CoreWeave CEO talks $12 billion OpenAI contract and AI-focused venture fund

The company, recognized as the AI company of the year, has launched a new venture fund to foster new AI companies. This initiative allows them to deepen their engagement with the AI ecosystem, providing crucial compute resources and support for startups to build, scale, and run experiments. This reciprocal relationship creates a feedback loop, guiding the company's resource allocation and product development.

The venture fund and the broader ecosystem are experiencing overwhelming demand for compute power, driven by AI labs, hyperscalers, enterprises, and governments. Despite the challenges and the inevitable rate of failure and success, there's a commitment to investing time and resources to support these companies. The company is actively building and delivering GPUs as quickly and as largely as possible to meet this demand.

Looking ahead, the speaker believes that while markets will eventually bring supply and demand for compute into alignment, the current global buildout of AI computing infrastructure is a monumental undertaking poised to define the economy and society for the next century. This "super cycle" is expected to last a long time, with periodic signals of sufficiency not diminishing the vast scale of what needs to be built.

• A new venture fund has been launched to support and invest in new AI companies, providing them access to compute resources.
• The demand for compute power is overwhelming the industry's capacity, with a specific mention of an $11.2 billion contract with OpenAI.
• The speaker believes that while markets will eventually balance supply and demand for compute, the current global AI infrastructure buildout is a massive, multi-generational project.
• Learnings from a visionary CEO include the importance of zooming out to understand the broader technological landscape and the many moving pieces required to scale a business.
• The current AI infrastructure buildout is described as a "super cycle" that will last a long time and impact various sectors like medicine and energy.

Related Video Summary

• Beat the 95%: Why AI Projects Fail—And How Builders Win
• Cohere Founder, Nick Frosst: How To Compete with OpenAI & Anthropic, and Sam Altman’s AI Disservice
• Aaron Levie and Steven Sinofsky on the AI-Worker Future
• Why Most AI Startups Are BAD Businesses

Launch of a New Venture Fund [00:26]

• The venture fund makes sense for the company as it allows them to get to know new AI companies being formed.
• It helps them get deeper into the ecosystem and provides new companies with access to compute, a fundamental building block for future technology.
• The fund enables investment in these companies and helps them build, scale, and run experiments.
• This initiative is considered an important part of building a robust and resilient ecosystem where diverse companies can develop and deliver solutions.
• The model provides an incredible feedback loop, informing the company about where to dedicate resources and what to build.

The venture fund makes a ton of sense for us, right? Because um it it allows us to help uh get to know new uh AI companies uh that are being formed. It gets us uh into uh the ecosystem deeper.

This section details the strategic rationale and benefits of the company's new venture fund, emphasizing its role in fostering AI innovation and strengthening the overall ecosystem.

Valuations and Market Dynamics [02:13]

• Valuations are challenging in a rapidly changing technological landscape.
• The focus should be on the potential impact these companies will have on the economy and society.
• The process involves small companies investing in ideas, building them, and delivering infrastructure to change the world.
• There's an acknowledgment of inherent failure and success rates in this process.
• The company is comfortable investing time, resources, and compute to drive these ventures forward.

You know um we know there's going to be a failure rate. We know that there's going to be a success rate and we're very very comfortable uh you know, investing time, resources, compute to to to driving them forward.

This part of the discussion addresses the complexities of startup valuations in the AI space, shifting the perspective to the long-term societal and economic contributions of these emerging companies.

Massive Compute Demand and Contracts [03:17]

• There was a conversation with the OpenAI CFO who stated their sales will 3x this year to $14 billion.
• The company has an $11.2 billion contract with OpenAI announced earlier in the year, described as a massive scaled contract for a large facility.
• The demand for compute in the AI space is overwhelming the industry's capacity across the board – for AI labs, hyperscalers, enterprises, and sovereign entities.
• The company is building as fast and as large as possible, delivering as many GPUs as they can.
• This high demand creates pressure from both customers and suppliers, necessitating efforts to deliver infrastructure to companies that can realize AI's potential.

The the the space is the depth of the demand, the scale of the demand, the breadth of the demand is overwhelming the industry's capacity to deliver the compute that is required by open AI, by the hyperscalers, by the enterprise, by the sovereigns.

This section highlights the immense demand for computing power in the AI sector, underscored by a significant contract and the company's efforts to meet this overwhelming need.

The Future of Compute Availability [04:56]

• The speaker, as a market operator, believes markets are good at solving for efficiency and bringing supply and demand into alignment.
• It is expected that at some point, there will be enough compute.
• The timing and exact nature of this equilibrium are subjects of discussion.
• The current situation is viewed as a global buildout of AI computing infrastructure that will define the economy and society for the next hundred years.
• The scale of what needs to be done globally to support AI's potential benefits across various sectors (medicine, energy, economy) is astounding.

Oh, so you know I'm a market operator. I've been a market operator my whole career right like markets are very good at solving for efficiency uh bringing supply and demand into alignment and so yes of course at some point there will be enough compute.

This segment explores the long-term outlook on compute availability, framing the current era as a critical, foundational period for AI infrastructure development.

Learning from Visionary Leadership [06:57]

• The speaker has had the opportunity to work alongside and observe a visionary CEO.
• This CEO has been a tremendous supporter of the company.
• Beyond just building great technology, an interesting aspect is the broad perspective the CEO uses to understand the technological landscape and business needs.
• It is crucial for a CEO to be able to "zoom out" and see the world from a high-level perspective, given the numerous moving pieces involved in running a large-scale business.
• The CEO's capacity for this comprehensive understanding is described as unmatched.

And um I I think it's really important as a CEO um to be able to zoom out and see the world that way because there are so many moving pieces that are required to drive a business of the size and scale that he runs.

This final topic focuses on the leadership qualities and insights gained from observing a visionary CEO, particularly the importance of a broad, strategic perspective in managing a large and complex business.