Financial data firms such as Bloomberg and FactSet have largely eased fears of being replaced by generative AI, but as AI agents consume far more data than human analysts, the question of who pays for the tokens has emerged as a new focus for investors.
AI Tokenomics · Financial Data Industry
Wall Street Data Giants Survived AI — Now the Fight Is Over Who Pays for the Tokens
Bloomberg, FactSet, LSEG and S&P are making their data agent-ready instead of being replaced by it. But as AI agents consume many times more data than humans, ballooning token bills are becoming the new competitive battleground.
$32K
Annual cost of a single Bloomberg Terminal
800+
Research providers reachable via Bloomberg's ASKB agent network
20–30%
Cost cut achieved with chargeback / showback (OpenText)
Agents Eat Far More Data Than Humans
Tokens consumed per user, per year
356M
Advanced
≈ 38× more tokens — usage can top 1 trillion tokens in half a year at 8–10% monthly growth
From "Will AI replace us?" → "Who pays for the tokens?"
1 · USER FIRM
Runs agents that retrieve, analyze & write reports
→
2 · DATA PROVIDER
Serves agent-ready data via MCP (FactSet+Gemini, Bloomberg ASKB)
→
3 · AI PLATFORM
Bills every input & output token — who absorbs it?
Data Firms' Stance
Embrace agent integration; differentiate on reliability, auditability and standardization . FactSet stresses interoperability; smaller players turn XBRL pipelines into MCP servers for traceability.
Developers' Warning
Uncontrolled agent usage is "like handing your end users a credit card." One case reported $6M+ in unplanned annual token costs, with mounting support burdens.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…