Nvidia loses two senior executives to Era, an investment firm raising a $250M fund

On March 12, 2026, it emerged that Nvidia had lost two prominent managers: Jeffrey Levin and Max Cohen are moving to Era, a New York investment firm betting on AI. For the market, this is no routine personnel move. It signals that the contest over artificial intelligence is shifting further from the models themselves toward capital, clouds and infrastructure.

Who left Nvidia

At Nvidia, Levin was responsible for global AI initiatives, while Cohen ran the company's cloud business in North America. Both worked at the intersection of compute demand, enterprise deployments and partnerships. They saw the AI boom not only from the chip side but also from the side of customers building services and data centers on those chips.

Their roles at Era are already fairly concrete: Levin is helping launch a new AI startup backed by the firm, and Cohen will look for deals among new cloud providers. Here, new cloud providers means companies that rent out computing infrastructure for training and running models without being classic hyperscalers such as Amazon, Microsoft or Google. This is the layer of the market where many new bets are emerging: demand for GPUs is enormous, the entry ticket to infrastructure keeps rising, and access to the right customers and contracts is becoming as valuable as the technology itself.

What Era is building

Era is a young New York firm, only four years old. It is led by 30-year-old investor Jasper Lau, and its backers reportedly include Taiwanese businessman Andre Ku Sr. The company operates not just as a venture investor: it also provides financial services to its clients and portfolio companies. Era already holds stakes in notable AI and defense-tech assets, which shows the firm's ambition to go beyond a typical family office.

• Is raising a $250M fund
• Has already received $140M in commitments, and the round is described as oversubscribed
• Portfolio includes Anduril Industries
• Its bets include Mira Murati's Thinking Machines Lab
• Focus: AI startups and related infrastructure

For Era, this is a convenient moment to scale. The AI market remains overheated, but money is increasingly looking not for abstract "next OpenAIs" but for teams and projects with access to compute, energy and enterprise customers. Against that backdrop, hiring people from Nvidia looks logical. They bring not just impressive résumés but a practical understanding of how demand for GPUs, cloud capacity and applied AI services works today.

Why infrastructure matters

The most interesting detail in the Era story is where its capital comes from. The firm is raising money not from traditional institutional LPs such as pension funds, but from very wealthy families, some of them tied to companies building global data-center infrastructure. According to the company, these investors control businesses with roughly 70 gigawatts of energy capacity and about 70 million acres of land. For AI, these are no longer just impressive numbers: without electricity, sites and permits, it is impossible to quickly deploy new campuses for model training and inference.

Era also plans to keep working closely with Nvidia. Earlier, the chipmaker helped the firm identify priority AI markets and investment strategies for the new fund. If that model holds, Era gets a rare combination of assets: money from families with real infrastructure resources, deal flow from the AI ecosystem, and executives who understand where the next wave of demand is forming. In effect, it is an attempt to combine venture capital, strategic consulting and infrastructure capital in one loop.

What it means

The move of two Nvidia managers to Era shows how quickly the AI market is becoming a game not only of technology but of control over supply, clouds, land and energy. The winners will not necessarily be those with the best model layer, but those who can bring together chips, financing and infrastructure. For investors, it is one more sign that the next phase of the AI boom will be far more capital-intensive.