Observed Signal · May 25, 2026 · Analysis · Source: Exponential View · Impact: 3/5 · Sentiment: Negative
AI Bills Rise Despite Falling Token Prices
This Exponential View analysis explains why AI expenses are difficult to forecast and continue rising even as per-token prices fall. Quarterly token throughput has exploded (estimated ~17,000× over four years) while token unit costs have collapsed, producing highly elastic demand. Cheaper tokens have made multi-step AI agents economically viable; agents perform dozens of tool calls and repeated context reads that create large hidden multipliers on token consumption. The author cites examples—coding agents can re-read context each turn and produce up to ~55× token amplification over single-turn queries—and references survey and usage data suggesting active inference accounts for only ~15–20% of total token use. China’s domestic demand and providers (notably ByteDance and Alibaba) are a major growth driver.
The piece highlights systemic cost and forecasting risks from agent-driven token consumption and hidden token multipliers—issues that affect budgeting, vendor choices, and operational design across businesses deploying LLMs and agentic systems.
Track ByteDance Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Estimated tokens processed per quarter have grown by around 17,000× over four years.
- Token prices have fallen substantially while overall consumption has grown, indicating highly elastic demand for machine intelligence.
- AI agents drive much higher token use: agents can make between 5 and 25 tool calls per task, each adding context and API cost.
- Token amplification example: a coding agent operating over 10 turns could re-read context and use up to 55× more tokens than a single-turn query.
- Survey/data cited indicate active inference may represent only 15–20% of total token consumption; the remainder is ‘invisible’ work (tool calls, context reads, retries).
Connected Companies & Entities
1 Entity mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
AI Token Costs Explode, Straining Engineering Budgets
Exponential View's Monday data brief examines rapidly rising token consumption — the variable cost unit for large language models — and its budgetary impact. The newsletter cites Uber CTO Praveen Neppalli Naga saying 5,000 Uber engineers exhausted the company's 2026 token budget in four months, and notes ServiceNow experienced similar overrun. Survey and market data show many organisations exceeded AI budgets in 2025 and enterprise AI spend is rising: nearly half of respondents report tech budgets up ~10%, while average monthly AI spend at large enterprises rose 36% to $85,000 year-over-year. The piece argues agentic AI adoption and diffusion of token usage are driving unpredictable costs, raising cost-management concerns for CFOs and prompting reassessments of tech budgets and finance controls.
Reading Your AI Token Bill and Managing Agent Costs
A June 2026 briefing argues that rising AI token bills mark a shift from AI as a purchased tool to AI as labor that companies must manage. Using Uber as an early concrete example, the piece notes that 95% of Uber engineers use AI monthly and an internal coding agent produces roughly 1,800 code changes per week. Uber reportedly exhausted its 2026 AI budget months early, and company leaders say token usage and commits are not yet clearly linked to customer-facing feature improvements. The author outlines a seven-part argument covering the AI cost curve, a routing rule called "minimum effective intelligence," why 2025 budgeting models break, and an operating model to replace blunt token caps with gates, permissions, and work objects.
AI Economy Shifts as Token Costs Bite
A developer essay by Hicham Douch (published 2026-05-01) argues the era of 'AI is almost free' is ending as providers move to token-based pricing and advanced capabilities become more expensive. The piece cites Anthropic removing Claude Code from a cheaper tier and GitHub Copilot moving from action‑based to token pricing as examples. It reports companies (including a claim about Uber) burning through AI budgets, and warns product teams to impose token budgets, use cheaper models for high-volume scaffolding, and treat AI calls like metered cloud compute. The author dubs the new phase the “tokenogen era,” where every AI call has explicit cost and product roadmaps must account for token economics.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.