Synthszr Charts — die großen AI-Marken im Wettkampf ums Podium
synthszr charts
amd

Amd · 7× · last seen Sep 01, 2026

47
Momentum

AMD Helios is a rack-scaled AI infrastructure solution that combines 72 AMD Instinct MI455X GPUs (CDNA-5 architecture), 18 AMD EPYC "Venice" CPUs (Zen 6), and Pensando networking technology in an open, OCP-standards (Open Rack Wide) based rack. The system is positioned as a direct competitor to Nvidia's Vera Rubin platform and delivers up to 2.9 ExaFLOPS FP4 Inference performance and 31 TB of pooled HBM4 memory per rack, according to AMD. Microsoft Azure is the anchor customer and will deploy Helios starting in the second half of 2026; customer shipments begin in the same timeframe.

Momentum trend
08.06.06.09.

Features

Manufacturing Process (nm)MI455X compute chiplets on TSMC 2nm (N2) and 3nm; EPYC Venice CCDs on TSMC 2nm (N2)
LicenseOpen standards (OCP, UALink, Ultra Ethernet); software stack ROCm (open source)
PlatformRack-scale system: 72x Instinct MI455X GPUs, EPYC 'Venice' CPUs, Pensando networking, ROCm software, OCP Open Rack Wide standard
Priceapprox. $5–5.5 million per rack (analyst estimate, not officially confirmed by AMD)
Compute Performance (FLOPS/TOPS)Up to 2.9 exaFLOPS FP4 (inference) and 1.4 exaFLOPS FP8 (training) per rack
Release DateUnveiled Oct. 2025, full specs July 2026, shipping H2 2026
Memory31 TB HBM4 total per rack, 1.4 PB/s aggregate memory bandwidth (432 GB HBM4 per GPU)
AvailabilityShipping to Microsoft Azure begins H2 2026; full market availability/mass production not until approx. Q2 2027 due to limited HBM4 supply

More products in this category: AI Inference Hardware

Sources (7)

Company Analysis: AMD

As of 26/08/2026
HOLDSynthszr Vote

HOLD is the most data-consistent stance at the current setup: fundamentals are strong (Q2 2026 revenue +50% Y/Y; Data Center +107% Y/Y) and management guided another sequential step-up in Q3 revenue (~$13.0B ±$0.3B). (sec.gov) However, valuation is elevated on trailing earnings (P/E ~123x at $479.18 on Aug 26, 2026), and near-term profitability guidance is not expanding (Q3 non-GAAP gross margin ~56% flat), which increases sensitivity to any change in growth expectations or investment/cash-generation concerns highlighted by recent analyst commentary. (spglobal.com) In EUR terms, the latest tool price is approximately €410.9 using the ECB USD/EUR reference rate (USD per EUR 1.1662 on Aug 25, 2026). (lb.lt)

Summary

Advanced Micro Devices (AMD) designs high-performance semiconductors and platforms across CPUs (EPYC for servers, Ryzen for PCs), GPUs/accelerators (Instinct for data center AI/HPC, Radeon for graphics), and adaptive/embedded solutions. AMD’s core competencies are chiplet-based design, leading CPU performance-per-watt, and an increasingly integrated data center stack that combines compute (EPYC), accelerators (Instinct), networking (Pensando), and software (ROCm) into system-level offerings. Market position has strengthened materially in data center. In Q2 2026, AMD reported record revenue of $11.5B (+50% Y/Y) with Data Center revenue of $6.7B (+107% Y/Y), indicating that AI and server demand are now the primary growth engine. Gross margin was 54% (non-GAAP 56%), reflecting a richer mix and normalization versus the prior-year period that included inventory-related charges tied to export restrictions. Diluted EPS was $1.38 (non-GAAP $1.66). Management guided Q3 2026 revenue to about $13.0B ± $0.3B and non-GAAP gross margin to ~56%, implying continued sequential growth but with margins held flat near current levels. (ir.amd.com) Recent months’ developments have been dominated by AMD’s AI platform roadmap and partnerships. At Advancing AI 2026, AMD highlighted a full-stack approach including Instinct MI400 series GPUs and Helios rack-scale solutions, positioning AMD to compete more directly for large AI infrastructure deployments. (ir.amd.com) External analyst commentary in August 2026 noted that Q2 results and raised outlook drove upward revisions to Q3 and FY2026 estimates, while also flagging that investment intensity and cash generation versus earnings are key investor focus areas as AMD scales AI. (spglobal.com) Valuation remains demanding on trailing metrics: the tool-reported P/E is ~123x at a price of $479.18 (Aug 26, 2026). Converted using the ECB reference rate (USD per EUR 1.1662 on Aug 25, 2026), this is approximately €410.9 per share and ~€681.7B market cap (approx.). (lb.lt) Short- to medium-term outlook is primarily a function of (1) the pace and profitability of the data center AI ramp (Instinct + systems), (2) EPYC share gains and pricing discipline, and (3) whether operating expense growth stays controlled relative to revenue growth as AMD executes its roadmap and partnerships.

