Invest Like the Best with Patrick O'Shaughnessy
Krishna Rao - Anthropic's CFO on Compute, Scaling to $30B ARR, and the Returns to Frontier Intelligence - [Invest Like the Best, EP.472]
· Summarized
Key takeaways
- Compute is central to AI development and must be managed with discipline, flexibility, and strategic planning to stay at the research frontier.
- Anthropic employs a robust, scenario-based approach to compute procurement, leveraging multiple chip platforms and constantly updating demand models to hedge against uncertainties.
- The company’s focus on frontier models yields exponentially increasing value, especially in enterprise, with high returns on investment as capabilities improve.
- Building a strong foundational platform allows for broad external adoption and the development of specialized vertical applications, exemplified by products like Claude and Claude Code.
- Stable and strategic pricing maximizes utilization and overall ROI, with recent model improvements allowing aggressive decreases in price without sacrificing efficiency.
- Deep collaboration with cloud providers and high internal compute fungibility enable Anthropic to rapidly scale and optimize AI infrastructure aligned with exponential demand growth.
- Internal AI tools significantly boost organizational productivity, deepen talent engagement, and foster a culture of transparency and rigorous debate.
- Society's perception and regulation of AI are critical, requiring responsible, phased releases and open dialogue about risks and benefits.
- The future of AI involves virtual collaborators with long-term memory and contextual understanding, promising to accelerate productivity across industries.
- Leadership must adopt exponential thinking, remain open to new paradigms, and continuously adapt to rapid technological advancements.
Summary
The Critical Role of Compute at Anthropic
Krishna Rao emphasizes that compute is the lifeblood of Anthropic, likening it to a canvas on which everything else is built. The decision to procure and allocate compute is highly consequential and challenging, balancing the risk of overbuying against under-serving customer demand. The company models demand carefully and builds flexibility into procurements, including utilizing three different chip platforms (TPUs, GPUs, Trainium) fungibly to maximize efficiency across model development, internal research, and customer service.
Navigating the Cone of Uncertainty in Compute Planning
Krishna introduces the concept of the cone of uncertainty, highlighting how exponential growth makes predictability difficult over short periods. To manage this, Anthropic forecasts multiple scenarios and plans for the upper bounds of demand, maintaining flexibility and high utilization of compute resources. This disciplined approach ensures they stay at the frontier, balancing growth and risk effectively.
Strategies for Efficient Compute Utilization and Investment
The company employs a sophisticated orchestration layer to optimize the use of various chip generations and platforms. They invest heavily in custom tooling, compilers, and collaboration with chip suppliers like Amazon, Google, and Nvidia to maximize the ROI on their compute spend. Krishna discusses how recent model improvements dramatically increase efficiency and capability, reinforcing the high returns of frontier AI and the importance of staying at the cutting edge.
The Business and Technological Evolution of Frontier AI
Krishna explains that the returns to frontier intelligence are increasing, especially in enterprise applications, where new models unlock exponential growth in capabilities and ROI. He describes the process of model scaling laws holding true, and how deploying the best models yields immediate benefits for customers. The internal focus on recursive self-improvement and research accelerates progress, with talent and tooling amplifying the company’s leap forward.
Platform vs. Application Layer and Building Products
Anthropic’s strategy is predominantly platform-oriented, providing foundational AI capabilities and tools like Claude. They collaborate with customers to develop vertical solutions such as Claude Code, which writes its own code, part of a broader effort to create virtual assistants that enhance productivity. Krishna highlights the importance of building broadly accessible platforms that empower external developers while also creating vertical products when it aligns with their vision of AI’s future.
Pricing, Market Dynamics, and Margins in AI
Krishna discusses the stability of model pricing, emphasizing how lowering prices can boost utilization and overall value—citing Jevons paradox. Despite the capital intensity, they prioritize long-term ROI over immediate margins, balancing customer value with the company’s return on compute investments. Strategic pricing and stable margins support scaling adoption and maintaining a competitive edge.
Collaboration with Chip Providers and Organizational Flexibility
Anthropic’s deep collaborations with cloud providers like Amazon, Google, and Microsoft are integral, involving capacity planning and joint development. Krishna explains the high fungibility of compute use within the company and their ability to rapidly deploy additional resources across different workloads, which is crucial given the exponential demand for compute infrastructure.
The Use of AI in Internal Operations and Talent Culture
Krishna details how Anthropic leverages AI tools like Claude internally for financial modeling, reporting, and decision-making, significantly increasing productivity. He emphasizes that AI has augmented talent rather than replaced it, enabling smarter decision-making and faster iteration. The company fosters a culture of transparency, rigorous debate, and collaboration, with leadership openly communicating and engaging with employees.
Perception, Regulation, and Ethical Considerations in AI Development
Krishna acknowledges society’s fears about AI and stresses the importance of transparency, collaboration with regulators, and responsible deployment. The episode highlights the phased release of models like Mythos, emphasizing safety measures. The company advocates for balanced discourse around AI’s risks and benefits, aiming to ensure technology supports human progress responsibly.
The Future of AI: Virtual Collaborators and Exponential Growth
Looking ahead, Krishna is excited about AI evolving into virtual collaborators that integrate seamlessly with organizational workflows, vastly boosting productivity. He envisions AI assisting in complex research, drug discovery, and long-term planning. The potential for AI to transform industries and improve lives is immense, driven by ongoing model improvements, research, and talent.
