In the rapidly evolving landscape of artificial intelligence, the transition from experimental AI chatbots to autonomous, task-oriented agents has created a new set of challenges for enterprise IT departments. As businesses scramble to integrate AI agents into their workflows, the need for robust oversight, security, and management tools has become a critical bottleneck. Enter Mintive, an early-stage startup that is positioning itself as a foundational layer for enterprise agent management. With a fresh $7 million in seed funding and a pedigree of founders steeped in the history of deep learning, the company is attempting to bring order to the chaotic proliferation of corporate AI.
Mintive’s total funding now stands at $11 million. The latest seed round was led by Cowboy Ventures, with participation from heavy hitters including Coatue, Operator Collective, and Hustle Fund. The startup has also attracted a notable roster of angel investors, including Okta co-founder Frederic Kerrest, former Google Chief Scientist and Discovery Loop co-founder Jeff Dean, and OpenAI co-founder Andrej Karpathy. This level of backing, particularly from figures who have shaped the trajectory of modern AI and enterprise identity management, underscores the industry’s recognition that the current "wild west" of agent deployment is unsustainable for large organizations.
The company was founded by Jiquan Ngiam and Vijay Vasudevan, two veterans of the AI revolution whose backgrounds lend significant credibility to the platform’s technical architecture. Ngiam, who serves as CEO, was the founding engineer at Coursera, where he was hired by AI luminary Andrew Ng. Following his tenure at Coursera, he moved to Google Brain to work on neural network research. His co-founder, CTO Vijay Vasudevan, possesses a similarly impressive technical lineage; he is widely recognized as one of the co-creators of TensorFlow during his time at Google Brain, a platform that served as the foundational training infrastructure for the modern AI era, long before the rise of alternatives like PyTorch.
At its core, Mintive aims to solve the problem of "agent sprawl." As enterprises deploy more agents to handle data analysis, customer support, and internal operations, they face increasing risks regarding data privacy, security, and cost management. Mintive’s platform acts as an orchestration and governance layer, offering an MCP (Model Context Protocol) gateway that allows IT professionals to define exactly which cloud services an agent can access and what specific actions it is authorized to perform.
Beyond access control, the platform provides an LLM gateway that can route requests to various models while simultaneously capping costs—a critical feature for companies worried about the "black box" nature of API spending. Perhaps most importantly for the enterprise, Mintive handles agent identities and authorization tokens that are distinct from the individual employees they serve. This allows companies to audit agent activity separately from human activity, providing an "agent monitor" that tracks the performance and behavior of all bots across the entire organization. Additionally, the platform enables the creation of long-running agents capable of executing complex, multi-step tasks without constant human intervention, effectively serving as an operating system for corporate AI.
The challenge Mintive faces is one of extreme competition. The market for agent management is becoming increasingly crowded as major incumbents and nimble startups alike race to claim the space. Tech giants like Amazon, Microsoft, OpenAI, and Anthropic are all building native tools to manage agents within their own ecosystems. Meanwhile, a flurry of point-product startups—such as Langchain, New Relic, and Rippling—are carving out niches in the orchestration, observability, and infrastructure sectors. A cursory look at the Y Combinator ecosystem alone reveals over 40 startups currently targeting some aspect of agent identity or management, suggesting that the barrier to entry for the general concept is relatively low, even if the barrier to high-quality execution remains high.
When asked about the "secret sauce" that will allow Mintive to survive in such a crowded arena, Ngiam points to simplicity. The company’s philosophy is rooted in the belief that for AI to be truly useful at the enterprise level, the barrier to adoption must be lowered significantly. He argues that the complexity of current systems—the need to manage sandboxes, security harnesses, and memory allocation manually—has slowed down internal deployments. Mintive promises a "click-and-deploy" experience that masks the heavy lifting of backend infrastructure.
"I think what our customers look at and they are a bit shocked at, is how quick it is and easy to deploy," Ngiam explains. "We have really thought through all the ways this can be done, all the access control, permissioning layer, to make it so easy for them to deploy it, that all the complexities are hidden away. You don’t think about sandboxes, you don’t think about harnesses, you don’t think about memory because we manage that for you."
This pitch appears to be resonating with the market. Since the product’s launch in February, Mintive reports having onboarded 50 enterprise customers, including notable names like Ro, Modern Treasury, the University of Michigan, Coursera, Apax, and Stability AI. This early traction suggests that while many companies have experimented with individual agents, they are now hitting a wall where they need a centralized system to manage them safely and efficiently.
For Aileen Lee, managing partner at Cowboy Ventures, the decision to invest was driven as much by the founders’ operational philosophy as by their technical expertise. In a market where many AI startups are securing massive, unicorn-valuation funding rounds without a clear path to sustainable growth, Lee was impressed by the disciplined approach Ngiam and his team have taken toward their finances.
"AI is going to be able to power incredible things, but we need more infrastructure that’s going to give us governance and visibility and control," Lee noted. She expressed particular respect for the fact that Mintive has limited its total fundraising to $11 million, a stark contrast to other AI infrastructure players that have raised hundreds of millions of dollars at billion-dollar-plus valuations—such as Prime Intellect, which recently secured $130 million for its Series A.
Lee views the current trend of "vanity rounds" as a potential trap for founders, arguing that over-capitalization can lead to bloated operations and a lack of focus on the actual needs of the customer. By choosing to grow "brick by brick," in the old-fashioned way, Mintive is positioning itself to be more agile and responsive to its user base.
"I feel really lucky to have won the chance to work with him," Lee said of Ngiam. "He is a grown-up and so he was not a person chasing a ‘vanity’ round. I think a lot of companies are being set up for failure by chasing round size, rather than, ‘how much do I need to actually really move the needle for the company?’ And so, yes, in a time when people are raising unprecedented amounts of money, we’re just kind of doing it the brick by brick, old-fashioned way."
As the industry moves into the next phase of the AI cycle, the focus is clearly shifting from the raw power of large language models to the practical reality of integrating those models into legacy corporate environments. The "agentic" future, where software works alongside humans to execute business processes, requires a level of trust that only robust infrastructure can provide. By focusing on governance and ease of use, Mintive is betting that the winning companies of the next decade will not necessarily be those with the most funding, but those that provide the most reliable "plumbing" for the AI-powered enterprise. Whether their platform can maintain its simplicity as it scales to meet the diverse and complex needs of larger enterprise clients remains to be seen, but the early signals from the market suggest that the demand for this kind of foundational oversight is higher than ever.
