Enterprise AI budgets are already stale as 2027 planning opens
sanjaykalra · x · 2026-08-04
Enterprise AI budgets are already out of date for 2027 planning
Sanjay Kalra argues that enterprise AI planning is getting overtaken by model progress and budget reality. He cites Gartner’s 2026 IT spending forecast at $6.37 trillion and says AI accounts for $2.59 trillion of that, with end-user spend on AI models and platforms rising from $39 billion to $64 billion.
He says Sam Altman’s claim that the next six months could outpace the prior two years of model progress means many 2027 budgets are already stale. His main planning priorities are:
- treat token usage as a variable operating cost with a named owner
- fund the context layer, since agents can only act on what they can read
- assign one accountable owner per agent, with P&L responsibility
- insist on model portability so providers can be swapped without rewriting integrations
Kalra also says Gartner still expects more than 40% of agentic AI projects to be canceled by 2027, while Forrester’s 2027 planning guide found 82% of tech leaders expect bigger budgets next year. His broader point: the winners in 2027 won’t necessarily be the biggest spenders, but the teams that manage AI like an operating discipline rather than a demo budget.
Related event: Gartner Forecasts Global IT Spending to Reach $6.37 Trillion in 2026(2 posts)→
More from Companies & People
- Interconnects launches Artifacts Hub to track key open AI models — natolambert · 2026-08-04
- Korea’s government-run LLM race has finished training, with unveilings due in early August — teortaxesTex · 2026-08-04
- Circleback says its July 2026 update shipped too much for a single email — ycombinator · 2026-08-04
- OpenAI names Codex winners and awards each team $10,000 — Gradio · 2026-08-04
- trydecideAI says it has finished training Dax-1 for deterministic spreadsheet editing — saheedniyi_02 · 2026-08-04
- OpenAI's First Influencer Brand Trip Sparks Backlash Amid AI Tensions — TechCrunch AI · 2026-08-04