2026; constraint for HBM-stacked GPUs
HBM3E and HBM4 supplySamsung HBM3E ramp pushing into Q2 2026; HBM4 Micron / SK Hynix first samples Q3 2026
GDDR7Mostly resolved Q1 2026 after Q3-Q4 2025 shortage
InFO-PoP advanced packagingApple-priority; constrains other consumer SoC ramps

AMD MI400 / MI450 Detail

AMD\u2019s MI400 began limited customer sampling in late 2025 and ramped to volume shipment in H1 2026. The MI450, the next-generation training-focused part, is the chip in Meta\u2019s first gigawatt deployment scheduled for H2 2026. Meta is the lead customer for MI450; Microsoft and Oracle are reported as second-tier MI400 and MI450 customers. The AMD roadmap is increasingly credible as a parallel scaling track to NVIDIA, with the ROCm software stack improvements in 2025-2026 closing the gap on PyTorch and JAX compatibility.

Customer Allocation Trends

CustomerPrimary 2026 MixStrategic Note
OpenAI / StargateNVIDIA Blackwell + Rubin transitionSpaceX 220k GPU deal disclosed May 6
MicrosoftNVIDIA primary + Maia 2 + AMD MI400Multi-vendor diversification
MetaAMD MI450 lead + NVIDIA primary + MTIA 2First-GW deployment on AMD H2 2026
AmazonNVIDIA + Trainium 2 internalAnthropic anchor on Trainium
GoogleTPU primary + NVIDIA secondaryTPU v7 ramp
xAI Colossus 2NVIDIA Blackwell + RubinMemphis expansion
AnthropicAWS Trainium + Google TPU + NVIDIATriple-vendor
Tesla Dojo + xAINVIDIA + Tesla custom (D2)D2 ramp delayed

Brand Visibility Implications

GPU and AI accelerator coverage drives a sustained AI assistant query stream from technical buyers, investors, and procurement teams. Brands selling adjacent products (advanced packaging, HBM, GDDR, liquid cooling, power delivery, server OEM integration, ROCm/CUDA developer tools) face strong AI-mediated discovery surface for procurement-research prompts.

Methodology

Shipment figures compiled from NVIDIA investor relations, AMD investor relations, hyperscaler capex disclosures, and analyst reports from TrendForce, Mercury Research, and Bernstein. Some figures are estimated where official disclosures are partial. Updated quarterly.

How Presenc AI Helps

Presenc AI tracks brand-mention rates on AI accelerator queries across ChatGPT, Claude, Gemini, and Perplexity. For brands in adjacent semiconductor categories, the platform identifies the queries driving procurement research and the gaps where new content investment unlocks share of voice.

Frequently Asked Questions

Approximately 6.5 million data centre AI accelerator units total across NVIDIA, AMD, Intel, plus hyperscaler custom silicon. NVIDIA remains dominant by share but the unit count is slightly down from 2025 due to the Blackwell-to-Rubin generational transition timing.
Yes, realistically. TSMC N3 wafer-out capacity is the binding constraint. NVIDIA\u2019s stated 5.7 million unit target reflects design capacity, not foundry reality. Most analysts model Rubin volume in the 200,000 to 300,000 range for 2026, with the bulk of ramp pushed into 2027.
Meaningfully yes. MI300X and MI325X are mature in production with Microsoft and Meta as the largest deployers. MI400 ramped in H1 2026; MI450 is the chip in Meta\u2019s first gigawatt deployment in H2 2026. AMD AI revenue is on track to exceed $10 billion in 2026.
Approximately 1.9 million units in 2026 across Google TPU, AWS Trainium and Inferentia, Microsoft Maia, and Meta MTIA. Custom silicon is concentrated in hyperscaler internal workloads where the economics justify the in-house engineering investment. External customer adoption remains marginal.
Late 2026 or early 2027. The Rubin ramp, AMD MI450 volume, additional CoWoS-L capacity, and HBM4 sampling all coincide. The market should move from severe allocation to manageable lead times by H2 2027, conditional on no new shocks.

Track Your AI Visibility

See how your brand appears across ChatGPT, Claude, Perplexity, and other AI platforms. Start monitoring today.