NVIDIA and AMD announced plans on October 27, 2026, to invest a combined $4 billion in Meta’s equity offering while expanding their separate partnerships to supply the computing infrastructure behind Meta’s AI models and agents. NVIDIA says it plans to invest $3 billion; AMD says it plans to invest $1 billion. The announcements are significant, but they are not evidence that either investment has already closed.
The unusual part is the combination: both companies sell technology that Meta uses to run AI workloads, and both are also becoming investors in the same customer. That can align the companies’ long-term interests, but it also makes the financial and commercial relationships worth examining separately.
What NVIDIA announced
In its October 27 announcement, NVIDIA said it would invest $3 billion in Meta’s equity offering and deepen work on Meta’s Muse assistant and related models. The stated technical goal is to improve training and inference performance while lowering the total cost of ownership of the systems running those models.
Meta also intends to expand its use of NVIDIA infrastructure, including Vera Rubin systems and Vera CPUs, Spectrum-X Ethernet networking and BlueField-4 networking technology. NVIDIA and Meta said they would consider further work involving NVIDIA OpenShell for safer agent deployment and NVIDIA Sentry for monitoring agent behaviour. Those extra collaborations are possibilities under consideration, not confirmed completed rollouts.
What AMD announced
AMD’s October 27 announcement separately said it would invest $1 billion in Meta’s equity offering. AMD described a broad technical partnership spanning processors, accelerators, rack-scale systems, networking and software.
AMD said initial Helios rack-scale systems are installed at Meta and being tested ahead of broader deployment. It also said EPYC processors support the dedicated computing environments behind Muse, while Instinct GPUs are already used in production for AI training, ranking and recommendation workloads. Those are distinct status points: systems being tested should not be described as a fully completed, company-wide deployment.
| Company | Announced investment | Technical status disclosed |
|---|---|---|
| NVIDIA | $3 billion | Expanded Muse collaboration and planned wider infrastructure deployment |
| AMD | $1 billion | Helios systems installed and being tested; EPYC and Instinct technology already supports some Meta workloads |
| Combined | $4 billion | Two separate announcements, not one joint investment agreement |
How the $4 billion fits Meta’s fundraising plan
The chipmakers’ announced investments are only part of Meta’s larger effort to fund AI infrastructure. Reuters reported on October 26 that the potential total of Meta’s equity-raising plans had reached $97.5 billion. The broader plan also includes a separate at-the-market share-sale programme that could raise additional capital over time.
For scale, $4 billion is about 4.1% of a $97.5 billion potential total (4 ÷ 97.5). That is a comparison of announced amounts, not a claim that all of Meta’s planned financing is complete or that both chipmakers’ investments have settled. NVIDIA’s statement explicitly describes its investment as proposed and refers to anticipated completion and closing conditions.
Why supplier investments in customers draw scrutiny
When a hardware supplier invests in a company that may spend heavily on that supplier’s products, the relationship can reinforce itself: the customer gains capital to build infrastructure, while the supplier may gain a stronger commercial position. Analysts and investors sometimes describe the wider pattern as circular financing. The concern is about how dependent future demand and revenue could become on financing arrangements that support one another; it is not, by itself, proof of misconduct or artificial sales.
Meta’s announcements do not say it will buy exclusively from NVIDIA or AMD. Meta has also described a portfolio approach to AI computing, including its own MTIA accelerators and a separate custom-silicon collaboration with Broadcom. The relevant question is therefore how Meta balances different chips and systems across different workloads—not whether one supplier has won every part of its infrastructure.
What this means for Meta users and creators
There is no immediate change to Instagram creator payouts, Facebook or Instagram recommendation rules, or Muse availability announced in these investment statements. More capable and efficient infrastructure could support AI services over time, but neither company provided a user-facing launch date or a guaranteed performance improvement for a specific Meta app in these releases.
For context, see our earlier explainer on Meta Muse AI, its tasks and privacy controls. The infrastructure side is also part of the wider build-out covered in our report on AI data-centre funding and the Nscale IPO filing.
What remains unconfirmed
The announcements do not establish the final closing dates or all transaction terms for the two investments. They also do not give a complete timetable for broader hardware deployment, quantify realised savings from the expanded partnerships, or promise that the possible OpenShell and Sentry work will proceed. Those details should be treated as open until the companies provide more information.
The clearest takeaway is narrower than a claim that either chipmaker has secured Meta’s entire AI business: on October 27, NVIDIA and AMD each announced an investment in Meta alongside deeper technical work. The combined $4 billion strengthens the financial ties, while the infrastructure details show that the companies’ roles and deployment stages remain different.
