Every week brings a new model, a new benchmark, a larger context window, a faster inference engine, a smarter chatbot. For the last three years, the AI conversation has been dominated by one question: which model will win?

It's the wrong question.

The more important one is where the economic value of AI will actually be created. History offers a useful clue here. During the industrial revolution, the companies that generated electricity didn't capture all the value. Neither did the manufacturers of steel, nor the builders of railways. The largest fortunes were ultimately created by the businesses that used those capabilities to reinvent entire industries.

AI appears to be following the same path. To see where the value will land, it helps to stop treating AI as a single product and start treating it as what it actually is: a five layer economic stack.

Layer 1: Energy

Every AI interaction starts with electricity. Unlike traditional software, which retrieves information from storage, AI generates intelligence in real time. Every prompt, every image generated, every voice conversation, and every automated decision draws power at the exact moment it happens, which means that without electricity, there simply is no intelligence to speak of.

As adoption accelerates, energy is becoming a strategic asset in its own right. Data centers now compete for power availability just as fiercely as they compete for compute capacity, and reliable, affordable energy is shaping up to be one of the biggest strategic advantages available to nations and enterprises alike.

Layer 2: Chips and Compute Infrastructure

Energy powers the machines that create intelligence, and modern AI workloads demand specialized hardware to do it: GPUs, AI accelerators, high-speed interconnects, and advanced cooling systems capable of processing trillions of calculations in parallel.

Where previous industrial revolutions were built on steel and manufacturing, this one is being built on silicon and compute. These are the factories of the digital age.

Layer 3: Cloud Infrastructure

Raw compute is only useful once businesses can access it easily, and that's the role cloud platforms play. Cloud providers turned AI from something only governments and research institutions could afford into something any startup can consume on demand. Organizations no longer need billion-dollar data centers of their own; they need minutes, not years, to reach world-class infrastructure.

Cloud is, in effect, the bridge between hardware and intelligence.

Layer 4: Foundation Models

This is the layer most people think of when they hear the word AI: large language models, vision models, speech models, and reasoning engines that can understand language, analyze images, generate code, and hold natural conversations. They are genuinely remarkable achievements.

But models alone don't create business value. A model sitting in a lab changes nothing; a model embedded inside a business process changes everything. This is also the layer where many executives mistakenly assume the AI story ends. It doesn't.

Layer 5: Applications and Digital Workers

This is where intelligence turns into action, and where the real economic value is captured. It's the layer where AI starts answering customer calls, processing invoices, scheduling appointments, approving claims, reconciling accounts, and managing inventory. It's where hospitals cut patient waiting times, banks accelerate lending decisions, manufacturers predict failures before they happen, retailers personalize experiences in real time, and governments improve citizen services.

Every industry, from healthcare and banking to manufacturing, retail, government, and logistics, will build its own application layer, and the companies that understand their industry's problems most deeply will create the greatest value.

This pattern isn't new. During the internet era, telecoms built the networks, chipmakers built the hardware, and cloud providers built the infrastructure, yet much of the economic value was ultimately captured by the companies that built applications on top of all of it. AI is following the same script.

The Next Workforce Is Hybrid

For the first time, organizations are preparing to manage two workforces at once: a human workforce and a digital one. That digital workforce is already taking shape across the business. Some digital workers greet customers and handle service requests, effectively acting as receptionists or customer service agents. Others sit closer to revenue and operations, working as sales executives, finance analysts, or procurement specialists. In healthcare settings, they're beginning to coordinate care itself, supporting patients and clinicians alike.

These digital workers won't replace organizations. They'll become part of them, operating around the clock, absorbing repetitive and transactional work, and freeing people to focus on judgment, relationships, creativity, and leadership. The organizations that learn to orchestrate both workforces well will build a real and lasting competitive advantage.

Technology Is Becoming a Commodity. Industry Knowledge Isn't.

Foundation models are improving so fast that capabilities which looked extraordinary twelve months ago are standard features today. That means the model advantage is shrinking, and the real differentiator is shifting toward domain expertise: understanding healthcare workflows, banking regulations, manufacturing operations, supply chains, and, more broadly, how work actually gets done.

The winners of the application era won't necessarily be the companies with the largest models. They'll be the companies that understand business problems deeply enough to embed intelligence directly into their operations.

From AI Experiments to AI Economics

Most organizations are still asking how they should use AI. The better question is which business outcome they're trying to improve, whether that's reducing customer acquisition costs, improving employee productivity, accelerating collections, reducing claims leakage, improving patient outcomes, or increasing manufacturing throughput.

The conversation needs to move from technology adoption to economic impact. The organizations that make that shift early will define the next decade of their industries.

Where 10xDS Plays

We're not trying to build chips, data centers, or foundation models. Those layers already have capable players, and that's not where we think we can add the most value.

What we focus on instead is the layer above all of it: Digital Workers that plug into a business and actually get things done, working alongside existing teams rather than replacing them, absorbing the operational drag that usually comes with growth.

It's a simple enough thesis. The layers below make intelligence possible. What we're building is meant to make it useful, day to day, inside real operations. That's the shift we think matters most over the next decade: not bigger models, but businesses that figure out how to run human expertise and digital workers as one team instead of two.

Talk to our experts today!