Macroplane › Investment Theses › Only credible merchant GPU alternative

Only credible merchant GPU alternative

AI supply-chain thesis — mapping bottlenecks, focus companies, and supply-chain exposure for investors.

**Bottleneck theme:** Custom Silicon **Focus:** $AMD — ADVANCED MICRO DEVICES INC AMD is the only credible merchant GPU alternative to NVIDIA at training and high-end inference scale. The MI300X / MI325X / MI355 / MI400 roadmap, paired with rapidly maturing ROCm software and meaningful HBM3E content per accelerator, has made AMD the procurement-backstop choice for hyperscalers and frontier labs unwilling to be 100% sole-sourced on NVIDIA. The OpenAI agreement to buy 6 GW of MI-series GPUs (announced 2025), combined with Microsoft, Meta, and Oracle MI355 deployments, validates that the bookings are real rather than narrative. The investment case is share recovery in data-center GPU plus margin mix expansion, with optionality from EPYC server CPUs continuing to take share from Intel and an embedded/Xilinx franchise providing baseline revenue. The bear case is structural: NVIDIA's CUDA software moat compounds over time, MI-series gross margins are below NVIDIA's by hundreds of basis points, and AMD has a capacity allocation problem (sharing TSMC CoWoS-L and HBM3E with NVIDIA in a constrained market). Treat as the highest-quality merchant alternative to NVIDIA, with continued execution required to convert design wins into sustained share.

Focus companies in this thesis (1)

  • ADVANCED MICRO DEVICES INC (AMD)

Supply-chain categories covered

  • AI GPUs — Compute accelerators and GPUs powering AI training, inference, and large language models.
  • Fabless Chip Design — Companies designing chips without owning fabs
  • Foundry / Fab Services — Contract semiconductor manufacturing — wafer fabrication for fabless and partially-fabless customers, spanning leading-edge logic, mature-node analog/mixed-signal, RF, and specialty processes (BCD, BiCMOS, SiC, SOI).
  • HBM — High Bandwidth Memory — 3D-stacked DRAM (HBM2E/HBM3/HBM3E/HBM4) connected via through-silicon vias, delivering 1+ TB/s of bandwidth per stack. Co-packaged with GPUs, TPUs, and custom AI accelerators for datacenter AI training/inference and HPC workloads.
  • Server & System Assembly (ODM/EMS) — ODMs and EMS firms assembling servers, racks, and AI systems from components.
  • Hyperscalers — Major cloud operators (AWS, Azure, GCP, Meta, Oracle, Alibaba, Tencent, Baidu, Naver) and tier-2 / neocloud providers (DigitalOcean, OVHcloud, Rackspace, Kingsoft) tracked as a demand signal across multiple theses (photonics, HBM, AI accelerators, power, cooling). Excludes SaaS apps, telcos, REITs, and IT services firms.
  • AI Training Accelerators — GPU and AI accelerator chips designed for training large AI models, the core demand drivers.
  • Advanced Packaging — 2.5D/3D packaging, CoWoS, chiplets, fan-out wafer-level packaging

Thesis milestones & bottleneck markers

  • AMD MI300 revenue inflection — AMD — MI300 series reaches meaningful revenue contribution
  • HBM supply stabilization — HBM pricing normalizes as capacity grows
  • AMD AI GPU market share — AMD — AMD captures datacenter GPU share
  • AMD hyperscaler wins — Major design wins beyond Microsoft
  • ROCm maturity vs CUDA — AMD — AMD software stack gains traction

Related baskets

  • Custom Silicon

More AI supply-chain theses

  • GE Vernova Stock ($GEV): The Power-and-Grid Pick for the AI Electricity Crunch
  • THE optical contract manufacturer
  • Utility transformers — the actual grid bottleneck (2–4yr lead times)
  • Comfort Systems (FIX) — MEP/HVAC contractor riding the AI data center super-cycle
  • Mission-critical electrical/mechanical construction
  • Only US rare-earth producer at scale
  • Celestica Stock (CLS): AI Hardware Manufacturing for Hyperscalers
  • POET Stock Thesis: The Photonic-Integrated CPO Engine Outlier
  • Boyd $9
  • Kulicke & Soffa (KLIC) — Strait of HBM TCB
  • Custom-silicon duopoly with AVGO
  • WaveLogic 6 Extreme 1

Frequently asked questions

What is the Only credible merchant GPU alternative thesis about?

**Bottleneck theme:** Custom Silicon **Focus:** $AMD — ADVANCED MICRO DEVICES INC AMD is the only credible merchant GPU alternative to NVIDIA at training and high-end inference scale. The MI300X / MI325X / MI355 / MI400 roadmap, paired with rapidly maturing ROCm software and meaningful HBM3E content per accelerator, has made AMD the procurement-backstop choice for hyperscalers and frontier labs unwilling to be 100% sole-sourced on NVIDIA. The OpenAI agreement to buy 6 GW of MI-series GPUs (announced 2025), combined with Microsoft, Meta, and Oracle MI355 deployments, validates that the bookings are real rather than narrative. The investment case is share recovery in data-center GPU plus margin mix expansion, with optionality from EPYC server CPUs continuing to take share from Intel and an embedded/Xilinx franchise providing baseline revenue. The bear case is structural: NVIDIA's CUDA software moat compounds over time, MI-series gross margins are below NVIDIA's by hundreds of basis points, and AMD has a capacity allocation problem (sharing TSMC CoWoS-L and HBM3E with NVIDIA in a constrained market). Treat as the highest-quality merchant alternative to NVIDIA, with continued execution required to convert design wins into sustained share.

Which companies does the Only credible merchant GPU alternative thesis focus on?

The Only credible merchant GPU alternative thesis on Macroplane focuses on ADVANCED MICRO DEVICES INC (AMD).

Which supply-chain categories does the Only credible merchant GPU alternative thesis cover?

It covers AI GPUs, Fabless Chip Design, Foundry / Fab Services, HBM, Server & System Assembly (ODM/EMS), Hyperscalers, AI Training Accelerators, Advanced Packaging.

Browse all AI supply-chain theses · Macro trends · Industries · Product categories · Baskets