Stack L2

AMD

Merchant #2 AI training silicon; OpenAI 6GW + Oracle 50k MI450 wins, ROCm still the watch-point

HQ US

Fit Score

avg of 4 segments

78/100

Growth

Comp

—

Learning

Stability

Culture

Recent Business Signals

No business signals recorded in the last 90 days.

Strategic Position

What they do best

Scaling a credible merchant GPU alternative at hyperscale — data-center revenue reached ~$4.3B in Q3 FY2025 (signal 4e06d83d) on MI350's ~4x generational compute uplift (signal 00c21fd4).

Their bet

MI450/'Helios' rack-scale systems backed by the OpenAI 6 GW warrant deal (signal 81014194) bet that AMD wins at rack level, not just chip level, against a projected ~$1T 2030 AI TAM (signal 65d05619).

Top risk

The 1 GW OpenAI / 50k-Oracle MI450 ramps (signals 81014194, 63f198be) slip past their stated Q3–2H 2026 starts, or ROCm gaps stall conversion — both checkable against deployment reporting by end of 2026.

Compared to peers

Direct competitor

Cerebras Systems

Cerebras sells wafer-scale single-chip systems into niche HPC/pharma; AMD competes as volume merchant GPU at hyperscaler scale (Oracle 50k MI450, signal 63f198be).

Substitute

Groq

Groq is inference-only deterministic LPU silicon; AMD Instinct targets training + inference with a broader ROCm software surface Groq does not address.

Why someone would join

1.Q3 FY2025 hit record ~$9.2B revenue (+36% YoY), data center ~$4.3B (signal 4e06d83d) — this is a funded revenue ramp, not a pre-revenue silicon bet.
2.OpenAI's 6 GW commitment (signal 81014194, Oct 2025) plus Oracle's 50,000 MI450 order (signal 63f198be) give MI450 work guaranteed 2026 deployment, matching the active chiplet/packaging and physical-design hiring.

Recent Hiring (60 days)

physical design
1
place and route
1
timing closure
1
chiplet architecture
1
advanced packaging
1

Reverse-Hype Watch

!Company-flagged ROCm 'watch-point' / NVIDIA-CUDA substitute narrative unbacked by hiring: sampled talent_signals are 5/5 silicon (ASIC RTL, physical design, DV/UVM, chiplet, power) with zero compiler/ML-systems software roles.
!FAD 2025-11-11 '35%+ annual revenue growth over 3-5 years' against a vendor-cited '~$1 trillion data center AI TAM by 2030' is an aspirational long-range management target, not booked revenue.
!Funding/capacity ceiling risk: FAD 2025-11-11 '35%+ annual revenue growth over 3-5 years' against vendor-cited '~$1T data center AI TAM by 2030' is an aspirational management target, not booked revenue.
!Product launch / hiring-specialty mismatch: ROCm named the explicit watch-point, yet 5/5 sampled reqs (posted 2026-02-25 → 2026-05-06) are silicon/hardware (RTL, physical design, DV/UVM, chiplet, power) with zero compiler/ML-systems roles.
!FAD 2025-11-11 '35%+ annual revenue growth over the next 3-5 years' against a vendor-cited '~$1 trillion data center AI TAM by 2030' is an aspirational management target, not booked revenue.

→ Get this data as JSONLast updated: May 16, 2026

Frequently asked about AMD →