20VC with Harry Stebbings
Showing 31–45 of 691 transcripts.
- 1h 0m
Why OpenAI and Anthropic Won't Win the App Layer | Glean Founder
Industry leaders project that enterprise AI will rapidly commoditize around open-source models, with Arvind Jain forecasting a three-year shift where most workloads run on such alternatives despite ongoing geopolitical concerns over Chinese capabilities. While organizations struggle to quantify ROI due to currently absurd token economics and inefficient inference costs, companies like Glean are betting on aggressive hiring and composite roles to build 10x superior products as the market matures. This strategy involves leveraging frontier model providers as assets rather than competitors, prioritizing specific problem-solving, and preparing for a rigorous audit of AI spending by 2026 to ensure genuine returns on investment.
- 1h 27m
Enterprises Fear Frontier Models | Sam Altman Offers Trump 5% of OpenAI | DeepSeek Builds Own Chips
Sam Altman, Trump, Jason Lemkin, Rory O'Driscoll, Harry Stebbings
OpenAI and the U.S. government are exploring regulatory frameworks involving pre-approval processes and potential equity stakes, while major tech firms like Meta and SpaceX pivot toward selling excess compute infrastructure to reshape market dynamics. In the enterprise sector, adoption bottlenecks arising from talent shortages are slowing diffusion rates despite aggressive service deployment strategies by Microsoft and Amazon. Concurrently, venture capital is normalizing high-dilution financing and secondary liquidity events as geopolitical tensions accelerate a bifurcation between U.S. closed-source models and China's advancing open-source alternatives.
- 1h 11m
Why Now is the Time for the App Layer | Why Startups Should be TokenMaxxing | Mike Mignano, USV
Former founder Mike Mignano joined USV as a General Partner to lead thesis-driven investments in the energy infrastructure layer and application-era AI, leveraging his experience selling Anchor to Spotify. Mignano forecasts a market shift toward autonomous AI agents and cost-optimized routing layers while criticizing current labor displacement narratives and the viability of mid-tier venture funds. Drawing on lessons from Fred Wilson, he emphasizes relationship-building and the "obliterate, don't automate" philosophy to guide USV's focus on reinventing industries rather than merely automating them.
- 1h 12m
Open Models vs Frontier Models: Who Actually Wins? | The $100K Token Budget Every Engineer Will Need
Clay Bavor, Bret, Harry Stebbings, Sundar
Sierra, a $16 billion company serving 40% of the Fortune 500, is redefining enterprise AI by deploying proprietary fine-tuned models that replace traditional customer support with lifecycle management systems for high-stakes industries. Founder Clay Pavore drives this strategy through an "AI-native" organizational model that integrates engineers directly into client environments, utilizes token budgets comparable to headcount, and mandates the use of coding agents for internal processes and hiring. As the firm anticipates market consolidation into a few dominant platforms, it prioritizes rapid iteration via six-week board cycles and a culture of craftsmanship to navigate the escalating competition in frontier intelligence and cybersecurity.
- 1h 18m
Coinbase Cuts AI Spend by 50% | Kalshi's $40B Valuation & Impending IPO | The Year for SaaS Roll-Ups
Jason Lemkin, Rory O'Driscoll, Harry Stebbings
Coinbase achieved a 50% reduction in AI token spend through open-source models, sparking a broader industry shift where CFOs now prioritize measurable revenue acceleration over speculative experimentation. This cost discipline threatens frontier model providers like Anthropic and Microsoft, as market volatility and skepticism regarding their standalone AI products risk decelerating Azure growth and future IPO valuations. Simultaneously, regulators and investors increasingly fear "regulatory capture" where US giants leverage national security narratives to eliminate low-cost open-source competitors, while venture capital firms tighten standards to favor only startups demonstrating rapid scale.
- 1h 6m
Bloom Energy CEO: Why We Aren’t in an AI Capex Bubble | Energy Sovereignty & The Future of Power
KR Sridhar, Leo Aschenbrenner, Harry Stebbings
Bloom Energy CEO Sridhar Ramaswamy leveraged a philosophy of exhaustive risk mitigation and direct operational empathy to scale the company to a $93 billion valuation amidst surging AI-driven electricity demand. The firm now commands a $20 billion order book by deploying modular solid-state fuel cells that deliver power faster than traditional infrastructure while ramping up manufacturing capacity to over 2 gigawatts. Ramaswamy warns that regulatory delays pose the primary threat to this growth trajectory, urging nations to prioritize energy sovereignty and maintain American entrepreneurial competitiveness against global rivals.
- 54 min
How We Got Fred Wilson, Benchmark and Index to Invest $94M | Why Robinhood's Strategy is Wrong
Fred Wilson, Paul Erlanger, Harry Stebbings
FOMO, a mobile trading app with a flat organizational structure and 17 employees, raised $75 million in Series B financing at a $550 million valuation led by Index Ventures and Union Square Ventures to fund vertical infrastructure integration. The company distinguishes itself through a social graph that enables global trading of crypto, equities, and prediction markets while utilizing AI tools to accelerate engineering output without expanding its junior workforce. By treating top performers as an extended founding team and rejecting traditional acquisition tactics for high-value creators, FOMO aims to become a compliant, decentralized alternative to legacy brokerage models.
- 1h 28m
DeepSeek Raises at $50B | The Rise of Open Source vs OpenAI & Anthropic | OpenAI Builds Own Chip
Jason Lemkin, Rory O'Driscoll, Harry Stebbings
Generational DeepMind researchers Noam Shazeer and John Jumper recently defected to Anthropic to escape bureaucratic constraints, highlighting a market shift where top talent prioritizes autonomy over massive financial packages. Simultaneously, China's DeepSeek secured $7.4 billion with exclusive state voting rights to drive AI sovereignty, while its open-source models are closing the performance gap with US closed-source competitors and threatening the profitability of the industry's middle tier. Amidst this geopolitical and technological realignment, hyperscalers face a $725 billion infrastructure spend against a $100 billion revenue gap, forcing a pivot toward strict ROI verification and labor displacement to justify future trillion-dollar investments.
