Quoted on: AI infrastructure and compute
AILatent Space
Evan Conrad, founder of SF Compute, joins swyx to explain why creating an open compute trading market requires building the operational stack of a cloud for physical settlement. They discuss the mechanics of GPU forward curves, multi-constraint order books, and the structural reason NVIDIA and hyperscalers cannot operate an exchange. Conrad also outlines how SF Compute programmatically virtualizes data centers and subdivides InfiniBand fabrics to enable sovereign compute and lower financing costs across the industry.
- Compute sellers dump unreserved near-term capacity at steep discounts because an unsold GPU hour instantly loses all value once the clock runs out. Read →
- Virtualized GPUs of the same SKU often suffer up to a 40 percent performance drop across different operators and configurations, which prevents compute from trading like a standard commodity contract. Read →
- Trading compute looks like software until you have to physically deliver it, forcing SF Compute to run physical data center operations instead of staying a clean marketplace broker. Read →
- Evan Conrad argues that pure cash-settled compute derivatives create severe basis risk because spot indexes track thin, unrepresentative transaction data. Read →
- SF Compute partners directly with power owners and regional data centers to turn raw megawatts into institutional GPU clouds without owning real estate. Read →
- Cloud operators like CoreWeave, Lambda, and Crusoe cannot build a GPU exchange because their direct hardware ownership invites market manipulation on forward contracts. Read →
BusinessTBPN
John Coogan and Jordi Hays discuss cultural milestones in AI including Dario Amodei's SNL parody and the hyper-growth of consumer agents like Instinct, alongside Meta's enterprise pivot poaching MongoDB's CEO. The episode features interviews with Plexo Capital's Lo Toney on technical fund managers, Gossip Goblin's Zack London on AI filmmaking, Andreessen Horowitz's Anish Acharya on agent commerce, SF Compute's Evan Conrad on compute commoditization, and Sean G (Gstaad Guy) on luxury consumption cycles.
- Running browser-use consumer agents currently costs thousands of dollars per user per year, creating a severe unit economic hurdle for consumer software. Read →
- Saturday Night Live parodied Anthropic CEO Dario Amodei, marking the moment frontier AI leaders crossed from tech obscurity into mainstream recognition. Read →
- SF Compute CEO Evan Conrad argues that financializing compute through cash-settled futures fails because compute requires physical settlement and direct infrastructure control. Read →
- Software optimization ruins high-end hospitality because true luxury relies on human friction, family heritage, and intentional gatekeeping rather than speed. Read →
- Instinct scaled to $1 billion in annualized transaction volume in roughly six months, beating Stripe, which took 18 months after two years in beta. Read →
- Product management at a top tech firm shares the exact operational playbook required to evaluate and support pre-seed and seed startups. Read →
- Meta introduced the Meta Enterprise Platform, recruiting former MongoDB chief executive Chirantan 'CJ' Desai as chief enterprise platform officer reporting directly to leadership. Read →