The termination shock: Where AI progress meets reality

At Progress Conference 2025 speakers from the AI frontier — including OpenAI’s Sam Altman — affirmed that transformative AI is technically within reach, but framed the real barrier as institutional friction: U.S. government AI use may be below 10% even as broader workplace adoption has roughly doubled to ~40%. Panelists argued that without either disruptive bypasses of legacy providers (examples cited include Boom Supersonic’s vertical manufacturing and rideshare’s displacement of taxi commissions) or deliberate, difficult institutional reform, technical progress will struggle to produce broad economic and social impact. For investors, the implication is clear: value realization will hinge on regulatory, procurement and organizational change, so opportunities lie with fast-adopting private players, regulatory arbitrage and efforts to modernize public-sector capabilities rather than on model capability alone.

Analysis

Progress Conference 2025 surfaced a clear bifurcation: technical AI capability is advancing rapidly — highlighted by frontier researchers and remarks from OpenAI CEO Sam Altman during the “AI Protopia” sessions — yet the conference emphasized that institutional and societal frictions will determine whether those capabilities produce broad economic or social impact. Multiple speakers framed the problem as a “termination shock” where fast-moving tech encounters slow-moving governments, legacy firms, and regulatory systems.

Evidence presented at the conference quantifies this gap: Gallup-like workplace adoption roughly doubled to ~40% over the last year while a senior policy advisor estimated U.S. government AI usage at “just short of double digits,” implying government adoption below 10%. Panelists illustrated two deployment paths — circumvention of legacy institutions (Boom Supersonic’s vertically integrated manufacturing, rideshare displacing taxi commissions) or difficult institutional reform — and warned that each path carries trade-offs between speed and public-good preservation.

For investors, the central takeaway is that value realization will be driven by deployment and adoption, not model capability alone. Opportunities are likely concentrated in AI-first competitors, companies that vertically integrate to speed time-to-market, and vendors that materially accelerate public-sector procurement and operations; conversely, expect delayed monetization where institutional bottlenecks persist and monitor regulation as a key catalyst or headwind.

Key Decisions for Investors

• Consider selective overweight exposure to AI-first platform companies and service providers that can bypass or displace slow incumbents, as exemplified by rideshare’s impact on taxi commissions
• Allocate to vendors and systems integrators that enable government and large-enterprise AI adoption (procurement, workflow automation, compliance), since public-sector modernization is a key potential catalyst
• Avoid assuming near-term revenue realization from pure-play model capability alone; require demonstrable deployment and adoption KPIs before increasing stakes
• Monitor government AI adoption metrics (current estimate <10%) and regulatory/procurement milestones as primary catalysts or risk events, and rebalance positions around inflection points
• Favor firms with vertically integrated production or in-house deployment capability (the Boom example) that can materially shorten design-to-product cycles and reduce deployment friction