Interview, Fireside Chat
Bucky Moore @ Lightspeed Venture Partners: Why You Cannot Do VC If You Do Not Do Pre-Seed
- Market outcomes are expected to evolve from the current $30 to $50 billion range to multi-trillion dollar scale within an emerging era defined by companies like SpaceX, OpenAI, or Anthropic.
- Revenue scale, growth rates, and margin profiles of model provider companies will likely differ significantly in the next three to five years as consumer and enterprise applications increasingly adopt AI.
- Companies are just beginning to optimize their business structures by layering products over capital-intensive models, with startups and incumbents identifying deep focus areas where model providers may struggle to prosecute every category.
- A "killer use case for LLMs" involving "Cogen" is anticipated, raising nervousness regarding how model providers might target this specific sector.
- The number of mega platforms is predicted to be limited due to commitment from large capital pools where sophisticated LPs invest with managers for at least three to four funds, and Lightspeed's capital supply is not expected to increase in size in the near term.
- Mega platforms may delay going public for a long time, sustained by private market demand, though the industry will determine over the next few years if blue-chip assets can scale beyond multi-trillion dollar levels.
- The super cycle in AI, defense, robotics, and space remains in its early stages, with potential for upside surprises regarding industrial applicability.
- While most observers expect only a few companies to reach unprecedented scale, the speaker remains open to the possibility of a different outcome.
- Best companies may appear expensive in the short term but are expected to compound for a long duration despite price sensitivity.
- Spreadsheet investors will be pushed to later stages as blue-chip AI companies raise significant early capital before traditional metrics exist, narrowing the window of access for such investors.
- Companies with a deterministic sense of market opportunity, such as those in cybersecurity, are better suited for large raises, while those fleshing out new markets require optionality and smaller capital.
- Solving deep technical problems will take most companies multiple years, requiring investors to remain open-minded about extended timelines.
- The bar for the best ideas is rising regarding growth rates and momentum, and the opportunity to pick winners is becoming harder for large multi-stage firms due to the vast number of founders and founders to evaluate.
- Execution capacity of founders is identified as critical, with past failures attributed to downweighting team execution in favor of product intrigue.
- The fund size range of $500 million to $2 billion is expected to become increasingly challenging as small specialist firms and large platforms gain dominance.
- Lightspeed is expected to continue asserting itself as a global multi-stage generalist platform over the next 10 years, with the next generation of venture relying on collaboration between multi-stage firms and seed firms for downstream capital.
- Model provider companies are viewed as venture-funded assets where the narrative is still being written, and every frontier model lab is expected to build both business and consumer-facing products beyond APIs.
- The arrival of AGI is felt to be present in many ways due to test time compute capabilities that allow models to solve problems beyond human capacity, even if overall progress halts.
- New scaling dimensions for AI are anticipated to persist as human ingenuity remains concentrated in research labs, making it difficult to imagine a world where ideas run out.
- Enterprise AI adoption is described as fundamentally different from cloud adoption, driven by existential fears of company demise that create an unprecedented appetite and urgency.
- Liquidity challenges are expected to be resolved by new investment products, specifically continuation funds devised by investment managers.