AI 2027 Milestone Tracker

Narrative

The VC and analyst lane tells a story of massive capital deployment alongside a persistent and widening gap between headline AI adoption and verifiable enterprise value creation. On the funding side, CB Insights documents a record $200B+ in AI venture investment in 2025, with OpenAI, Anthropic, and xAI alone raising $86.3B — 38% of all AI funding. Crunchbase confirms Sequoia and a16z dominated unicorn formation, backing 21 and 20 new unicorn companies respectively, with AI-native firms comprising 25% of all 2025 unicorns. Revenue growth at the frontier labs has also outpaced the AI 2027 report's own predictions: OpenAI reached ~$20B annualised revenue (slightly ahead of the $18B AI 2027 forecast), while Anthropic grew from $1B to $5B ARR between late 2024 and July 2025. At the application layer, CB Insights documents concrete shipped milestones: Anysphere (Cursor) reached $500M ARR by June 2025, Anthropic's Claude Code hit $400M ARR in five months, and GitHub Copilot leads coding AI with ~$800M ARR. Gartner's 2025 Hype Cycle positions AI agents at the Peak of Inflated Expectations, projecting 33% of enterprise software will incorporate agentic AI by 2028 (up from <1% in 2024), while Forrester warns 75% of firms will fail at building advanced agentic architectures independently.

However, every major analyst house — McKinsey, Bain, Gartner, and the a16z enterprise survey — converges on the same structural finding: adoption is wide but value creation is narrow. McKinsey's 1,993-respondent global survey (published November 2025) finds 88% of organisations use AI in at least one function, yet only 39% report any enterprise-level EBIT impact, and only ~6% of respondents qualify as genuine 'AI high performers' (>5% of EBIT from AI). Bain's Technology Report 2025 documents that tech-forward AI leaders have achieved 10–25% EBITDA gains, but most organisations remain stuck in experimentation. Sequoia's December 2025 '2026 outlook' explicitly forecasts AGI timeline delays and data-centre construction slippage, noting the end revenue from AI 'remains limited (on the order of tens of billions per year) relative to the scale of data center and energy investments (on the order of trillions over the coming five years).' The AI Futures Blog's direct scorecard of AI 2027's 2025 predictions (February 2026) finds that while revenue is roughly on pace, AI software R&D uplift is 'behind pace' and no leading AI company has conducted a substantially larger training run than GPT-4.5, suggesting compute scaling is not accelerating as the report assumed. These findings collectively support the 'Fant-AI-sia' critique that regulatory, adoption, and infrastructure frictions are real and consequential, that enterprise transformation is far from uniform, and that the gap between tool deployment and organisational rewiring is the dominant constraint on realising AI 2027-scale outcomes.