Four CEOs on the Future of AI: CoreWeave, Perplexity, Mistral, and IREN

The video comprises a series of high-level interviews conducted by Jason Calacanis at the Nvidia GTC conference. The conversation transitions from the physical layer of AI—the massive data centers, GPUs, and energy requirements discussed by the CEOs of CoreWeave and IREN—to the software and application layers represented by Perplexity and Mistral AI.

The overarching narrative explores how the "compute" industry is evolving from a speculative crypto-mining foundation into a mission-critical infrastructure for global intelligence. While the first half of the program focuses on the "hardware" side—addressing the longevity of GPUs, the innovation of GPU-based financing, and the pressing need for power—the second half looks at the "intelligence" side, focusing on how AI is becoming an autonomous agent capable of executing complex, multi-step objectives. A recurring theme throughout is the tension between the potential for mass job displacement and the new era of "solopreneurship" enabled by low-cost, high-capability AI tools.

• AI Infrastructure & GPUs: The transition from crypto-mining to high-performance AI compute 1:01, the lifespan and depreciation of GPUs 10:30, and the shift from general-purpose to specialized models 1:09:51.
• The Energy Bottleneck: The critical importance of power availability, the use of renewable energy in data centers 1:26:27, and the rise of small modular nuclear reactors 1:32:45.
• Agentic Computing: The evolution of AI from search engines to "computers" that can execute tasks via browser and system access 35:50; the concept of "AI as an Operating System" 42:11.
• Financing the AI Boom: CoreWeave's "Box" method of using client contracts to secure massive infrastructure debt 19:27.
• Economic Implications: The Jevons Paradox in compute demand 1:29:56 and the impact of AI on the future of labor and entrepreneurship 1:00:22.

• Jason Calacanis: Host of the All-In Podcast.
• Michael Intrade (CoreWeave): CEO focusing on GPU infrastructure and "neocloud" services 0:41.
• Arvin Shri Nas (Perplexity): CEO of the AI-native search and agentic computing company 32:58.
• Arthur Mensch (Mistral AI): CEO of the leading European open-source model provider 1:07:26.
• Daniel Roberts (IREN): Co-CEO and Co-founder of IREN, a large-scale data center and Bitcoin mining company 1:19:02.

• CoreWeave's "The Box": A unique financing mechanism where the company uses client contracts (e.g., with Microsoft) and GPU collateral to create a "box" that governs cash flow, allowing them to raise $35 billion in 18 months 19:27.
• Perplexity Personal Computer: An upcoming product concept that synchronizes Perplexity's agentic capabilities with local hardware (like a Mac Mini) to provide a private, hybrid computing experience 37:36.
• Mistral's "Forge": A product designed to allow enterprises to specialize and customize open-source models for specific verticals like finance or engineering 1:08:08.
• The Power Shift: The transition in the data center industry from focusing on the number of GPUs to focusing on the availability of power and "time to compute" 1:21:52.

• The Move to Specialization: Mistral AI focuses on specialized models because enterprises need models that respect data sovereignty and can be fine-tuned for specific industry tasks 1:10:33.
• The Move to Agents: Perplexity is moving toward agentic browsing and "computer" capabilities because the next frontier is AI that can perform actions, not just provide information 35:50.
• The Need for Scale: CoreWeave's strategy is to become enormous to drive down the cost of capital and gain a wider flow of information from all parts of the market 23:20.

• Michael Intrade (CoreWeable): Describes the company's evolution from crypto-mining to being a "tip of the spear" for Nvidia's new architectures 9:28. He emphasizes the importance of risk management and the long-term value of GPUs, refuting claims of rapid obsolescence 10:30.
• Arvin Shri Nas (Perplexity): Outlines the roadmap from AI search to "AI as the Operating System" 42:11. He highlights the importance of "accuracy" as the foundation of trust and discusses the technical advantages of being a "model orchestrator" rather than a model creator 48:09.
• Arthur Mensch (Mistral AI): Discusses the strategic importance of open-source models and the necessity of "deterministic gates" for enterprise use 1:15:30. He explains how Mistral provides a "portable platform" to ensure data stays with the customer, addressing privacy concerns 1:11:59.
• Daniel Roberts (IREN): Explains the massive scale of physical infrastructure required, noting that the primary bottleneck is now power and "time to compute" 1:21:52. He touches on the economic "Jevons Paradox," where increased efficiency in compute leads to even higher aggregate demand 1:29:56.

• The Future of the Workforce: A debate on whether AI will cause mass unemployment or empower a new generation of "rugged individualists" and solo entrepreneurs 1:00:22.
• The Role of Open Source vs. Closed Source: The competition between massive, resource-heavy models (OpenAI, Google) and the efficiency and specialization of open-source/orchestration models (Mistral, Perplexity) 48:09.
• The Infrastructure Bottleneck: The transition of focus from the scarcity of GPUs to the scarcity of electricity, transmission lines, and trained tradespeople 1:22:34.

Overall Sentiment

The sentiment is mixed, characterized by high excitement for the featured companies—particularly IREN and Mistral—balanced by frustration with the host's interviewing style. While many viewers find the technical and economic insights valuable, there is significant skepticism regarding the long-term viability of certain AI business models and anxiety about the economic implications of AI-driven job displacement.

Recurring Themes

• Strong bullish sentiment toward IREN and its potential in the AI economy.
• Criticism of Jason Calacanis for interrupting guests or dominating the conversation.
• Anxiety and debate surrounding AI's impact on employment and the future of labor.
• Requests to feature other companies in the space, specifically Nebius.
• Skepticism regarding data scraping practices and the sustainability of Perplexity.

Notable Comments

"I am perplexed that perplexity has not gone out of business lol" — @Elementzi, 96 likes

"The way he is explaining Coreweaves Box sounds like Enrons SPV." — @ajphoenix1, 28 likes

"IREN is in an extremely good position to benefit from the new AI economy." — @mldway1, 16 likes

"When Jason is in his element as a moderator or an interviewer, he shines." — @marthajohnson2775, 10 likes

Questions Raised

• How will AI simplify life for the elderly?
• Why does the NYSE need to advertise, especially to potential clients?
• Why would hyperscalers want to use outdated chips, and what other companies besides AI firms need H100s?
• What will people who are involuntarily freed from monotonous tasks by AI do to support themselves?

Dissent / Disagreement

There is noticeable pushback against the host's moderation, with commenters suggesting he needs to "dial back" and let guests speak more freely. Additionally, some viewers presented counterarguments to the presented business optimism, specifically criticizing the "garbage-in-garbage-out" nature of AI and the potential "circular money loop" in CoreWeave's business model.