Each archetype represents their different path to power.

Windows crowned Microsoft, gatekeeper of the computer operating system.

Search crowned Google, gatekeeper of consumer intent.

Chrome crowned Google, gatekeeper of the internet.

Android crowned Google, gatekeeper of the mobile device operating system.

iPhone and iPad crowned Apple, the gatekeeper of the mobile device.

Social Media crowned Meta, the gatekeeper of online attention.

GPUs crowned Nvidia, the gatekeeper of compute.

And then came the cloud, the invisible infrastructure.

AWS built the scaffolding.

Azure the enterprise bridge.

Google Cloud the computational brain for AI.

Each indispensable. None crowned.

It powered everything, but was ruled by no one.

History never crowns them all.

Each seeks to crown itself the ultimate king of the artificial intelligence era.

Only one will wear the true crown of the AI age.

A new war has begun.

Is ChatGPT the default interface of artificial intelligence itself? It now serves more than 700 million weekly active users and processes over 2.5 billion prompts each day, making it the first consumer app to reach mass scale. ChatGPT has become the starting point for how hundreds of millions research, plan, write, edit, and think. OpenAI’s strategy is clear: shape and redirect consumer behavior through generative AI to control distribution, and then expand vertically into the AI supply chain. They have no operating system, no hardware, and no infrastructure, yet OpenAI sits at the center of the AI economy. Through ChatGPT, OpenAI has built what every other crowned incumbent believed was impenetrable, choke point distribution, without owning a single layer of the traditional tech stack.

OpenAI continues to expand accessibility and functionality at a rapid pace. The Apps SDK and the underlying Model Context Protocol (MCP) laid the groundwork for ChatGPT to evolve from a single application into a platform that can orchestrate external services and agentic workflows across third-party apps, strategically blurring the line between model, app, and interface. To scale usage even further, the GPT Store serves as the distribution layer for that ecosystem, enabling brands to publish their own customized agents through the GPT Builder interface inside OpenAI’s platform. With each new GPT released, OpenAI’s reach expands, turning its model into a hosted environment where distribution is driven by rapidly compounding capabilities and functionality.

OpenAI has now extended that ecosystem from cognition to commerce. On September 29, 2025, OpenAI and Stripe announced the co-development of the Agentic Commerce Protocol (ACP), an open standard that enables AI agents to complete transactions seamlessly and securely inside ChatGPT. ACP lets users move directly from chat to checkout, allowing ChatGPT to facilitate payments, execute transactions, and interact with merchants directly through conversation itself. Less than a month later, on October 28, 2025, PayPal joined the rollout, introducing an additional payment option and expanding transaction coverage within ChatGPT. What began as an interface for reasoning is becoming an economic platform for action, one that fuses artificial intelligence with intent and execution.

Sora is the bridge between language and light. Released on September 30th, 2025 and powered by the Sora 2 model, a fusion of text, image, and video generation into one creative app. It can generate photorealistic video, simulate movement, and recreate entire physical scenes in seconds from a single prompt. Sora isn’t just generative, it’s generational. Within days, it skyrocketed to the top of Apple’s Free Apps chart, cementing OpenAI’s foothold in generative media and marking the moment multimodal creation became mainstream. It gives OpenAI the power to render imagination itself, eliminating the boundary between cognition and perception. If ChatGPT made OpenAI the interface for reasoning, Sora makes it the interface for creation. What began as words and code now extends into motion, color, and light, allowing OpenAI to shape not only how creativity is expressed, but how it is experienced.

Atlas completes the trilogy. With the release of ChatGPT Atlas on October 21, 2025, OpenAI extends its dominance into navigation itself, transforming the browser into the next interface of artificial intelligence. Available today on macOS, with Windows, iOS, and Android versions on the way, Atlas brings AI to the internet’s front door. Together ChatGPT, Sora 2, and Atlas form the foundation of OpenAI’s new AI stack of cognition, synthesizing intelligence, functionality, commerce, creativity, and navigation. In the short term, OpenAI is focused on broadening access and distribution to meet consumers at their preferred points of interaction.

But over the long term, artificial intelligence isn’t just about creation, it’s about controlling distribution and access. Desktop clients for macOS and Windows elevate ChatGPT from a website to a cognitive layer that integrates one level deeper than every other application. Through its partnership with Apple, OpenAI gains something only Google has had, default access to the world’s premium devices. When Siri or Apple Intelligence hands off a query to ChatGPT, OpenAI captures distribution without owning the operating system. That reach expanded on October 23, 2025, when OpenAI acquired Software Applications, Inc., the maker of Sky, an AI-powered natural-language interface for macOS. The acquisition gives OpenAI native presence within Apple’s ecosystem, marking its first tangible step toward an AI-native operating layer that merges interface and artificial intelligence.

The Sky acquisition signals a deeper intent. Sky isn’t just an interface, it’s a foothold in the operating layer. By acquiring Sky, OpenAI now has a potential entry point to control its own distribution through its own operating system, pushing them one layer deeper into the device stack. Marking the beginning of an experiment: what happens when the interface itself becomes the OS. By embedding a multimodal assistant directly into the device layer, OpenAI gains access far more durable than an app embedded or integrated into a partner’s ecosystem. This is the first visible step toward what could become an AI-native operating system, that controls every task, query, and workflow through generative AI rather than software hierarchies. OpenAI doesn’t need to build its own device to control distribution, Sky is its gateway for the interface to become the operating system.