Key Takeaways

  1. Q2 2026 showed a step-change in scale: $11.5B revenue (+50% Y/Y) with Data Center at $6.7B (+107% Y/Y), confirming data center AI/server as the dominant growth driver. (sec.gov)
  2. Profitability improved with mix: GAAP gross margin 54% and non-GAAP 56% in Q2 2026; Q3 non-GAAP gross margin guided ~56% despite higher revenue, suggesting early AI ramp economics are not yet expanding margins. (sec.gov)
  3. Guidance implies continued momentum: Q3 2026 revenue guided ~$13.0B ±$0.3B, indicating sequential growth and management confidence in 2H 2026 acceleration. (reddit.com)
  4. Roadmap and system-level positioning are central: AMD is marketing rack-scale solutions (Helios) and next-gen Instinct (MI400 series) as a full-stack alternative for AI infrastructure buyers. (ir.amd.com)
  5. Valuation is high on trailing earnings (P/E ~123x at $479.18 on Aug 26, 2026), increasing sensitivity to any deceleration in AI ramp, margin disappointment, or higher-than-expected operating expense growth.

Action Ideas

BUY

Action: accumulate on pullbacks with a 12–18 month horizon, anchored to execution against Q3 guidance and evidence of sustained data center acceleration into 2H 2026. The data-driven basis is (a) Q2 2026’s scale and mix shift (Data Center +107% Y/Y) and (b) management’s Q3 revenue guide (~$13.0B) that implies continued sequential growth. This is appropriate only if an investor accepts that near-term margins may remain around ~56% non-GAAP while the AI product/system ramp scales, and is underwriting the thesis primarily on revenue/segment mix and operating leverage over time rather than immediate margin expansion.

Horizon: 18 mo.

HOLD

Action: hold existing positions for 6–12 months while requiring confirmation that earnings growth translates into durable cash generation and that operating expense growth remains controlled. This stance is supported by (1) strong Q2 results and raised outlook commentary, but (2) explicit market focus on investment intensity and cash generation as AMD scales AI. The hold is justified if the investor wants additional quarters of evidence that the AI ramp improves cash conversion and does not structurally cap margins near current levels.

Horizon: 9 mo.

SELL

Action: reduce exposure over a 3–6 month horizon if the investment mandate cannot tolerate high-multiple volatility or if the position is sized assuming near-term margin expansion. The data-driven rationale is that AMD is priced at a very high trailing multiple (P/E ~123x) while management’s near-term margin guide is flat (~56% non-GAAP) even as revenue steps up, which can create asymmetric downside if the market re-prices growth or profitability expectations.

Horizon: 6 mo.

Contrarian Insights

  • Contrarian (more cautious than consensus): Even with strong revenue growth, the near-term profitability profile may not expand as quickly as investors expect. Management guided Q3 non-GAAP gross margin at ~56% (flat sequentially) despite a sizable revenue step-up, which can indicate that early AI/system ramps are margin-neutral to dilutive until scale improves. (reddit.com)
  • Contrarian (more constructive than consensus): AMD’s differentiation may increasingly be system-level rather than chip-level. The company is explicitly positioning Helios rack-scale solutions and MI400-series accelerators as an integrated stack, which can improve strategic relevance in large deployments and potentially support longer-duration revenue streams than discrete component sales alone. (ir.amd.com)

Sources (7)

Subscribe free. Unsubscribe the second it sucks.

High-signal news across AI, business, UX, and tech. Every morning.