AI's S-Curve: Mapping the Infrastructure for the Next Work Paradigm

AI is no longer a futuristic concept; it is a present-day work tool, and its adoption is following a classic S-curve. The inflection point is here. According to recent analysis, at least half of workers are now using AI in their jobs, often without formal employer approval. This grassroots, bottom-up adoption-where people use it at home, at work, and across all demographics-shows the technology has "caught on incredibly quickly." The rapid spread from niche experiment to everyday utility is the hallmark of a technology entering its steep growth phase.

Yet this early phase is defined by a stark unevenness. While adoption is broad among individuals, it is highly concentrated among firms. Data reveals that half of the top 1% of firms with the most job postings have adopted AI, while the vast majority of smaller companies have not. This creates a bifurcated landscape: a small cohort of large, tech-forward employers is already integrating AI into their hiring and operations, while the broader economy lags. The economic impact of AI depends critically on whether this initial, elite adoption can cascade down to the middle and long tail of businesses. If it remains confined to a few giants, its aggregate effects will be limited and uneven.

This structural tension defines the next investment frontier. The opportunity is not in the companies merely using AI as a peripheral tool, but in those building the fundamental rails that will enable its widespread, effective integration. This is the infrastructure layer-the platforms, systems, and tools that transform AI from a fragmented, risky experiment into core, reliable infrastructure. The current setup, where employees raised data security concerns and many lack proper training, highlights the critical need for robust, user-friendly, and secure foundational systems.

. The companies that succeed will be those that solve the practical problems of scaling AI safely and efficiently, turning it from a force multiplier for a few into a standard operating system for work. The paradigm shift is underway, and the infrastructure for the next work paradigm is being built now.

The Paradigm Shift: Collaboration vs. Automation

The economic story of AI is not a simple tale of machines replacing people. It is a more complex narrative of reshaping work, where the outcome hinges on a critical distinction: whether AI tools are designed to automate or to collaborate. This is the core mechanism driving the next phase of the adoption S-curve. As economist David Autor argues, an automation tool is something that eliminates expertise, while a collaboration tool acts as a force multiplier for it. The current trajectory leans heavily toward the former, but the most promising economic path lies in the latter.

This distinction explains the counterintuitive wage patterns we are already seeing. When automation targets the simpler, more routine aspects of a job, it can paradoxically increase wages for the remaining workers. The evidence is clear: bookkeepers and accounting clerks saw computers take over many of their tasks between 1980 and 2018. Even as total employment in these roles fell by a third, their real hourly wages rose by nearly 40%. The reason is that automation concentrated the work around harder-to-replace, expert tasks, making those who remained more valuable. This is the paradigm shift in action: AI isn't just displacing labor; it's redefining the skill premium. When automation instead targets the specialized tasks, however, the opposite happens. As seen with inventory clerks, where employment more than doubled but wages fell, the job becomes easier to enter, increasing competition and driving pay down.

The result is a long, uneven transition. This is not a sudden wave of mass unemployment, but a continuous process of job reshaping. As Autor notes, the benefits of technological change have not always been widely shared, and the same applies to AI. The labor market is being split, with demand surging for new skills and tasks that complement AI. Our analysis shows that one in 10 job postings in advanced economies now require at least one new skill, with IT and healthcare leading the charge. This creates a dual pressure: workers must continuously upgrade, while firms must adapt their hiring and training to build teams that can effectively collaborate with AI.

The investment implication is straightforward. The companies that will capture value are not those building the next generation of automation tools that eliminate expertise. They are the ones developing the collaboration platforms, the interfaces, and the systems that amplify human judgment and domain-specific knowledge. In this new paradigm, the currency is not just data or compute power, but the ability to integrate AI as a true partner in complex, creative, and strategic work. The infrastructure for the next work paradigm is being built for collaboration, not replacement.

The Infrastructure Layer: Where Capital Flows

The capital markets are already moving. As AI adoption accelerates, the focus of investment is shifting decisively from the finished AI products themselves to the foundational infrastructure that makes them usable, safe, and collaborative. This is the logical next step on the S-curve: once a technology is proven, capital flows to the systems that enable its mass integration. The evidence shows a clear bifurcation. While firms race to explore AI for efficiency, employees raised data security concerns and many lack proper training. This gap between ambition and readiness is the precise opening for infrastructure builders.

The specific types of infrastructure being funded are the 'force multipliers' for human expertise, as economist David Autor defines them. This includes platforms for seamless AI integration into existing workflows, tools for rapid upskilling to meet the new skill demands, and systems that manage the complex collaboration between humans and AI. The financial implication is a new wave of venture capital and corporate investment targeting companies that solve these practical problems. The market is signaling that the most valuable assets are not the AI models, but the interfaces, guardrails, and training systems that turn them into reliable tools for the workforce.

Policy is now actively accelerating this build-out. The U.S. AI Action Plan, unveiled in July 2025, is a major catalyst. It directs federal agencies to review and eliminate regulations that could impede AI development and favors states with lighter regulatory touch for federal funding. This creates a powerful incentive for companies to build their infrastructure in supportive environments. The plan also expresses a strong preference for open-source AI, which can reduce costs and speed up development for infrastructure firms. In essence, the federal government is cutting red tape to ignite a new wave of public and private investment in the rails of the new paradigm.

The bottom line is that the infrastructure layer is becoming the new frontier for exponential growth. Companies that provide the platforms, tools, and systems to manage the human-AI collaboration will capture value as the adoption curve steepens. This is where capital is flowing because it is where the real bottlenecks to scaling AI are being solved. The paradigm shift is not just about what AI does, but about how it is built and deployed. The infrastructure for the next work paradigm is being funded now.

Catalysts and Risks: The Path to Exponential Adoption

The path from today's uneven adoption to a truly transformative, economy-wide AI paradigm is defined by a clear set of catalysts and guardrails. The primary driver is the widening chasm in readiness between large and small firms. While AI use among U.S. employees grew to 45% in Q3 2025, the underlying corporate adoption is starkly concentrated. Data shows that half of the top 1% of firms with the most job postings have adopted AI, while the vast majority of smaller companies have not. This creates a massive, underserved market for enabling infrastructure. The companies that build the platforms, tools, and training systems to bridge this gap-from secure, user-friendly integration to rapid upskilling-will capture value as the S-curve steepens.

Yet this concentration also poses the central risk: the uneven diffusion of productivity gains. If AI's benefits remain confined to a small cohort of elite employers, its aggregate economic impact will be limited and potentially exacerbate inequality. The risk is not just economic; it's political and social. As employees raised data security concerns and many lack proper training, the current setup is fragile. Without a concerted push to democratize access and capability, the technology could entrench existing market power rather than create broad-based prosperity.

Policy decisions will be the critical lever that either accelerates or constrains this diffusion. The U.S. AI Action Plan, unveiled in July 2025, is a major catalyst that favors rapid deployment. By directing agencies to eliminate regulations that could impede AI development and favoring states with lighter regulatory touch for federal funding, it creates a powerful incentive to build infrastructure quickly. However, the plan's focus on open-source AI and its stance on export controls will also shape the cost and accessibility of the foundational compute layer. Watch for how these policies evolve, as they will directly influence the pace and inclusivity of the adoption S-curve.

The investment thesis is clear. The exponential growth opportunity lies in the infrastructure that solves the practical problems of scaling AI safely and efficiently. This means investing in the companies that provide the rails for collaboration, not just the next automation tool. The catalyst is the readiness gap; the risk is the concentration of benefits; the policy environment will determine the speed and fairness of the climb up the S-curve.