OpenAI’s reach, however, comes with dependence. OpenAI doesn’t own the cloud infrastructure or software stack that powers its compute. Its models are hosted, trained, and served primarily through Microsoft’s Azure cloud. From 2019 through August 2025, Azure was essentially OpenAI’s exclusive cloud provider, effectively supplying its entire infrastructure. Under the original terms of their partnership, Microsoft didn’t hold equity in OpenAI’s for-profit arm. Instead, it participated through a four-phase profit-sharing structure, receiving up to 49% of OpenAI’s profits until a predetermined capped return was reached. The agreement gave Microsoft a powerful economic interest and strategic influence, but not ownership or governance rights, positioning it as a preferred capital partner rather than a shareholder. On September 11th, 2025, the companies announced a non-binding restructuring to rebalance ownership, revenue commitments, model access, and cloud exclusivity, opening the door for OpenAI to diversify its infrastructure beyond Azure.

And that is exactly what OpenAI is doing. The company moved quickly to unwind its reliance on Azure, both to preempt potential future DOJ antitrust scrutiny and to build an independent supply chain for compute. In less than two months, OpenAI entered a co-design partnership with Broadcom to develop its first in-house AI processors and networking systems. It also signed a multi-year agreement with AMD to secure up to 6 gigawatts of GPU capacity, a deal expected to generate up to $100 billion in revenue for AMD through 2030 while granting OpenAI warrants for a 10% equity stake. OpenAI also signed two major agreements with Oracle: a $300 billion, five-year cloud services contract for independent data center capacity and a multi-billion dollar joint venture with Oracle and Related Digital to build a Michigan data center with over one gigawatt of computing capacity as part of the larger Stargate initiative. OpenAI further expanded its CoreWeave contract by adding $6.5 billion in new GPU commitments, and deepened its relationship with Nvidia through a proposed $100 billion strategic investment and the deployment of 10 gigawatts of Nvidia systems across its next-generation data center network. This rapid evolution, from renting compute to architecting it, from distributing artificial intelligence to manufacturing the processors and capacity that sustain AI, marks one of the fastest and most consequential vertical expansions in the history of technology.

During this period, OpenAI and Microsoft began redefining the terms of their strategic alliance, shifting from dependency to interdependence. That realignment was formalized on October 28, 2025, when a new agreement was finalized. The revised terms granted Microsoft an estimated 27% equity stake, worth ~$135 billion, in OpenAI Global LLC, valuing the capped-profit entity at nearly $500 billion. This updated agreement also committed OpenAI to $250 billion in Azure compute purchases and formally ended Azure’s exclusivity as its cloud provider. The result was a partnership no longer defined by control, but by mutual leverage, with a new balance of power between a frontier AI model and infrastructure. A true masterclass in how strategic partnerships should work, even during a period of massive realignment, competitive tension, and structural change across the entire AI ecosystem.

OpenAI is the gravitational disruptor of the AI industry, bending every layer through partnership, not infrastructure. Its leverage lies in its mass adoption of synthetic cognition, consumer interfaces that eventually will abstract away browsers, websites, and apps, making an AI native OS chatbot the default layer of human-technology interaction. OpenAI seeks to control the interface of thought itself, replacing the friction of human touch points with seamless cognition. OpenAI isn’t fighting for screen time, it’s replacing it, collapsing the distance between intent and outcome, and turning conversation itself into the new access and distribution channel of the digital world.

Directing demand.

Redefining distribution.

Dictating the tempo of an industry built to serve it.

Every empire in technology is built on ecosystem control and reinforced through a competitive advantage and distribution. The AI Model Distribution Wars are not just about who builds the first AGI model, they’re about who integrates AI deepest into the government, enterprise, and consumer layers of the digital ecosystem. Every company is racing to fuse models into the access points that define daily digital life: the cloud, devices, operating systems, browsers, websites, feeds, and apps. Whoever owns the default surfaces of human and device interaction controls the new flow of cognition.

In this era, distribution isn’t about where attention goes, it’s about where reasoning happens. The interface will shift from the app-centric experiences to the operating system, from navigation to delegation, and from searching to thinking. The kings of the last generation’s built empires by capturing digital attention and transactions. This next generation will build them by capturing decision-making itself, both human and synthetic.

OpenAI is the market driver of the soon to be collapsing attention layer of the digital ecosystem. Buckle up, its about to be rewired from fragmented attention to centralized cognition.

Every company is converging toward the same goal: owning the first true AGI model, integrating the deepest vertically in the AI supply chain, scaling and optimizing their cloud infrastructure the most effectively, and building the broadest built-in device distribution. That combination of AGI model, cloud infrastructure, supply chain, and device distribution, defines the new crown of power in the next phase of the AI economy.

This is the great inversion of power: infrastructure was built, connection was wired, information was extracted, attention was directed, and cognition will be orchestrated. And the companies that master that orchestration, from the model to the cloud to the device, will control not just what we see, but how we think, act, and decide.

Next comes the orchestration of movement itself: vehicles, robots, planes, and ships, the physical extensions of AI that will translate cognition into control in the real world. And in time, even those systems will evolve from being orchestrated to being autonomous.

Welcome to the age of AI factories, where machines no longer follow instructions, they execute their own intelligence.

This was Edition #3 of Techonomics.

Stay tuned, we’re just getting started.

— Craig Sweeney-Draude

Founder, Techonomics

👉 Click Here to Subscribe