Conference Presentation, Fireside Chat
The State of AI: Growth, Fragmentation, and the Next Wave
- AI companies are projected to grow faster and larger than previously anticipated, with value accruing across all stack layers including models, infrastructure, and applications.
- The landscape is characterized by a paradox of accelerated value creation and higher risk of rapid market wipeout within shorter timeframes.
- Foundation model and AI application markets are expected to fragment while growing significantly, with top-tier frontier labs projected to surpass the early revenue ramps of historical SaaS leaders and major hyperscalers.
- Model inference costs are projected to decrease by 10x year-over-year with a trend of continuous price declines, driven by factors including Chinese open-source models.
- AI-native companies are projected to far outpace traditional SaaS counterparts in time-to-$100M ARR on an average basis, delivering over 10x customer experience improvements compared to typical 25-50% SaaS 2.0 increments.
- Specialized AI apps with deep customer data integrations are expected to outperform foundation model layers in complex workflows, while the concept of "GPT wrappers" is dismissed as a valid category in favor of major infrastructure evolution.
- Founder behavior is expected to shift toward intensive office presence, with most AI-native founders working six to seven days a week to avoid the 2021 remote culture.
- Post-initial model-based growth, companies are projected to revert to traditional software business models, such as two-sided marketplaces and long-tail integrations, to address retention.
- A return of strong brand effects similar to the early internet era is anticipated, where brand recognition drives adoption even with equal competitor technology.
- Customer purchasing behavior has shifted over the last 12 months from "vibe-based" buying to a focus on tangible ROI, necessitating that investors be "in the field" with higher strategic precision.
- Specific market examples include Cursor gaining traction via VS Code behavior and RL waves, with enterprises reporting 30-50% productivity gains and some achieving a 10x lift; 90% of code within Cursor users' companies is expected to be AI-generated.
- Decagon customers are projected to slash support costs by up to 80%, increase deflection rates from 30% to 60-80%, and double customer satisfaction scores.
- The AI super cycle is expected to follow a prosumer-to-enterprise adoption pattern, with enterprise pipelines projected to be higher than ever, including million-dollar deals closing on the first day of an Account Executive's tenure.
- Enterprise valuation requirements now demand demonstrated customer stickiness and the ability to convert top-of-funnel traffic into revenue, with foundation model startups facing immense pressure to prove execution over storytelling.
- The market is not expected to be a rising tide scenario, requiring winners to be picked carefully due to limited capital and significant conflicts in early investment decisions.
- Investment strategy plans to avoid the state-of-the-art model market except for primary names with prior success and premium teams capable of securing capital, focusing instead on market leaders with momentum and vertical application vision.
- The opportunity is described as incredibly bullish with no inherent endemic mode in the stack other than solving the customer bootstrap problem, urging firms to remain active in the market to avoid irrelevance.