AI in 2026: 3 Predictions For What's To Come (a16z Big Ideas) - YouTube
a16z partners argue AI's biggest 2026 opportunities aren't in automation or productivity—they're in reinforcing business models that drive revenue, facilitating human connection, and human-AI collaboration in scientific discovery.
Read Original Summary used for search
TLDR
• AI that reinforces business models (drives revenue) has unlimited market pull vs. cost reduction—EVE's plaintiff law platform and Salient's voice agents improve outcomes, not just efficiency
• Consumer AI shifts from productivity to connectivity in 2026—the core emotion is "wanting to be seen," with AI-to-AI communication potentially facilitating relationships
• Autonomous labs will emerge through human-AI collaboration, not full automation—interpretability matters more than closed-loop systems in the near term
• Compounding advantages come from proprietary outcomes data that can't be scraped from the public internet—EVE's intake-to-outcome data creates defensibility
• Startups can beat incumbents if they create net new interaction models that don't fit existing platforms
In Detail
Three a16z partners present contrarian takes on AI's 2026 trajectory. David Haber argues the real opportunity is AI that reinforces business models rather than just cutting costs. EVE, operating in plaintiff law where attorneys work on contingency, uses AI to help lawyers take on more clients and win more cases—reinforcing their revenue model rather than eroding billable hours. Salient's voice agents in loan servicing don't just reduce call center costs; they drive better collection rates. This creates unlimited market pull because customers see direct revenue impact, not just efficiency gains.
Brian Kim predicts consumer AI will shift from productivity tools to connectivity applications. The core human emotion is "wanting to be seen by others," and AI's deepening understanding of users—through digital footprints, photos, conversations—enables new relationship dynamics. He envisions AI-to-AI communication facilitating human connections: "my AI coming to your AI" to suggest conversations or check-ins. Startups can win against incumbent platforms if they create genuinely new interaction models and atomic units that don't fit existing social networks.
Oliver Shu focuses on autonomous labs, arguing the near-term is human-AI collaboration, not fully closed-loop systems. The combination of AI reasoning capabilities and robotics is new, but interpretability is critical—scientists need to understand why the system plans experiments a certain way. Market dynamics will drive adoption: pharma, chemicals, and materials science have established buyers for research outputs. Companies like Periodic Labs, Medra, and Chemifi are building toward this, alongside public-private collaborations like the DOE's Genesis mission. The compounding advantage in all three domains comes from proprietary data—EVE's outcomes data, relationship graphs in consumer apps, and experimental results in autonomous labs—that can't be replicated from public sources.