Nvidia and AMD Accelerators: Roadmaps, Pricing, Supply — 2026-10-08
Nvidia enters October with record-high stock driven by a $108 billion Q3 revenue guidance and the ramp-up of its Vera Rubin platform, while AMD faces a 10% price hike on Instinct chips due to rising TSMC costs. Supply chain data indicates GPU lead times have normalized to 30–40 weeks, though HBM4 shortages remain a critical bottleneck for Rubin production. Meanwhile, Chinese domestic AI chips are gaining significant market share, reducing reliance on restricted Nvidia hardware.
Nvidia and AMD Accelerators: Roadmaps, Pricing, Supply — 2026-10-08
Top developments
Nvidia’s $108B Guidance and Rubin Ramp Drive Stock Highs
Nvidia reported strong fiscal Q2 results with revenue hitting $96.2 billion, beating expectations, and issued a staggering Q3 guidance of $108 billion. The company confirms that its next-generation Vera Rubin platform has entered full production, with shipments expected in Fall 2026 offering 10x agent throughput compared to Blackwell. Analysts forecast Rubin shipments to reach approximately 5.7 million GPUs in 2026 as Blackwell volumes taper, signaling a rapid transition in Nvidia’s data center revenue mix. This aggressive growth outlook keeps AI infrastructure spending at historic highs, despite emerging concerns about financing risks in the sector.
AMD Instinct MI400 Ships Amid 10% Price Hikes
AMD is preparing to ship its Instinct MI400 series, featuring 432GB of HBM4 and targeting direct competition with Nvidia’s Blackwell and Rubin platforms. However, partners have been warned of a 10% price increase on Instinct, Radeon, and chipset products for Q4 2026, attributed to rising TSMC wafer costs. Despite the price hikes, AMD’s stock surged to a historical high, supported by Citi raising its target price to $800 due to strong CPU and AI chip demand. The MI400 promises up to 40 petaflops FP4 performance, aiming to capture a larger slice of the AI accelerator market from Nvidia’s dominant position.
GPU Supply Normalizes but HBM4 Remains Constrained
TrendForce reports that GPUs have become the least constrained part of the AI hardware supply chain, with lead times stabilizing at 30–40 weeks, which is considered the normal range. However, the ramp-up of Nvidia’s Rubin systems is draining supplies of HBM4, DRAM, MLCCs, and substrates, creating new bottlenecks for OEM buyers. Samsung and other memory manufacturers are increasing HBM4 output, but supply remains a critical dependency for Rubin system delivery. This shift from general GPU scarcity to specific component shortages suggests that memory bandwidth and capacity are now the primary limiting factors for next-gen AI clusters.
China’s Domestic AI Chips Surpass Nvidia in Market Share
Domestic AI chip makers in China, led by Huawei’s Ascend series, have captured over 50% of the local market share, surpassing Nvidia’s restricted offerings. Nvidia acknowledged in earnings calls that H200 shipments to China now account for less than 1% of its data center revenue, reflecting the impact of sustained US export controls and China’s push for self-sufficiency. Despite these restrictions, leaked documents reveal that Chinese state-backed entities are still financing purchases of restricted Nvidia Blackwell chips, indicating continued demand for top-tier hardware where possible. This bifurcation means Nvidia’s China strategy is increasingly limited to compliant variants, while domestic rivals fill the volume gap.
Local view
In South Korea, media outlets like Maeil Business Newspaper highlight AMD’s stock surge and market cap exceeding $1 trillion, attributing it to increased CPU demand alongside AI GPUs. Newspim reports that AMD plans to significantly expand semiconductor supply in 2027 to meet growing AI demand, positioning itself more aggressively against Nvidia. In China, Sina Finance discusses Nvidia’s partnership with SpaceX and Microsoft, noting how equity stakes and exclusive supply deals are creating "chip-equity-ecosystem" closed loops that lock in future demand.
Context & numbers
- Nvidia Q3 Guidance: $108 billion, significantly above street estimates.
- AMD Price Hike: 10% increase on Instinct and Radeon products effective Q4 2026 due to TSMC costs.
- GPU Lead Times: Normalized to 30–40 weeks per TrendForce.
- DGX B300 Pricing: Listed at approximately $646,878, with DGX B200 at $562,154.
- China Market Share: Domestic AI chips now hold >50% of the Chinese market, displacing Nvidia.
On the radar
- Rubin Shipment Ramp: Monitor Q4 2026 reports for actual Rubin shipment volumes against the 5.7 million unit forecast.
- HBM4 Supply Tightness: Watch for further component shortages in DRAM and substrates as Rubin production scales up.
- SpaceX Financing: Track the progress of SpaceX’s proposed $40 billion financing round intended to secure 1.21 million Nvidia GPUs.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
