In a world gripped by the rapid evolution of AI, few investors have navigated the intersection of autonomy and innovation with the conviction and foresight of Andrew Medjuck. A backer of foundational players like OpenAI, xAI, and Anthropic, Medjuck is not merely following the future—he’s helping define it. His approach, rooted in private capital and long-term vision, offers a rare lens into how AI agents are transforming industries, disrupting workflows, and redrawing the economic map.
Beyond hype: the real traction behind AI agents
When asked what first signaled the paradigm shift of AI agents, Medjuck doesn’t cite theory or headlines. He points directly to function. From legal tech startups in the U.S. like Harvey and EvenUp to international players like Leya in Europe and Enter in Latin America, these tools aren’t just shiny new interfaces. They’re getting real work done. For Medjuck, that cross-border, cross-regulation effectiveness is more than validation—it’s confirmation that AI agents are delivering tangible, scalable outcomes.
This traction, he emphasizes, is redefining the very economics of knowledge work. AI platforms are no longer selling software—they’re selling labor output. A boutique law firm paying $200,000 to $300,000 a year for an AI platform may sound extravagant—until you compare it to the cost of hiring another associate. That delta represents a fundamental shift in the total addressable market. Software is no longer just a tool; it's a workforce multiplier.
Autonomy versus automation: calibrating the investor lens
The narrative around AI agents often leans on grand ideas of full autonomy and AGI. But Medjuck cautions against that binary thinking. Most AI agents today are not replacing humans wholesale—they're integrating into workflows to augment productivity. He points to code-generation platforms like Cursor as a rare example of true autonomy, where agents generate and deploy code with minimal human involvement.
This distinction isn’t just academic. For investors, it reshapes how risk is evaluated and how returns are forecasted. A tool that helps professionals complete a task faster sits in the realm of SaaS and workflow optimization. An autonomous agent that handles tasks continuously with minimal oversight enters the terrain of compounding value. For Medjuck, understanding where an AI product falls on that spectrum is critical to underwriting long-term potential.
Crypto, capital, and the new economic actors
Medjuck’s perspective takes a radical turn when discussing the implications of crypto-native AI agents—autonomous systems capable of holding and transferring value without human intermediaries. In his view, the traditional financial system is ill-equipped to support these digital entities. Bank accounts, KYC protocols, and compliance checks are built for people, not agents.
Enter crypto. Medjuck believes digital currencies and decentralized infrastructure will become the de facto financial rails for agent-based economies. The emergence of “AI agent millionaires” may still be anecdotal, but for Medjuck, it's the early signal of a broader reimagining of ownership and revenue models. Machine-to-machine transactions, untethered from human bottlenecks, could unlock entirely new forms of commerce.
Backing the builders before the boom
Medjuck’s approach is shaped by the freedom of operating outside traditional venture constraints. As a family office investor, he prioritizes long-term outcomes over quarterly metrics. That perspective allowed him to bet early on longevity tech, crypto, and fintech—and it’s what’s driving his current conviction in AI agents.
For him, the surest signal of a sector’s readiness is the migration of elite talent. When top-tier engineers abandon stable careers to build in a new space, something deeper is happening. That pattern repeated itself in crypto during 2016 and 2017, and Medjuck sees the same gravitational pull around AI agents today. In his playbook, timing the talent migration is timing the market.
Rethinking moats in an open-source world
With open-source models accelerating and foundational models improving overnight, defensibility in the AI agent space can feel tenuous. But Medjuck is clear-eyed about what creates real staying power. It’s not just better code—it’s layered moats: data network effects, switching costs, proprietary infrastructure. Startups that stack these advantages become not just tools, but ecosystems.
Take EvenUp. Its growing dominance in specific legal case categories allows it to better price and benchmark outcomes. That data advantage compounds with every use, making it harder for competitors to replicate the insights or match the speed. In a market where agility matters, this form of defensibility isn’t just protection—it’s propulsion.
The rise of machine-native business models
Looking at the road ahead, Medjuck sees business models evolving to match the capabilities of agents. The old SaaS model—pricing per seat or license—doesn’t fit when AI agents are completing entire workflows. Instead, he sees growing potential in models tied to output: usage-based pricing, agent-to-agent transactions, and crypto micropayments.
This is a world where machines transact with machines, autonomously and in real time. It’s not just an operational shift—it’s an economic reordering. Protocols that enable this machine-native commerce, in Medjuck’s view, represent the foundational infrastructure for the next phase of the internet.
Navigating risk while chasing upside
Of course, investing in such a fast-moving space isn’t without its existential risks. Medjuck points to the potential arrival of AGI by a major player like OpenAI, Meta, or Google as the ultimate black swan. If foundational models leapfrog entire layers of abstraction, hundreds of startups could be rendered obsolete overnight.
To hedge that risk, Medjuck takes a diversified approach—investing both in foundational model providers and in startups building defensible applications atop them. That layered exposure gives him resilience in the face of volatility, while still capturing upside.
The next decade of autonomous value creation
Looking five to ten years out, Medjuck sees a world where AI agents have rewired the flow of capital, labor, and value. Investment decisions will be made in real time by agents, with diligence and allocation executed autonomously. Work will transition from human-executed to human-orchestrated, or even become entirely machine-driven. And value will derive not just from intellectual property or labor, but from the continuous output of agent networks.
The early adopters—those building or backing the platforms that facilitate these interactions—stand to gain the most. In Medjuck’s eyes, that’s not speculative fiction. It’s a pragmatic bet on where the digital economy is already heading.
As industries brace for the full impact of AI agents, Andrew Medjuck’s insights offer a rare, grounded look at the tectonic shifts underway. His strategy—patient, precise, and bold—underscores a fundamental truth of this era: in the age of autonomous agents, the future belongs to those who act before it’s obvious.
Investing involves risk and your investment may lose value. Past performance gives no indication of future results. These statements do not constitute and cannot replace investment or financial advice.
VentureBeat newsroom and editorial staff were not involved in the creation of this content.