AMD Invests $1 Billion in Meta as Helios Rack Systems Begin Testing for AI Infrastructure

7 Sources

Share

AMD commits $1 billion to Meta's equity offering as its first Helios rack systems enter testing at Meta's facilities. The expanded partnership positions AMD as a key supplier for Meta's superintelligence infrastructure, with millions of EPYC processors deployed and Instinct GPUs powering AI workloads.

AMD Commits $1 Billion to Meta's Equity Offering

AMD announced it will invest $1 billion in Meta's equity offering, deepening a strategic partnership that spans years of technical collaboration

1

. The investment comes as part of Meta's massive $35 billion share sale launched to fund AI infrastructure spending, with the broader plan reaching up to $95 billion

1

. The common stock portion was raised to $25 billion, with investor demand reaching five times the shares on offer. Shares were priced at $625 each, representing a 2.7% discount to Friday's close

1

. Combined with Nvidia's $3 billion pledge, the two semiconductor giants are contributing $4 billion to Meta's fundraising efforts

1

. AMD shares rallied 5.4% following the announcement, making it among the top gainers in the semiconductor complex

5

.

Helios Rack Systems Enter Testing Phase at Meta

AMD's first Helios rack systems are now installed and undergoing testing at Meta's facilities, marking a significant milestone in the AMD and Meta partnership

1

. The chipmaker expects broader deployment in the near future as testing progresses

4

. Helios rack systems represent AMD's rack-scale solution, packing 72 GPUs per rack to deliver high-density computing power

1

. "With Helios now running at Meta, we are taking that partnership to a new scale in the superintelligence era," said Lisa Su, chair and CEO of AMD

2

. The companies are building and deploying AI infrastructure at gigawatt scale, shaping future generations of silicon, software and systems to enable Meta's superintelligence models and services

3

.

Meta Deploys Millions of AMD EPYC Processors

Source: Wccftech

Source: Wccftech

Meta has deployed millions of EPYC processors across multiple generations, with AMD EPYC processors forming the foundation of Meta's infrastructure

3

. The social media giant serves as the lead customer for AMD's 6th Gen EPYC CPUs, based on the Zen 6 core architecture

3

. These 6th Gen EPYC CPUs power the computing behind Meta's Muse personal AI agent, handling the processing requirements for the AI assistant

1

4

. Meta is also collaborating with AMD on future processor designs, including a chip called Verano optimized for Agentic AI workloads

3

. This positions AMD as a critical supplier across Meta's computing environment, from general-purpose processing to specialized AI acceleration.

AMD Instinct GPUs Power Meta's AI Models

Meta is using AMD Instinct GPUs in production for training and operating recommendation engines, two core functions that drive user experience across Meta's platforms

1

3

. The company served as an early design customer for the Instinct MI450 series and will deploy a custom Instinct GPU built specifically for its workloads

1

3

. This collaboration will shape AMD's next-generation GPUs, the Instinct MI500 series, as the companies work together on hardware-model co-design

1

. AMD delivered day-zero support for Meta's Muse Spark Open models on Instinct GPUs, along with Muse Glimmer support on AMD's Ryzen AI PCs and Radeon GPUs

1

3

. Meta's technical paper details optimizations that delivered up to 1.5x improvement in end-to-end training throughput on AMD hardware

3

.

Software Collaboration Strengthens Open-Source Ecosystem

The AMD and Meta partnership extends beyond hardware into software optimization, with both companies advancing ROCm through PyTorch contributions

3

. The companies have expanded their software collaboration to large-scale AI training, optimizing recommendation models on Instinct MI350X GPUs and contributing improvements to PyTorch, Triton and ROCm

3

. Meta's technical paper represents the first hyperscaler report on productionizing AMD GPUs for large-scale training, with contributions upstreamed to open-source ecosystems

3

. This work ensures developers can leverage AMD GPUs out of the box, lowering barriers to adoption. Meta deploys a heterogeneous fleet comprising Nvidia GPUs, AMD GPUs and proprietary MTIA accelerators to reduce costs and mitigate vendor lock-in

3

.

Chipmakers Deepen Ties with Hyperscale Customers

Source: The Next Web

Source: The Next Web

The AMD investment in Meta reflects a broader trend of semiconductor companies taking equity stakes in their largest customers. AMD has participated in 25 funding rounds in 2026, up from just two in 2022, while Nvidia has joined 121 rounds this year

1

. AMD previously committed up to $5 billion to invest in Anthropic, which plans to run Claude on Helios systems

1

. These strategic investments cement balance-sheet and operational ties between chip suppliers and AI developers, securing long-term hardware allocations as hyperscaler demand intensifies

5

. The multi-year co-design roadmap positions Meta as an early-stage partner for AMD's future architectures, giving both companies visibility into upcoming technology requirements

5

. "AMD has been a key partner in building out the compute that's running our frontier models," said Mark Zuckerberg, Meta's founder and CEO

2

. As technology companies race to secure merchant silicon allocations, AMD's expanded agreement with Meta signals sustained tier-one hyperscale demand and positions the chipmaker to compete more effectively against rivals in the AI infrastructure market

5

.

© 2026 TheOutpost.AI All rights reserved