- 1h 17m
Nikesh Arora on the Frontier Model Problem: Breadth vs Depth | The Future of Token Costs
Palo Alto Networks CEO Nikesh Arora outlines a strategic vision where enterprise AI evolves from mere augmentation to agentic judgment, fundamentally reimagining workflows and driving a predicted 50% reduction in administrative headcount. Addressing the cybersecurity landscape, Arora explains how frontier models like "Mythos" accelerate vulnerability discovery while necessitating a shift toward unified, AI-infused gateways to manage autonomous agent traffic. Complementing this technical roadmap, his investment philosophy advocates for early engagement in unproven technologies and warns against current market euphoria, predicting a tenfold drop in token costs that will force a realignment of model valuations based on actual developer utility.
- 1h 21m
Flexport CEO: Two Questions Every Founder Needs to Ask
Ryan Peterson, Harry Stebbings, Peter Thiel, Masayoshi Son
Flexport founder Ryan Peterson advocates for a return to five-day in-office work and outlines an aggressive AI-driven strategy to automate logistics workflows, projecting $600 million in net revenue by next year. The company plans to pursue an IPO based on consistent profitability rather than market trends, leveraging open-source models and building proprietary tools to replace traditional SaaS vendors. Peterson also details an investment thesis centered on motivating founders through "revenge and patriotism" while warning of potential labor arbitrage pitfalls and the strategic risks of relying on a concentrated tech ecosystem.
- 1h 28m
Anthropic's Fable Banned by US Government | Wix & Adobe Hit All-Time Lows | Mistral Raising at $20BN
Everett Randle, Rory O'Driscoll, Harry Stebbings
The transcript analyzes a record-breaking $2.7 trillion SpaceX IPO that generated a $1.2 trillion surge in Elon Musk's net worth while establishing a "gamma squeeze" market dynamic alongside the U.S. government's first AI model ban on Anthropic's "Claude Fable." These financial and regulatory shifts coincide with Salesforce's $3.6 billion acquisition of Finn, which validates outcomes-based pricing models for legacy SaaS firms, contrasting sharply with the market struggles of Adobe and Wix due to AI-native competition. Investors note a strategic pivot away from pre-AI SaaS toward semiconductor stocks and specialized robotics, marking a critical transition in capital allocation as the industry grapples with sovereign AI constraints and hardware scaling challenges.
- 1h 35m
Perplexity CEO: Micron Will Be More Valuable Than Meta & How Export Controls Helped Not Hurt China
Aravind Srinivas, Harry Stebbings
Perplexity AI CEO Aravind Srinivas highlights the company's rapid rise to a $20 billion valuation with 45 million users, attributing their success to a shift from answering questions to orchestrating autonomous agents that drive significant revenue efficiency. Srinivas contends that Perplexity has forced major industry redesigns, including Google's new AI interface, by proving that high-value enterprise workloads can replace traditional advertising models as the primary revenue driver. Looking ahead, he forecasts a market dominated by smaller, highly productive teams managing agent loops, with infrastructure bottlenecks in power and memory becoming the critical constraints for AI's next growth phase.
- 1h 25m
OpenAI vs Anthropic vs Open-Source | Token Maxing, AI Hangovers & The Coming ROI Reckoning
Matan Grinberg, Harry Stebbings
Mattan Grinberg and industry experts analyze a shifting AI landscape where macroeconomic productivity gains and strategic resource reallocation will redefine enterprise team structures and valuation cycles. They predict a transition toward model-agnostic routing and open-source competition, while warning of short-term labor displacement and security challenges that necessitate a focus on high-stakes outcomes rather than traditional output metrics. The discussion further contrasts divergent views on automation's impact, emphasizing that future success depends on building autonomous software factories and aligning engineering talent with critical global problems like healthcare and climate change.
- 1h 16m
SpaceX Launches Largest Ever IPO | OpenAI Files to Go Public | Uber Cuts 23% of HR
Jason Lemkin, Rory O'Driscoll, Harry Stebbings
SpaceX initiated a $75 billion IPO with a fixed $135 share price to value the company at nearly $1.8 trillion, a move that bypassed traditional book-building and led analysts to predict a potential first-day decline despite high institutional interest. Concurrently, OpenAI has filed for its own public offering to fund massive foundation model development as its strategy diverges from competitors like Anthropic, while Apple pragmatically integrated Google's Gemini into Siri to prioritize immediate consumer utility. The broader market context features record efficiency in AI-native startups like Lovable and Cursor challenging traditional SaaS models, though valuation risks remain high as major tech firms ramp up capital expenditures for the AI arms race and traditional tech stocks face correction.
- 1h 15m
Nebius Co-Founder on AI Infrastructure Bubbles | How Price Elastic is Demand for Compute
Roman Chernin, Harry Stebbings
Nebius navigates a critical 6–12 month execution bottleneck by leveraging a four-layer product strategy that evolves from bare metal infrastructure to agentic orchestration, aiming to diversify its client base beyond hyperscalers to thousands of enterprises. Co-founder Roman defends the sector's growth trajectory against bubble concerns, arguing that the Jevons Paradox will drive surging demand for optimized open-source models while European sovereignty initiatives foster a distinct ecosystem of builders. With a 2025 capital expenditure program of $20–25 billion and recent institutional validation from a 5.3% stake disclosure, the company positions itself as a cost-efficient alternative to hyperscalers by prioritizing Total Cost of Ownership reductions over raw hardware pricing.