The Architect of the Social Era Steps Down: Inside the Leadership Transition at Meta’s Infrastructure Core
MENLO PARK, CA — In a move that signals a profound shift in the internal mechanics of the world’s largest social media empire, Meta Platforms Inc. has confirmed that Santosh Janardhan, the longtime head of infrastructure and co-head of engineering, will step down from his role. Janardhan, a foundational figure who joined the company in 2009 when it was still known as Facebook, will be succeeded by Dave Brown, a veteran of Amazon Web Services (AWS).
The transition comes at a critical juncture for Meta. As the company pivots its entire business model toward generative artificial intelligence and the "Metaverse," its underlying physical architecture—the vast network of data centers, custom silicon, and undersea cables—has become its most valuable, and most expensive, asset. Janardhan’s departure marks the end of an era defined by hyper-scale social networking and the beginning of a new chapter focused on industrial-grade AI compute.
The Transition: Passing the Baton in a High-Stakes Environment
Meta officially confirmed the leadership change following the circulation of an internal memo on November 11, 2026. According to company spokespeople, Janardhan will not vanish overnight. Instead, he will transition into an advisory role within the newly formed "Meta Compute" division, ensuring a stable handover before his final departure, which is slated for some time in 2027.
The appointment of Dave Brown as Janardhan’s successor is a calculated strategic move. Brown joined Meta earlier in 2024 as Vice President of Infrastructure, bringing with him a nearly two-decade pedigree from Amazon. During his 19-year tenure at AWS, Brown was instrumental in scaling the world’s most successful cloud infrastructure, specifically focusing on elastic compute (EC2), cloud networking, and specialized AI hardware.
By tapping an AWS veteran to lead its engineering and infrastructure, Meta is signaling that its internal operations must now mirror the efficiency and scalability of a top-tier public cloud provider. As Meta scales its Llama family of large language models (LLMs) and integrates AI across its "Family of Apps," the complexity of its hardware requirements has transcended traditional social media needs.
A Chronology of Scale: The Janardhan Era (2009–2026)
To understand the weight of this leadership change, one must look at the evolution of Meta’s infrastructure over the last 17 years. When Santosh Janardhan joined in 2009, Facebook was a burgeoning social network struggling to stay online under the weight of hundreds of millions of users.
1. The Early Years and the Move to Custom Hardware
Before Janardhan’s influence took full hold, Facebook relied largely on leased data center space and off-the-shelf servers. Under his leadership, the company made the radical decision to design its own hardware. This culminated in the 2011 launch of the Open Compute Project (OCP). By open-sourcing its hardware designs—ranging from server racks to power supplies—Meta forced the entire industry to innovate, significantly lowering the cost of hyper-scale computing.
2. Global Expansion and Connectivity
Through the 2010s, Janardhan oversaw the construction of Meta’s massive global data center footprint, which now spans dozens of sites across North America, Europe, and Asia. He was also a key proponent of Meta’s investment in subsea fiber optic cables, ensuring that the company controlled the "pipes" through which its data traveled, reducing latency for billions of users on WhatsApp, Instagram, and Facebook.
3. The Great AI Pivot (2022–Present)
The most challenging phase of Janardhan’s career began in late 2022 with the explosion of generative AI. Meta found itself in a frantic race to acquire H100 GPUs from Nvidia and reconfigure its data centers to handle the massive thermal and power demands of AI training. Earlier this year, Janardhan was tasked with leading Meta Compute, an initiative designed to treat Meta’s vast infrastructure as a strategic business unit rather than just a cost center.
Supporting Data: The Cost of Staying Competitive
The scale of the task awaiting Dave Brown is best illustrated by Meta’s recent financial disclosures. The company’s capital expenditure (Capex) has reached unprecedented levels, driven almost entirely by AI infrastructure.

- Capex Explosion: In 2024 and 2025, Meta’s annual capital expenditure hovered between $35 billion and $40 billion. Projections for 2027 suggest this figure could rise further as the company breaks ground on "next-generation" AI data centers designed specifically for liquid cooling and massive power densities.
- The GPU Moat: CEO Mark Zuckerberg previously noted that Meta’s "compute stockpile" would include the equivalent of 600,000 Nvidia H100 GPUs by the end of 2024. Managing this fleet requires a level of orchestration that Janardhan pioneered but Brown must now industrialize.
- Custom Silicon (MTIA): Meta has been increasingly vocal about its Meta Training and Inference Accelerator (MTIA). Reducing reliance on third-party silicon providers like Nvidia is a top priority for the infrastructure team to manage long-term margins.
Official Responses and Internal Sentiment
In his internal memo, Janardhan reportedly characterized his decision to step down as a natural progression, stating it was time to "pass the baton" to a new generation of leadership.
"Santosh’s contributions to this company are immeasurable," a Meta spokesperson said in a statement to the press. "He helped build the backbone of the modern internet. As we look toward a future where AI is integrated into every fiber of our products, Dave Brown’s experience in cloud-scale computing makes him the ideal leader to take us forward."
While the memo did not cite specific personal reasons for the departure, industry analysts suggest that the grueling pace of the AI race may be a factor. The transition from managing a social media backend to building a global AI engine is a "once-in-a-generation" technical hurdle that requires fresh perspectives on cloud architecture—precisely what Dave Brown provides.
Implications: Is Meta Becoming a Cloud Company?
The leadership change raises several critical questions about Meta’s long-term strategy. The appointment of an AWS veteran and the creation of the "Meta Compute" division suggest that Meta may be looking beyond its own internal needs.
1. Selling Excess Capacity
There has been persistent speculation that Meta might eventually follow in the footsteps of Amazon. Just as Amazon built AWS to monetize its internal infrastructure, Meta may seek to sell its excess AI compute to third-party developers or partners. With its massive stockpile of GPUs and its own custom silicon, Meta is uniquely positioned to offer "AI-as-a-Service," particularly for developers already using the Llama ecosystem.
2. The Llama 4 and 5 Roadmap
The success of Meta’s future AI models depends entirely on the efficiency of the infrastructure Dave Brown will now manage. Training the next generation of models (Llama 4 and beyond) will require clusters of GPUs that are larger and more tightly integrated than anything currently in existence. Brown’s background in elastic compute will be vital in ensuring these clusters can scale without massive downtime or energy inefficiency.
3. Structural Reorganization
Janardhan’s departure might be the first of several leadership shifts as Meta flattens its organization. The company has been in a "year of efficiency" mode for some time, and the consolidation of engineering and infrastructure under a cloud-centric leader like Brown suggests a move toward a more streamlined, "product-first" engineering culture.
4. Energy and Sustainability
One of the greatest challenges Brown will face is the power constraint. AI data centers consume vastly more electricity than traditional ones. Meta’s ability to hit its net-zero targets while doubling its power consumption will be a major test of the new leadership’s ingenuity in renewable energy procurement and cooling technology.
Conclusion: A New Era of Engineering at Meta
The departure of Santosh Janardhan is more than a simple executive reshuffle; it is a symbolic milestone in the history of Silicon Valley. Janardhan helped build the "Social Web," a world of likes, shares, and photos. Dave Brown is being brought in to build the "Intelligence Web," a world of autonomous agents, synthetic media, and real-time AI reasoning.
As Janardhan moves into his advisory role over the next year, the tech industry will be watching closely to see how Brown adapts the lessons of AWS to the unique, high-velocity environment of Meta. For a company that has staked its entire future on the belief that "compute is the new oil," the man running the rigs is now the most important person in the building.
