Fireside Chat, Interview
Are Burn Multiples BS in an AI World? & Sam Altman Needs $1TRN of Energy
- Pricing Philosophy in 2025: Venture capital deals are now priced either on "hope" (growth trajectories) or on "multiples" (fundamentals); VCs are increasingly rejecting companies under $100M ARR that cannot demonstrate a path to a massive IPO, even if they have reasonable growth, because they offer no "upside option value" at subscale.
- The AI Burn Multiple Anomaly: AI-native companies under $100M ARR exhibit severe negative free cash flow margins (-126%) compared to non-AI peers (-56%), yet maintain lower (more efficient) burn multiples because they generate revenue at such a rapid velocity that capital efficiency per dollar of burn remains high.
- Limitations of Capital Efficiency Metrics: The "burn multiple" metric (ARR added per dollar of burn) relies on assumptions that are breaking down in the current market, specifically that ARR is real, churn is accurately reported (often hidden by hyper-growth), and that CapEx is negligible; the metric is no longer a universal truth and requires scrutiny of underlying data honesty.
- Funding Rejections for Non-AI Storylines: Founders with strong growth (e.g., 2x-3x year-over-year) and good burn multiples are struggling to raise funds if their TAM is not perceived as "enormous" or if they lack an explicit AI narrative, forcing advice to founders to "take the deal" immediately rather than optimizing for valuation.
- The "Kingmaker" Effect: Investor syndicates are exhibiting increased aversion to funding "number two" or "number three" companies in segments where a "kingmaker" competitor (e.g., Harvey, a16z-backed firms) has secured dominant funding, creating a binary funding environment where proximity to a funded leader acts as a deterrent.
- M&A Consolidation Strategy: With 15 IPOs occurring year-to-date but hundreds of unicorns remaining, firms like Andreessen Horowitz are aggressively consolidating portfolio companies (e.g., 5tran buying dbt) to create critical mass; this involves accepting dilution of individual stakes (e.g., dropping from 20% to 8%) to achieve a public exit.
- OpenAI's Energy and Capital Ambitions: OpenAI's growth trajectory implies a need for 125x current energy capacity and hundreds of billions in CapEx within eight years; while the technical path is viewed as feasible, the financial scale (potentially trillions) may lead to a reality check where growth slows significantly before the "trillion-dollar" vision is realized.
- Valuation Risk and Public Market Multiples: Public B2B software companies trade at historically elevated multiples (e.g., 20x-26x revenue for 30% growth), creating a risk that if these benchmarks revert to historical norms (6-8x revenue), private valuations built on "hope" will face severe downward pressure.
- Private Equity (PE) Disruption: The static product lifecycle of the past decade (where software models remained valid for 10 years) is ending; PE firms face new risks as AI accelerates product obsolescence and AI agents reduce the need for per-seat licensing, threatening the traditional "buy-and-optim" LBO model.
- Monetization Experiments: AI models (e.g., ChatGPT) are exploring direct commerce integrations and advertising to monetize free users, though historical data suggests users are more comfortable with ads on social platforms than purchasing directly within AI interfaces.
- CEO vs. CFO Tension on AI Scale: While visionary CEOs are "willing" trillion-dollar projects into existence, there is a growing concern that the market is borrowing against these metaphors; a scenario where AI growth slows to 50% (still impressive) but requires less capital than projected is deemed a likely, though less celebrated, outcome.
- Political Expression and Fiduciary Duty: The consensus among VCs is that companies should generally avoid taking public political stances to stay out of culture wars, though individual CEOs retain the right to personal expression; boards rarely intervene unless personal opinions directly alienate a significant portion of the customer base or workforce.