DreamWorks & the Science of Storytelling | Jeffrey Katzenberg & ChenLi Wang, WndrCo

Discussion of 2025 as the year AI moved from hype to tactical deployment, with real productivity gains emerging. Katzenberg warns of an impending reckoning as $700B in hyperscaler capital investment over 24 months must produce ROI. Both investors see continued opportunity even if model development froze today, emphasizing the decade of work needed to productize AI effectively.

• •$700B capital investment in AI infrastructure over 24 months requires demonstrable ROI by 2026-2027
• •2025 marked shift from 'show your friend' moments to real business transformation with AI
• •Even freezing model development now would leave a decade of productization work
• •Not all hyperscalers will win - expect market consolidation rather than winner-take-all
• •Real value creation is emerging but enthusiasm got ahead of tactical reality

ChenLi describes the most rewarding investments as those where they supported founders through dark moments before breakthrough success. Highlights portfolio companies transforming professional workflows - Harvey for legal, Abridge for healthcare, Cursor becoming ubiquitous in startup offices. Emphasizes the importance of building relationships early and supporting through setbacks.

• •Cursor went from zero to ubiquitous (visible in every startup office) in 12 months
• •Most rewarding investments involve supporting founders through 'dark moments' before success
• •Harvey and Abridge are transforming how lawyers and doctors work daily using AI
• •1Password and Granola achieved near-ubiquity in startup environments
• •Focus on inputs, grit, and founder support rather than just outcomes given macro/luck factors

Katzenberg reveals the recurring theme of his career: using technology as competitive advantage for storytelling. From trying to buy Pixar from Steve Jobs for $50M to building Shrek in Silicon Valley with 800 employees, he's always imported innovation from tech into Hollywood. This pattern led to founding WndrCo as a third act, combining his storytelling expertise with technical partners.

• •Attempted to buy Pixar from Steve Jobs for $50M in 1989-1990 when Jobs was financially stretched
• •First Shrek movie made in Redwood Shores with 800-person studio, establishing Silicon Valley presence
• •Every career chapter involved importing latest tech innovations from Silicon Valley to Hollywood
• •Recognized common patterns between finding great founders and great storytellers - both are 'truffle hunting'
• •Partnership is foundational - sought technical partners (ChenLi, Sujay) to complement storytelling expertise

WndrCo evolved from investment holding company to hybrid model inspired by Founders Fund, a16z, and Thrive. Started with conviction in building companies based on Dropbox scaling experience. Now operates three strategies: building 1 company/year with $50-75M investment, venture investments of $5-15M, and $50M seed fund with $500K checks. The holding company generates $400M annual free cash flow.

• •Built 8 companies since 2018; holding company will generate ~$400M free cash flow in 2026
• •Core fund does ~1 build per year ($50-75M investment) plus venture investments ($5-15M)
• •Separate $50M seed fund makes $500K investments alongside other seed investors
• •Model inspired by Founders Fund, Andreessen Horowitz, and Thrive Capital
• •Started as holding company in 2016, evolved strategy after 1Password opportunity was too large to acquire

Case study of WndrCo's company-building process using Wonder Health. Identified longevity as massive market opportunity where 95% is 'snake oil' not driven by actual science. Built company by assembling 70 leading research scientists globally under one umbrella to share data across verticals (mental acuity, bone density, metabolism, etc.). Process involves relentless founder/scientist meetings and cold emails.

• •Longevity market is 95% snake oil sales not driven by actual science - saw gap for science-based approach
• •Assembled 70 leading research scientists globally under Wonder Health umbrella
• •Longevity requires expertise across multiple verticals: mental acuity, bone density, skin, metabolism
• •Company building process: identify big idea, relentless meetings with experts, cold emails, storytelling to recruit talent
• •Sometimes unique market view or underfunded areas require building rather than investing

Katzenberg argues storytelling is fundamental to every business aspect, not just entertainment. Required to recruit top talent (70 scientists for Wonder Health), raise capital, sell to enterprise/consumers. Harvey and Abridge founders specifically sought WndrCo's storytelling help despite having great products. Distinguishes between good storytelling to bad outcomes (snake oil) versus good storytelling to good outcomes.

• •No business can succeed without storytelling - recruiting talent, raising capital, selling products all require it
• •Harvey founder explicitly said 'I need help with storytelling' despite having crushing product
• •Abridge needed storytelling for both enterprise sales and doctor adoption (demand side)
• •Snake oil in health is 'good storytelling to bad outcome' - evil storytelling
• •WndrCo's differentiation: 'our money is same color as everyone else's' - storytelling is the edge

ChenLi argues AI benchmarks fail to capture what makes great products, just as standardized tests fail to assess human potential. Emphasizes product craft through Granola example: auto-correcting esoteric company names by doing web searches shows obsessive attention to detail. This Steve Jobs-level craft (making circuit boards beautiful) is rare and more important than benchmark performance.

• •Benchmarks for AI are like standardized tests for kids - they miss what actually matters
• •Granola automatically corrects esoteric company names via web search - edge case obsession shows craft
• •Steve Jobs obsessed over circuit board aesthetics nobody would see - same craft mentality needed
• •70% of use cases don't require frontier models - last mile product work creates magic
• •Passion is infectious and can't be hidden behind PowerPoint - do live demos and product installs

Core investment thesis centers on people and generational founders rather than ownership percentages or specific rounds. ChenLi learned from NEA's Dick Kramlich and Scott Sandell to make irrational bets on founders even when models predict failure. Dropbox taught first-principles talent evaluation - hiring people without fancy resumes who became successful. This approach shaped WndrCo's founder selection and portfolio company hiring advice.

• •No strong view on ownership percentage or round - invest in generational founders regardless
• •Best returns come from earliest investments, but betting on great founders works across stages
• •NEA mentors made 'irrational' bets on founders when models predicted failure - gave founders confidence
• •Dropbox hired talented people without pedigrees who became successful - first principles talent evaluation
• •Seed fund ($50M, $500K checks) allows working with former Dropbox colleagues now VCs

Katzenberg credits Barry Diller as most impactful mentor, making him studio head at 30 despite being 'irascible, tough, grinding, demanding.' Learned storytelling from Spielberg, animation principles from Walt Disney archives ('movies only as good as their villains'). ChenLi cites NEA's Kramlich/Sandell for integrity under pressure, Dropbox founders for talent-first hiring, and Charlie Munger for even-keeled curiosity.

• •Barry Diller made Katzenberg studio head at 30 - invested in potential he didn't see in himself
• •Walt Disney principle: 'movies only as good as their villains' - Ursula, Scar, Farquad all great antagonists
• •NEA's Dick Kramlich and Scott Sandell showed integrity and irrational founder support during tough times
• •Dropbox taught first-principles hiring - bet on talent without pedigrees, many became successful elsewhere
• •Charlie Munger exemplifies even-keeled gratitude with fierce curiosity - evergreen learning mentality