Personal Reflection and Leadership in Rapidly Evolving AI Landscape
Krishna discusses how he personally adapts to exponential change, emphasizing first principles thinking, hiring exceptional talent, and maintaining curiosity about future possibilities. He recalls a transformative walk with Tom Brown, who predicted rapid model advances, reinforcing the importance of optimism and openness while managing uncertainties.
Sponsors
- Ramp · Ramp is the only platform built to make your finance team leaner, faster, and better, saving businesses 5% annually on average so you can stay focused on growth. Ramp customers grew revenue 3.2 times faster than the average American business. Visit ramp.com slash invest for more.
- Vanta · Vanta automates security and compliance for over 16,000 fast-moving companies, helping them stay audit-ready and monitor risks across vendors and AI tools. It acts like a 24/7 GRC engineer to find issues, draft fixes, and cut assessment time by up to 50%. Get $1,000 off at vanta.com slash invest.
- Colossus · Become a Colossus member to get our quarterly print magazine and private audio experience, with in-depth profiles and early access to episodes. Learn more at colossus.com/subscribe.
- RidgeLine · RidgeLine is redefining asset management technology as a true partner, enabling faster growth and smarter operations for investment firms. Visit ridgelineapps.com to discover what they can unlock for your firm.
- Rogo · Rogo is the AI platform built specifically for finance, creating agents trained to understand and support Wall Street workflows from diligence to modeling. Learn more at rogo.ai/invest.
Show notes
My guest today is Krishna Rao, the CFO of Anthropic. The center of our conversation is how he navigates the decision around procuring and allocating compute, which he describes as the canvas on which everything else gets built.
We talk about what he calls the cone of uncertainty, the three chip platforms Anthropic uses fungibly across Trainium, TPUs, and GPUs, and the daily meetings they run to allocate compute between model development, internal use, and serving customer demand.
He explains why the returns to frontier intelligence keep getting higher, especially in enterprise, and how Anthropic thinks about the line between platform and application and why they choose to build their own products like Claude Code.
Krishna has such a unique seat watching one of the fastest growing businesses in history, and he is generous in sharing what he has learned since joining the company two years ago.
For the full show notes, transcript, and links to mentioned content, check out the episode page here.
-----
Become a Colossus member to get our quarterly print magazine and private audio experience, including exclusive profiles and early access to select episodes. Subscribe at colossus.com/subscribe.
-----
Ramp’s mission is to help companies manage their spend in a way that reduces expenses and frees up time for teams to work on more valuable projects. Go to ramp.com/invest to sign up for free and get a $250 welcome bonus.
-----
Trusted by thousands of businesses, Vanta continuously monitors your security posture and streamlines audits so you can win enterprise deals and build customer trust without the traditional overhead. Invest Like the Best listeners get a special offer of $1,000 off Vanta when you go to vanta.com/invest.
-----
WorkOS is the infrastructure B2B and AI-native companies use to sell to enterprise. It covers everything enterprise security requires: SSO, SCIM, RBAC, Audit Logs, AI governance, and more. Trusted by 2,000+ fast-growing companies, including OpenAI, Anthropic, Cursor, and Vercel.
-----
Rogo is the AI platform for finance. They're building agents for Wall Street that are trained to understand how bankers and investors actually do work: from diligence and modeling, to turning analysis into deliverables. To learn more, visit rogo.ai/invest.
-----
Ridgeline has built a complete, real-time, modern operating system for investment managers. It handles trading, portfolio management, compliance, customer reporting, and much more through an all-in-one real-time cloud platform. Visit ridgelineapps.com.
-----
Editing and post-production work for this episode was provided by The Podcast Consultant (https://thepodcastconsultant.com).
Timestamps:
(00:00:00) Welcome to Invest Like The Best
(00:02:29) Episode Intro: Krishna Rao
(00:03:14) Compute as Anthropic's Lifeblood
(00:05:17) Three Fungible Chip Platforms
(00:07:31) The Cone of Uncertainty
(00:09:08) Competing Ways to Allocate Compute
(00:10:36) What Drives Compute Efficiency
(00:12:38) Why Frontier Returns Are So High
(00:16:32) How Claude Code Writes Its Own Code
(00:18:46) Will Talent Become Obsolete?
(00:20:07) How Scaling Laws Are Holding
(00:21:54) Exponential Thinking
(00:23:17) The Layer Cake of Compute
(00:26:36) How Anthropic Deploys New Compute
(00:27:53) Platform v. Application Layer
(00:32:42) Why Model Pricing Has Stayed Stable
(00:35:26) Measuring Return on Compute
(00:37:22) Working With Chip Providers
(00:38:32) How Anthropic's Finance Team Uses Claude
(00:41:32) The Jevons Paradox for Labor
(00:43:08) Anthropic's Fundraising & Growth Journey
(00:47:31) The Exponential Revenue Curve
(00:49:02) The Hardest Thing to Explain to Investors
(00:52:15) AI's Public Perception Problem
(00:55:38) Mythos
(00:57:31) Relationship With Government
(00:58:51) Inside Anthropic's Culture
(01:03:48) The Next Frontier: Virtual Collaborators
(01:06:22) How Leaders Scale With a Business
(01:10:55) The Biggest Risks to Continued Progress
(01:12:09) What Krishna is Excited About
(01:13:45) The Kindest Thing
Follow this show on AudioBrevity and new episodes get summarized for you. Create a free account