20VC Deep Dive: Canva's Growth Downgrade, AI Compute Cost Dilemma, Google Talent Exodus, and Musk's $55 Billion Chip Factory

"Canva Slashes Growth | Talent Exodus at Google | Revolut's $50B CEO Package | Musk's $55B Terrafab" (20VC with Harry Stebbings, featuring guest investors Rory O'Driscoll and Jason Lemkin), here are the key points summarized: 1. Canva's Growth Slowdown and AI Inference Cost Shock • Growth under pressure and soaring costs: • Canva (annual revenue of about $3.6 billion) expects its growth rate to slow from 30% to 20% by year-end, primarily due to the integration of cutting-edge AI features leading to a surge in inference costs, making it difficult for the company to continue subsidizing low-cost subscriptions with high-cost compute. • "End of the No-Code Era": • No-code tools like Airtable, Notion, and Canva were once disruptive products, but with the rise of ChatGPT, Claude, and various AI agents, ordinary users (Prosumers) are more inclined to generate designs, posters, or even complete websites directly using natural language. • Many teams' self-developed agent marketing systems have completely bypassed traditional UI software like Canva when scheduling tasks. 2. Exodus of Top Talent from Google: Jeff Dean and Demis Hassabis Adjust Roles • Departure of a 27-year veteran and commercialization friction: • Google AI founder Jeff Dean has left Google to start his own venture with his team, while DeepMind founder Demis Hassabis has transitioned to a non-full-time management/chairman role. • The core conflict lies in resource allocation: the capital market forces Google to prioritize compute resources for cloud business (Google Cloud) and quickly monetizable coding/chat large models, while scientists prefer to allocate compute to biomedical, fundamental science, and ultimate exploration. • Extreme demand for "science-based AI": • Top researchers no longer need to rely on large bureaucratic companies; they can independently raise billions for frontier exploration based solely on their personal reputation. 3. Competition for Top AI Talent and "God-Tier Compensation" • Complete differentiation in compensation levels: • The tech industry has formed three completely different compensation systems: regular employees, conventional AI engineers, and a very small number of "God-Tier" core talents with experience in core pre-training and supercomputer scheduling (annual salaries or equity reaching millions to tens of millions of dollars). • Mature companies must establish special zones or incentive programs to prevent top teams from being poached by giants like Meta, Anthropic, and OpenAI. 4. Musk's TerraFab Chip Factory and Data Center Resistance • Breaking free from TSMC's capacity bottleneck: • Musk announced an investment of over $10 billion to build the TerraFab chip manufacturing and energy integration base, aiming to bypass TSMC's multi-year order restrictions and achieve complete vertical integration from chips, energy to compute. • Challenges in data center infrastructure and power constraints: • Some regions in the U.S. are showing resistance to high-energy, high-water-use data centers; however, due to regulatory competition among states, compute construction is shifting to areas with cheap electricity and policy support. 5. Revolut's $50 Billion CEO Equity Incentive and Founder Model • Benchmarking Musk's super ladder incentive: • Revolut plans to introduce a large incentive scheme tied to valuation milestones (up to 40% equity if a $500 billion valuation is achieved). • Irreplaceability of founders: During the tumultuous AI transformation period, companies led by professional managers are at risk of becoming victims of disruption; granting exceptional founders strong control and excessive incentives is a necessary cost to protect the company's continuous evolution. 6. Reality Differentiation in Enterprise Software (SaaS) • Atlassian and Palantir's strong comeback: • Atlassian has dispelled doubts with solid financial reports and execution; Palantir has achieved nearly 100% revenue explosion due to its strong frontline engineer (FDE) team and delivery model deeply tied to business results. • Intensified competition in SMB software: • Traditional CRMs (like HubSpot) are facing a barrage from lightweight native new competitors that rapidly iterate using AI tools in the low-end market.