Race to Scale: AI-Driven Transformation

EdgeConneX Chief Transformation Officer Jenny Zhan explains the organizational transformation required to deliver data center capacity at ever-rising scale and ever-accelerating pace.

AI has dramatically changed the scale of data center capacity we’re delivering, and the pace we’re delivering at.

Ten years ago, a 30-megawatt (MW) data center was large. Today, 200-MW facilities are common, and every hyperscaler has plans for gigawatt-plus campuses. And just as the scale of development has grown, the pace of development has dramatically accelerated as well, with the speed of delivery today in some cases half the traditional timeframes.

Scale is driven by increasingly powerful AI models that require increasingly large clusters of connected graphics processing units (GPUs) for training, which quickly advanced Nvidia to four trillion valuation as the most valuable company. Just when the market deemed the traditional chip makers such as AMD, ARMS and others are far left in the dust, all of them came roaring back in the recent Q2 2026 earnings, with Intel leading the pack as shown in almost 500% of incredible stock growth.

Three years since ChatGPT’s emergence, the AI led development has quickly moved from training to inference, which led to massive CPU resurgence. This is Jarvon’s Paradox at full display – increased AI use led to ever faster agentic AI adoption, leading to unimaginable token consumption and the resulting need for power and compute. The unsatiable demand also led to divergent use and development of alternatives to GPU – Amazon’s Trainium, Google’s TPU, Meta’s MTIA300, Microsoft ‘s Maia, and Apple/Broadcom’s Baltra, not to mention increasing scaling of startup in inference chips such as the recent high-profile IPO of Cerebras. None of these have gained nearly as close a scale as Nvidia, but the market is taking note of the rapid development.

AI interconnect technology is rapidly shifting too, with synthetic silicon photonics increasing applied to supplement copper, as AI clusters scale from tens of thousands to hundreds of thousands of accelerators. Such trends will impact power consumption and requirements at data centers. Technology such as co-packaged optics became more prevalently produced and applied.

Meta’s Llama 3.1, for example, has 405 billion parameters and was trained using a cluster of 16,000 GPUs; its next-generation model, Llama 4, has 2 trillion parameters and is trained on a cluster of over 100,000 GPUs. On average over the last decade, AI models’ computational power has doubled every five months.

For data center providers, continuing to thrive at this new scale and pace requires transformation.

AI is here to continue explosive growth, and the world is racing for power and compute. Delivering data center capacity at an ever-rising scale and ever-accelerating pace while vigilant for technology changes requires organizational transformation. McKinsey articulated it well: “Capturing the scale-up opportunity will require data center players across the value chain to adopt new approaches and technologies while learning from other industries that have experienced similar breakthrough moments.”

EdgeConneX is creating a stronger, more agile organization with transformation built around three interdependent pillars: People, Process, and Performance.

People are the heart of the company, shaping its culture, vision, and innovation; when we focus on collaboration and cultivating positive behaviors, we create the momentum to drive meaningful progress and impact across the board. Processes enabled by technology are the backbone, providing the structure and consistency we need to grow effectively; and freeing enterprise time to focus on business intelligence and analytics. Performance is the ultimate measure of success, ensuring we can timely course correctly, and are effective to meet and exceed the expectations of our customers, our employees, our shareholders and all other stakeholders.

There are three key ingredients for transforming the organization to win the race to scale.

Agility

One of the primary goals of our transformation is to place agility in our company’s DNA, equipping people with the confidence and clarity to operate in uncharted territory. Agility enables us to adapt to support our diversifying customers’ requirements as they change. It’s the ability to align all internal organizations to react quickly; it’s also an iterative feedback mechanism that enables us to course correct quickly, to recalibrate as changes come. This iterative loop drives operational effectiveness.

Thriving in a world of shifting challenges requires ingenuity, but in many organizations, human ingenuity is constrained by clutter – data distractions and inefficient processes. Clearing the clutter unleashes our people, enabling their collective intellectual power and ingenuity to uncover insights and create value. It requires understanding what the true hurdles are that impede forward progress, and what are the behaviors, however small, that need to be replicated and scaled across the organization.

Mental flexibility is one aspect of agility. At EdgeConneX, we have a lot of smart people with very clear views of our vision from different angles. Transformation involves aligning our culture to allow people to express divergent views and then productively reconcile those differences. That can be quite difficult, compounded by a very diverse set of backgrounds, views, and working styles, and sometimes emotions. Empathy and the ability to communicate with a sense of humility are essential qualities for our leaders.

In this AI-enabled data center landscape, agility needs to be a core principle of data center design. A modern data center should be designed as a flexible platform, capable of supporting everything from lower-density cloud workloads to high-density AI and HPC environments. That flexibility extends to cooling, with options ranging from air to liquid, hybrid, and immersion as needs evolve. With constant advances in AI models, architectures, and chip innovation, the ability to adapt quickly more than an advantage, it’s a critical differentiator.

Collaboration

Collaborative innovation is essential to scaling at the speed required to support our customers. This is collaboration with each other, as well as with our customers and our vendors. For example, we collaborate with customers to understand their capacity forecasts, design requirements, and required delivery schedules, to ensure we are ready for service on time and on budget. We collaborate with chipmakers to understand their roadmaps and how the data center needs to support new technologies. And we collaborate with our supply chain partners to ensure access to materials, power, and talent to erect and run our sites. The path to success, especially in today’s world, is never linear, especially with the plethora of workstreams, SMEs, and other stakeholders involved in the complete end-to-end delivery of our integrated solutions. Any nuanced perspectives conducive to synchronized collaboration on a shifting landscape is essential.

Central to this collaboration is data, which threads together People, Process, and Performance. Ultimately, information flow and full visibility drive execution, and another reason why ‘decluttering’ to remove data distractions is so important.

Standardization

So much has changed about how data centers are built and operated; The need for speed is never as more prominent in today’s world which makes it imperative to holistically evaluate our processes, determine what still works and what needs to be changed, and standardize new processes.

Across the industry, data center operators are advancing standardization efforts to scale more efficiently, digitizing workflows, centralizing documentation, and investing in structured training programs. These initiatives strengthen operational consistency (Process), enable more effective onboarding and cross-training (People), and ultimately improve execution at scale (Performance). Modern training programs are increasingly role-based, creating a common language and shared standards across teams. This approach supports faster ramp-up, greater workforce flexibility, and continuous professional development—critical capabilities as demand for talent accelerates alongside the rapid expansion of digital infrastructure.

With agility, collaboration, and standardization, we’re transforming the organization to win the race to scale

Such DNA creates ‘anti-fragility’, quoting a property coined by Nassim Taleb. It allows the organization to thrive, improve, and increase capability when exposed to stressors, volatility, and disorders, attributes that can be well manifested in the ever-morphing AI led market developments.

I feel so lucky to be at the forefront experiencing AI as it transforms the world. We’ve been through technology shifts before, of course, but AI is clearly fundamentally different in how it is transforming society. As a data center provider, we’re right in the center of supporting social evolution. It’s hard to imagine a more exciting position to be in.

What makes this work particularly rewarding is knowing that we’re leading the industry into uncharted territory. We’re developing expertise and systems that no one has seen before, and that’s both exciting and challenging.

 _____________________

Jenny Zhan is Chief Transformation Officer for EdgeConneX, a global data center provider focused on driving innovation. Contact EdgeConneX to learn more about their 100% customer-defined data center and infrastructure solutions.

If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!