MetaLab CEO: AI Is a 'Race to the Average'
Luke Des Cotes, CEO of design firm MetaLab, characterized the current AI landscape as a “race to the average,” arguing that models are by definition mediocre because they reflect the average of their training data. He contended that defensibility for AI startups no longer comes from proprietary models but from deep workflow integrations, unique data feedback loops, and strong user experience design. Des Cotes stated that what matters is how companies build *around* AI for their customers, not just *with* it.
- Venture capital funding for AI companies surpassed $100 billion in 2024, an 80% increase from 2023, with nearly one-third of all global venture funding now directed toward the AI sector. In the Bay Area alone, AI-related funding reached over $122 billion in 2025, representing more than 75% of all U.S. AI investment. - Enterprise Chief Risk Officers (CROs) are increasingly involved in AI procurement to manage risks associated with the technology; a recent survey found that 91% of middle-market executives are using AI, but 70% of those require outside help to maximize its value. This involvement often lengthens sales cycles as it adds layers of security, compliance, and data governance review to the procurement process. - Agentic AI architectures are moving beyond single Large Language Models (LLMs) to multi-agent systems that orchestrate specialized agents, similar to microservices. Common orchestration patterns include sequential handoffs, concurrent (or parallel) processing, and hierarchical supervision, which allow for more complex and robust workflow automation. - When selling to enterprise sales leaders, productivity metrics are key; they increasingly focus on the quality of customer interactions and conversion rates through the sales funnel, rather than just the volume of calls or meetings. AI tools are being positioned as a way to enhance sales effectiveness by identifying high-potential opportunities and automating tasks like lead routing and account plan creation. - To secure a Series A in the current Bay Area climate, startups need to demonstrate a burn multiple under 2.0, year-over-year growth of at least 50%, and net revenue retention above 120%. The era of "growth-at-all-costs" is over, with investors now prioritizing capital efficiency and a clear path to profitability. - For early-stage founders, personal productivity frameworks like "time blocking" for deep work and managing energy levels—not just time—are critical for avoiding burnout. Leadership in a scaling AI company shifts from direct execution to designing intelligent systems that can handle operational complexity, freeing up founders to focus on strategic decisions. - Emerging hardware trends are focused on enabling AI to interact with the physical world, exemplified by San Francisco-based Bedrock Robotics, which recently raised $270 million to deploy fleets of autonomous construction equipment. This signals a shift from pure software AI to embodied AI that can perform tasks in real-world environments.
