SuperAI showcased a broad mix of startups and enterprises spanning bare metal GPUs, robotics, agentic commerce, observability, and workflow infrastructure. A major theme was the rise of solopreneurs building early products with subscription tools such as Claude or Codex before shifting toward open models and cheaper startup services. Because models have become more capable, a wave of solopreneurs and 2-3 person teams showcased MVPs and raised early pre-seed rounds at the event. Students and non-technical professionals were using Artificial Intelligence tools to build products, including language-learning games for children and narrative intelligence platforms for storytellers.
The market also appears to be moving from tokenmaxxing to tokenoptimizing. Major enterprises burn millions & billions on frontier models while individual developers and startups opt for open models that cost 70-90% cheaper while having 70-90% the intelligence of frontier models. Enterprises are beginning to limit frontier model use or consider open models as spending rises. Engineers and developer relations teams highlighted infrastructure such as ClickHouse and Temporal, pointing to growing demand for Artificial Intelligence observability, analytics, and durable execution. These tools help teams track agent performance, monitor inference spend, identify bugs, and prevent long-running workflows from failing after rate limits or network crashes.
Product design is shifting from generative Artificial Intelligence to copilots and then toward autopilot systems. Instead of users prompting models for every task, more systems are being embedded into existing workflows to deliver outcomes with less active direction. Examples included Buildables, described as Artificial Intelligence copilots for industrial hardware, and ReadyAI, a living knowledge graph that maps relationships, venture capital activity, fundraising news, and scattered data sources. The trend was especially visible in agentic commerce, where systems are being designed to help users discover, evaluate, and buy products directly inside conversational interfaces.
Agentic commerce is emerging in both near-term consumer experiences and longer-term machine-to-machine payment systems. OpenAI and Google are working on search-and-retrieval shopping flows, Meta is building closed-loop commerce across Facebook, Instagram, and Whatsapp, and Stripe has developed an agentic commerce protocol with OpenAI and Meta. Future-facing startups are exploring pay-per-use agentic commerce and payment standards across vending machines, electric cars, and Artificial Intelligence agents. What is not early and is gaining a lot of popularity in 2026 is the Artificial Intelligence business of wholesaling, reselling, aggregation. Demand for cheaper Artificial Intelligence access is supporting gateway providers and AWS resellers that pool startup demand to secure discounts.
