Meta Platforms Does It Again, Signs a Multi-Billion Deal With Google

Meta Platforms (NASDAQ:META | META Price Prediction) has been aggressively forging major agreements with leading chipmakers to supercharge its AI infrastructure expansion. Deals with Nvidia (NASDAQ:NVDA) and Advanced Micro Devices (NASDAQ:AMD) have already positioned the company to deploy millions of advanced GPUs for training massive AI models like Llama. This diversification strategy not only provides leverage in negotiating better pricing but also mitigates risks associated with over-reliance on a single supplier, such as supply chain disruptions or monopolistic pricing pressures. 

True to form, Meta has done it again, reportedly inking another multi-billion-dollar deal with Google to access specialized AI hardware, further broadening its ecosystem amid surging demand for computational power in generative AI and metaverse applications.

Bolstering the AI Infrastructure

Meta’s infrastructure buildout is nothing short of monumental, aimed at supporting its ambitious goals in artificial intelligence, including open-source models and real-time applications across its platforms like Facebook, Instagram, and WhatsApp. The company is investing heavily in data centers, servers, and networking to handle the exponential growth in AI workloads. 

For 2026, Meta has committed to capital expenditures ranging from $115 billion to $135 billion, a sharp increase from the $72 billion spent in 2025, primarily driven by AI-related infrastructure. This spending spree underscores CEO Mark Zuckerberg’s vision for “superintelligence,” but it also strains the balance sheet with rising costs.

Meta’s deal with Nvidia involves a multiyear partnership for millions of Blackwell and Rubin GPUs, ensuring a steady supply of high-performance chips for on-premises and cloud-based AI training. Similarly, the agreement with AMD — valued at up to $60 billion over five years (with some estimates saying $100 billion) — includes deploying 6 gigawatts of Instinct GPUs, and gives Meta the option to acquire a 10% stake in the chipmaker. These pacts complement each other, allowing Meta to mix and match hardware for optimal efficiency.

Now, adding Google to the fold introduces tensor processing units (TPUs), custom AI accelerators designed for machine learning tasks. This expands Meta’s toolkit beyond GPUs, potentially optimizing costs for specific workloads like inference.

The Google TPU Partnership

The latest deal with Google is a multi-year lease agreement worth billions, granting Meta access to TPUs for training and running new AI models. Discussions are underway for Meta to purchase TPUs outright starting in 2027, installing them in its own data centers for greater control. This follows Google’s formation of a joint venture with an investment firm to lease TPUs to external clients, with Meta as an early adopter.

For Meta, the gains are clear: further diversification reduces dependency on Nvidia’s dominant GPUs, potentially lowering costs through competition and enabling specialized optimizations for its AI ecosystem. It also hedges against shortages in the GPU market. 

Google benefits by scaling its TPU business, generating new cloud revenue streams, and challenging Nvidia’s market lead. This partnership could accelerate Google’s chip ambitions, proving TPUs’ viability in large-scale deployments.

Overall, this means Meta is solidifying its position as an AI powerhouse, but at the expense of ballooning capex that could pressure margins if AI monetization lags.

Key Takeaways

Investors remain anxious about Meta’s escalating capex, viewing it as a high-stakes bet on unproven AI returns. They still have nightmarish visions about Zuckerberg’s all-in bet on the metaverse and its Reality Labs unit that continues to drain billions of dollars in resources.

The stock has tumbled nearly 20% from its all-time high of around $796 in August, reflecting broader concerns over profitability amid economic uncertainty. While incorporating Google’s TPUs advances Meta’s diversification aims and enhances infrastructure resilience, the market is still skeptical, as the stock is dipping another 2% in midday trading today on the deal reports. Ultimately, this arrangement may favor Google more than Meta, cementing TPUs as a cornerstone of the AI landscape and establishing Alphabet (NASDAQ:GOOG)(NASDAQ:GOOGL) as a formidable rival to Nvidia in custom silicon.