Semiconductor giants NVIDIA Corp. and Advanced Micro Devices Inc. have pledged a combined $4 billion to back Meta Platforms Inc.’s frantic push to finance its artificial intelligence (AI) infrastructure, cementing closer architectural ties between Silicon Valley’s leading chipmakers and its social media behemoth.
The investments follow Meta’s announcement of an aggressive capital raise totaling up to $97.5 billion.
Meta’s financing roadmap opens with an initial $35 billion tranche split across $22.5 billion in Class A common stock and $12.5 billion in mandatory convertible preferred stock, followed by an additional $60 billion equity offering slated for early 2027.
NVIDIA committed $3 billion to the initial equity offering, while AMD is chipping in $1 billion. Key hardware supplier Broadcom Inc. was also named as part of a collective supplier syndicate eyeing up to $5 billion in aggregate shares.
“Chip suppliers putting $4 billion behind a customer’s $97.5 billion AI buildout are helping finance their own order book,” said Mitch Ashley, vice president and practice lead for Software Lifecycle Engineering and AI-Native Software Engineering at The Futurum Group. “NVIDIA and AMD are locking in hyperscale demand, and AI infrastructure growth is leaning harder on capital that circulates between suppliers and customers.”
“CIOs should read it as a supply signal,” Ashley said. “When the largest buyers secure capacity this way, everyone else competes for what remains, and enterprises will feel it in GPU availability and inference pricing at their next contract renewal.”
The funding round deepens technical integrations between the chip titans and Meta’s infrastructure pipeline, led by Muse, Meta’s newly unveiled consumer-facing AI agent.
Muse acts as an autonomous assistant capable of executing complex everyday workflows across third-party applications, anchoring Meta CEO Mark Zuckerberg’s quest for personal superintelligence.
Under the expanded alliance, Meta will widen its deployment of NVIDIA’s next-generation Vera Rubin systems and Vera CPUs, alongside Spectrum-X Ethernet networking and BlueField-4 data processing units, aiming to lower inference costs and optimize power efficiency per watt. The companies will also collaborate on NVIDIA’s OpenShell safety frameworks and Sentry monitoring tools.
“AI agents represent a new computing paradigm — software that understands our goals, reasons, uses tools, and gets work done on our behalf,” NVIDIA CEO Jensen Huang said in a statement. “This changes how we use computers, from operating applications to working with intelligent agents.”
Zuckerberg noted the collaboration is structured to “deliver the best personal agents and frontier models for billions of people.”
AMD is simultaneously expanding its operational presence within Meta’s data centers. Meta currently operates AMD’s Helios rack scale architectures, utilizing EPYC server processors to run inference and training workloads for Muse.
“Together, we are building and deploying infrastructure at gigawatt scale and shaping future generations of silicon, software, and systems to enable Meta’s most ambitious superintelligence models,” AMD CEO Lisa Su said.
Wall Street delivered a mixed initial verdict on the news. Shares of AMD gained 3% on Tuesday, while NVIDIA was flat. Meta dipped 1% as markets absorbed the dilution risks tied to its historic share offering.
Meta is building its AI stack with two chipmakers that are competing to run its models. For developers, that could mean more choice in where they deploy Meta’s open-weight models, especially if NVIDIA and AMD both make them easier to run on their hardware, according to Stephanie Walter, practice leader for AI Stack & Enterprise Application Development at HyperFRAME Research.
“But choice on paper only goes so far,” Walter said. “Developers need models they can deploy and maintain without rewriting their software every time they change infrastructure.”
The transactions underscore a broader strategic trend across the semiconductor sector, where dominant hardware suppliers reinvest windfall profits directly into their largest hyperscale customers.
Dealroom data reveals NVIDIA has participated in 121 funding rounds this year alone — up from just 13 in 2022 — backing high-profile AI developers including OpenAI, Anthropic, Intel Corp., and xAI.
AMD has adopted a parallel strategy, engaging in 25 deals this year, headlined by an equity stake of up to $5 billion in Anthropic.
