Fireside Chat, Interview
AI Sticker Shock: Open Source as the Answer to Broken Tokenomics | Vipul Ved Prakash | RAISE 2026
- Revenue is projected to sustain growth beyond the $1 billion threshold achieved earlier this year, underpinned by a continued AI boom expected to last for many years despite global compute and power headwinds.
- The production AI design space is being evolved to handle agentic workloads capable of serving hundreds of trillions of tokens, with a strategic shift away from vertically integrated closed models toward a disaggregated ecosystem of thousands of software vendors.
- Open source models are predicted to become the industry's general intelligence platform, with usage growth rates expected to reach 10,000x compared to the 7x growth observed in closed models like Google Gemini, while the company anticipates closed models failing to dominate as the industry matures.
- A significant market migration is expected as digital natives and financial institutions shift workloads to open models to avoid data sovereignty risks and prevent proprietary intellectual property from being used by major labs that may enter their specific sectors.
- The company plans to deploy full platforms for post-training and scaling while guaranteeing 100% assurance that customer data remains within specific residency constraints and is never used to train new frontier models.
- Infrastructure investments will prioritize U.S. and European facilities, including hosting in Europe for GDPR compliance, to ensure data control and avoid the concentration of market control by two closed model providers.
- More frontier open source models beyond the current five or six, including Nemotron as the U.S. entry, will be released, fostering the emergence of specialized ecosystems for operating open weights, such as those providing reinforcement learning and post-training services.
- The company is building global infrastructure through partnerships with other clouds and neo-clouds alongside its own facilities, aiming to host workloads locally before local demand materializes to ensure infrastructure viability.
- Sovereign partnerships across two continents are being initiated to address compute and power access challenges, with infrastructure investment underwritten to ensure positive returns from the start and modeled to resemble energy independence.
- Strategic considerations include potentially distributing equity to a GDP-weighted group of sovereign nations (France, UK, U.S., Spain, Australia) to secure global access and curry favor, alongside a recommendation for market participants to continuously update priors regarding infrastructure build-out.