Rob May on Why Enterprise AI Is Moving Beyond Standalone Models

The Neurometric co-founder and CEO built his latest company around a conviction that the next phase of artificial intelligence would belong to systems, not standalone models.

Rob May, co-founder and chief executive officer of Neurometric, has spent much of his career recognizing changes before the market has settled on the language to describe them.

That pattern began long before artificial intelligence became the focus of enterprise technology. May was among the early bloggers who found a way to earn money through online writing, at a time when there was no established model for doing so. He later carried that instinct into cloud software, founding Backupify as businesses began moving more of their operations online.

Backupify focused on cloud-to-cloud backup, helping companies protect data stored in software services rather than on local systems. Datto acquired the company in 2014 for roughly $90 million. The exit gave May a close view of how technology categories develop, where customers find practical value, and how infrastructure companies become embedded in everyday operations.

Moving Early Into Artificial Intelligence

After Backupify, May turned his attention to AI and co-founded Talla, an AI-enabled digital assistant. The move came well before generative AI became a standard topic in boardrooms.

Talla gave May direct experience building within the category, but his perspective also expanded through investing. As an angel investor and venture capitalist, he has invested in over 75 AI companies. He has described the scar tissue from his own founder experience as an advantage in evaluating other businesses, particularly when separating compelling technology from a company that can create lasting value.

That broad view also shaped Investing in AI, the newsletter May has written on Substack for several years. Through it, he has examined AI companies, infrastructure, investment patterns, and the economics behind the industry.

May is also developing a book on AI investing, a project that has been in progress for approximately 2.5 years. The newsletter became a place for him to test and refine the thesis that eventually led to Neurometric.

The Engine and the Car

May’s view of the AI market can be reduced to a straightforward analogy. The model is the engine, but the customer buys the car. While an engine may be technically impressive, most buyers care about the full system surrounding it. They are concerned with whether it performs the required task, integrates with existing operations, controls costs, and remains dependable.

That distinction became more important as enterprises began using multiple models for different purposes. By late 2024, May saw signs that the industry was moving toward a post-model world, where companies would compete less on access to a single model and more on how effectively they coordinated multiple models.

For enterprise buyers, that creates practical questions around inference economics. A model may produce a strong result, but the cost, latency, reliability, and vendor requirements can vary significantly depending on the task.

Building Around Fragmentation

May founded Neurometric a little over a year ago as a direct bet on model fragmentation. His thesis held that enterprises would increasingly need systems capable of selecting, combining, and managing different models rather than relying on one provider for every use case.

That premise placed model routing at the center of the company. Instead of sending every request through the same model, a routing system can match a task with an option based on factors such as performance, speed, and cost.

The broader category of AI infrastructure becomes more important as those choices multiply. Enterprises may need to account for different vendors, model capabilities, pricing structures, and operational requirements across a growing number of applications.

As enterprise AI adoption grows, organisations are increasingly evaluating interoperability, governance, security, vendor flexibility, and operational cost alongside model performance. Industry research from Gartner and McKinsey & Company suggests that AI success increasingly depends on integrating multiple technologies into core business processes rather than relying on a single foundation model. Gartner has identified AI engineering and AI governance as key priorities as organisations move from experimentation to large-scale deployment, while McKinsey reports that enterprises are placing greater emphasis on scalable deployment, governance, and measurable business value. Together, these trends suggest that the next phase of enterprise AI will depend not only on advances in model performance, but also on the systems that enable organisations to deploy, manage, and govern AI effectively.

For May, the frontier laboratories moving higher into the software stack only reinforced his earlier conclusion. Models would remain important, but many industry observers expect increasing value to come from the systems that help organisations integrate, manage, and govern multiple AI models.

From Thesis to Operating Company

Neurometric represents the latest version of a pattern that has run through May’s career. He identified cloud data protection before it became a routine concern, entered AI before the current wave of attention, and then recognized model fragmentation before it became a widely discussed enterprise problem.

The company's goal is to develop infrastructure that supports enterprise AI orchestration as organisations adopt increasingly complex multi-model environments.

For companies evaluating AI spending, the immediate concern is often enterprise AI cost optimization. Yet May’s argument goes further. Cost depends on architecture, vendor selection, routing, and how well a company matches each task with the appropriate model.

That is why Neurometric does not read as a reaction to the latest AI cycle. It follows a thesis May developed through building, investing, and writing. He saw the market moving from engines to cars, then started a company designed for the road ahead.