AI's Cash Splash Continues: Meta Platforms and Broadcom Expand Custom Chip Partnership

Hyperscalers plan to spend between $635 billion and $665 billion on AI infrastructure in 2026 alone, a 67% jump from 2025. That cash is fueling deals at a record pace. Around Nvidia's (NVDA) GTC 2026 in March, the chipmaker expanded agreements with AWS, Google Cloud, Microsoft (MSFT) Azure, and Oracle (ORCL) OCI.

Google also deepened its compute partnership with Anthropic. AWS struck new ground with OpenAI, while Intel (INTC) teamed up with Google. Last week, Meta Platforms (META) committed an additional $21 billion to AI cloud provider CoreWeave (CRWV) through 2032. Now comes the latest move: Meta and Broadcom’s (AVGO) expanded custom-chip partnership.

Let’s look at what this agreement actually delivers, what it means for both companies, and why it matters for investors who own META or AVGO shares.

The AI Infrastructure Frenzy in Numbers

Four hyperscalers—Microsoft, Alphabet (GOOG) (GOOGL), Amazon (AMZN), and Meta—guided for combined capital expenditures of roughly $650 billion this year, with some estimates reaching $720 billion when Oracle is included. Meta alone expects $115 billion to $135 billion in 2026 capex, nearly double its 2025 spend. 

OpenAI closed a $122 billion funding round on March 31 at an $852 billion valuation, backed by Amazon, Nvidia, SoftBank (SFTBY), and Microsoft participation. Anthropic secured expansions with Google, Microsoft (about $5 billion), and Nvidia (about $10 billion). No matter how you slice it, the ecosystem is all-in on scaling compute.

META stock jumped 4.4% on Tuesday and has been on the rise throughout trading today, while AVGO rose less than 1% yesterday but is up over 3% today.

What the Meta-Broadcom Deal Actually Says

The two companies are extending their collaboration to co-develop multiple generations of Meta’s MTIA (Meta Training and Inference Accelerator) chips through 2029. Broadcom will supply technology based on its XPU custom-accelerator platform, advanced packaging, and high-speed Ethernet networking. The initial phase commits more than 1 gigawatt of computing capacity—enough to power roughly 750,000 average U.S. homes—with a roadmap for multiple gigawatts over time.

Meta CEO Mark Zuckerberg called it the foundation needed “to deliver personal superintelligence to billions of people.” Broadcom CEO Hock Tan described it as the start of a sustained, multi-generation roadmap that plays to the company’s leadership in custom accelerators and networking. In short, Meta gets purpose-built silicon optimized for inference, recommendations, and generative AI workloads. Broadcom gets a multi-year revenue stream from one of the biggest AI spenders on the planet.

What It Means for META, AVGO, and Investors

For Meta, this expands its portfolio approach to silicon. The company already runs a mix of accelerators; MTIA chips target specific workloads where custom hardware can deliver better performance per dollar. With 2026 capex guidance at $115 billion to $135 billion, every efficiency gain counts. Meta’s trailing 12-month revenue hit $200.97 billion with 23.8% year-over-year (YoY) quarterly growth, and its forward P/E sits at 21.12—reasonable for a company turning AI spend into user-facing features across Instagram, WhatsApp, and Threads.

For Broadcom, the deal builds on momentum. In its fiscal Q1 2026, AI revenue doubled 106% to $8.4 billion and made up 43% of the company’s $19.31 billion total revenue. CEO Hock Tan has a line of sight to more than $100 billion in annual AI chip revenue by 2027. Adding Meta’s multi-gigawatt, multi-generation commitment strengthens that outlook and diversifies beyond any single hyperscaler.

That said, risks remain. Meta’s heavy capex will pressure free cash flow in the near term. Broadcom faces the usual semiconductor-cycle volatility and execution risk on 2nm-class chips. Granted, both companies have delivered strong results so far: Meta’s revenue growth outpaced the broader market, and Broadcom’s AI segment is accelerating faster than its non-AI business.

What Do Analysts Think About Meta Platforms

According to analysts followed by Barchart, 56 analysts cover the stock and give it a consensus "Strong Buy" rating with an average score of 4.66 out of 5. That breaks down to 45 "Strong Buy" recommendations, three "Moderate Buy," and seven "Hold." 

Their mean 12-month price target stands at $855.96, with a high of $1,015 and a low of $676. That points to roughly 30% upside potential from recent trading levels around $660. In short, analysts see the AI infrastructure push—and this Broadcom partnership—as a clear tailwind.

Key Takeaway

In any case, the Meta-Broadcom partnership is another data point that hyperscaler AI spend is not slowing—it is broadening. Savvy investors who already hold META or AVGO get reinforced exposure to the infrastructure layer without needing to chase every new funding round or cloud contract. The deal does not change near-term earnings estimates dramatically, but it confirms both companies sit squarely in the middle of the multi-year AI build-out. 

If your portfolio needs more direct AI-infrastructure exposure, this expanded partnership gives you a clear, measurable reason to keep—or add—positions in both stocks. The cash splash continues, and these two just turned up the volume.