The Setup

SPGI reported Q4 2025 earnings on February 10, 2026 — the exact day of the AI disruption panic selloff. Stock closed -9.7% at RSI 10.4 (extreme oversold). The irony: management spent the call explaining in specific detail why they're an AI beneficiary, not a victim.

And they're backing it up. CFO Eric Aboaf announced $1B in buybacks for Q1 2026 (vs $650M prior year Q1), explicitly citing "stock price levels that we're seeing." Management is buying the panic.

What the Call Actually Said

1. SPGI Is the Data Layer for AI Platforms

CEO Martina Cheung directly addressed AI fears:

"We see AI really is a net tailwind for the business."

SPGI is platform-agnostic data substrate for:

  • Claude financial services (Anthropic)
  • MCP connector for OpenAI (announced Feb 9)
  • Gemini Enterprise (Google, Dec 2025)

Key moat defense:

"We don't allow the LLM providers to train on S&P Global data."

Commercial licensing model preserved. Hundreds of distribution partners — LLMs are just another channel.

What customers are saying:

"We prefer to work with you guys, we want you to put the functionality into your tools, we want to use single pipe to get our data."

Customers want fewer vendors, not more. SPGI becomes the consolidated data layer.

2. AI = Internal Productivity Lever, Not Threat

$100M EDO savings target by end-2027 (20% reduction on ≈$500M expense base):

  • Automated >50% of data workflows
  • Eliminated >10% of applications
  • Software development tools saving $10M+/year
  • "Peak headcount" reached in some areas

CEO: "Revenue growth will outstrip headcount growth" for next several years.

This isn't cost-cutting under pressure. This is operating leverage from being an AI productivity winner.

3. Workflow Moat Defense

Cheung addressed "can LLMs replace workflows?" head-on:

"Workflows at S&P Global are critical systems of record for customers. iLEVEL, ClearPar, Cap IQ Pro, Platts Connect — not simple apps you can replace rapidly. In fact, they get smarter as we embed AI technology in them."

Key attributes that create moat:

  • Compliance-embedded: DORA regulation in financial services
  • Data integration: workflows only work with SPGI's proprietary data sets (loan reference data for WSO)
  • Network effects: connectivity to industry networks, enable capital flows, trading, reporting
  • Mission-critical: systems customers rely on for regulated functions

LLMs don't replace regulated, data-integrated, network-connected systems of record. They make them better.

4. 2026 Guidance: Solid Execution

EPS: $19.40-$19.65 (9-10% YoY growth)

  • Organic constant currency revenue: 6-8%
  • Margin expansion: 50-75bp

Every division beat original 2025 guidance:

  • Revenue growth at/above high end
  • Margins at/above high end
  • Q4 subscription ACV accelerating to 6.5-7% in MI (up from 6-6.5% H1)

Division guidance for 2026:

  • Market Intelligence: 5.5-7% (subscription top half of range)
  • Ratings: 4-7% (assumes Billed Issuance +low-to-mid single digits)
  • Energy: 5.5-7%
  • Indices: 10-12%
  • Mobility: 7.5-9%

No emergency cuts. No guidance pulls. Business executing.

5. Mobility Spin Q2 2026 = Repricing Catalyst

Form 10 filed confidentially with SEC. Q2 2026 milestones:

  • Public Form 10 filing
  • Investor Day
  • Equity roadshow
  • Public debt offering (targeting investment-grade)

Mobility growing 7.5-9% organic with 35.4% margins, CARFAX subscription base.

Separation unlocks pure-play repricing of RemainCo at higher multiples (financial data/ratings) and Mobility (auto data).

The Bear Case

ICE flat (-2.8% same day) while SPGI crushed (-9.7%) suggests market IS discriminating:

  • Market may be worried about Market Intelligence (32.6% of revenue) vulnerability vs ICE's exchange moats

FactSet (FDS) weakness is bearish read-through:

  • FDS at low multiples signals Market Intelligence pressure
  • MI is SPGI's largest segment, most exposed to "AI替代" fears

Valuation not obviously cheap:

  • 32x P/E at $401
  • If growth decelerates, multiple compression risk

Segment exposure matters:

  • Ratings + Indices (fortress moats): 41% of revenue
  • Market Intelligence (competitive, FDS comp weak): 33% of revenue
  • Energy/Mobility: 26% of revenue

The market may be rationally pricing MI risk, not irrationally panicking across all segments.

What Needs Validation

  1. SOTP valuation by segment — what's MI actually worth vs Ratings/Indices fortress?
  2. FDS competitive read — is MI vulnerability real or overdone?
  3. ICE anomaly explanation — why flat when SPGI crushed?
  4. Alpha calculation — what's the mispricing edge after segment-level analysis?

The Signal

RSI 10.4 on a quality compounder with:

  • Management accelerating buybacks citing price
  • Earnings call directly contradicting AI victim narrative
  • Fortress moats in 41% of revenue (Ratings + Indices)
  • AI driving internal cost savings ($100M by 2027)
  • Catalyst calendar: Mobility spin Q2 2026

The conviction gap:

  • Market: "AI disrupts data companies"
  • Management (with customer conversations): "AI is net tailwind, customers want more from us"

One of them is wrong. Management has customer relationships. Market has momentum.

The timing doesn't wait. RSI 10 quality compounders with management buying + catalyst calendar don't sit around for perfect analysis.

IC needs to decide: resource the deeper work (SOTP, FDS comp, alpha calc) or pass on the setup while it's still dislocated.