OpenAI Signals a ‘New AI Era’ — and Warns Markets Are Behind the Curve
TREE NEWS reports: OpenAI capped its annual developer conference with more than 20 product launches, headlined by a new personal agent platform called Dots. In the hours that followed, Tibo Sottiaux — the executive who leads ChatGPT, Codex and Dots — delivered a blunt message: investors have not yet priced in the second-order effects of AI’s next leap. His central claim is that model capabilities will improve roughly tenfold within a year, and that this compounding curve will rewire how the internet works before most portfolios adjust.
The Three Trends the Market Is Missing
Sottiaux laid out three shifts he believes are underappreciated:
- Agent-driven traffic: The majority of online actions will be executed by AI agents rather than humans, forcing every product and API to be rebuilt for machine-scale volume.
- Relentless cost deflation: Models will get dramatically cheaper and faster, collapsing the economics of AI-native applications and destroying margin assumptions built on today’s pricing.
- Seamless multimodality: Voice, vision, text and action will merge into a single continuous interface, making today’s app-centric UX obsolete.
He pointed to Notion as a case study: once the productivity platform exposed an MCP (Model Context Protocol) interface to agents, traffic surged and strained its infrastructure and unit economics. In his view, most AI startups are building for a world that is about to disappear.
The ‘Sleeper Hit’: An Open Ecosystem With Revenue Sharing
While Dots grabbed the headlines, Sottiaux argued the most consequential — and least understood — announcement is OpenAI’s push toward an open ecosystem with shared economics. Sixteen partners have already integrated “Sign in with ChatGPT,” and OpenAI is introducing revenue sharing for high-usage, high-retention plugins distributed across its roughly 1.2 billion users. Discovery will be driven by retention and quality metrics rather than SEO or branding — a direct challenge to the app-store playbook and to platform incumbents that monetize distribution.
Roadmap Details and a Deliberate Restraint on Model Releases
Sottiaux confirmed that ChatGPT’s Chat and Work modes will merge, that Dots’ capabilities will be embedded directly into ChatGPT, and that multi-agent “dots” — virtual teams with specialized roles — are coming soon. He also disclosed that OpenAI has an internal model, “6.1 Astra,” that it chose not to release. Instead, the company is devoting a large share of compute to a dual-track safety stack: one track executes tasks, the other monitors the primary agent for risky actions or prompt injection.
Market Implications
Equities
The clearest beneficiaries remain AI infrastructure — GPU suppliers, hyperscalers, networking and power providers — because a tenfold capability jump implies sustained compute demand even as per-token costs fall. The risk sits with application-layer software companies whose moats depend on interface complexity, manual workflows, or distribution through app stores. SaaS firms that fail to expose agent-ready interfaces could face both traffic shocks and pricing pressure. OpenAI’s revenue-sharing model also threatens the economics of incumbent app marketplaces.
Bonds and Rates
Cheaper, faster models reinforce the disinflationary productivity narrative that has supported long-duration bonds, but the near-term capex boom in data centers and power infrastructure is inflationary for rates at the margin. Expect continued tension between AI-driven efficiency gains and AI-driven capital spending.
Crypto and Digital Assets
An agent-dominated internet strengthens the case for machine-to-machine payments, programmable wallets and decentralized compute markets. Agent-native economies need settlement rails that operate without human intervention, which is precisely the thesis behind on-chain agent frameworks, GPU networks and AI data marketplaces. Tokens tied to decentralized inference and compute could see renewed narrative momentum, though fundamentals remain thin.
Commodities and Currencies
Rising compute demand translates into structural pressure on electricity, uranium, copper and natural gas. On currencies, the dollar retains its advantage as the denomination of AI infrastructure contracts and cloud billing, while economies that can supply cheap power and talent stand to attract capital flows.
Key Takeaways for Investors
- Position for cost deflation, not just capability gains — falling inference prices will compress margins across the AI application layer.
- Underwrite agent-readiness as a core diligence item for any software holding; products without machine-accessible interfaces face structural risk.
- Watch OpenAI’s revenue-sharing ecosystem as a potential redistribution of platform economics away from traditional app stores.
- Treat decentralized compute and agent-payment networks as early-stage, narrative-driven exposure with high volatility.
- Monitor power, grid and commodity markets as the physical bottleneck of the AI buildout.
The through-line of Sottiaux’s message is that the AI trade is transitioning from a model-capability story to an infrastructure-and-distribution story. Investors who price in only the first phase may find themselves on the wrong side of the reset.
