AI BATTLEGROUNDS · EDITORIAL ASSESSMENT
AI Power Rankings
RAW PEC's assessment of who holds power in each AI battleground, scored on Position, Scale and Momentum. First edition: three battlegrounds, as of October 24, 2026. Rankings are re-scored monthly.
Leader
Anthropic
5.00Why: Preparing an IPO; Broadcom agreed to finance up to $42B of a TPU lease; Claude runs inside Google's Gemini agent.
Google (Alphabet)
4.60Why: Gemini 4 Argon generally available; launched the Gemini agent; OpenAI's GPT-6 models offered on Google Cloud; Gemini coming to Oracle applications.
OpenAI
4.60Why: Launched GPT-6.1 Sol at one-fifth of GPT-6 Astra's price; GPT-6 models on Google Cloud; IBM built an OpenAI practice.
Contender
Meta
3.40Why: Released Muse Wonder, its first frontier model, and the Muse Code agent.
Microsoft
3.00Why: Completed pre-training of MAI-Base-2, a 2.8-trillion-parameter in-house model.
SpaceX (incl. xAI)
3.00Why: Released Grok 4.6 and Grok 4.7.
Challenger
DeepSeek
2.80Why: Released V4-Pro and V4.1-Flash; raising about $12B from Tencent and CATL.
Low confidenceMistral AI
2.60Why: Raised a EUR 3B Series D at over EUR 21B; released Mistral Large 4.
Moonshot AI
2.60Why: Raised at about a $50B valuation ahead of a planned Hong Kong IPO; released Kimi K3.1 Pro.
Watch
Safe Superintelligence
1.80Why: NVIDIA alliance to raise compute about tenfold; no product released.
Leader
NVIDIA
5.00Why: Agreed to acquire Hugging Face for about $11.9B; compute alliance with Safe Superintelligence; $1B commitment to Nscale.
TSMC
4.20Why: Approved about $29.4B for new capacity; Q3 revenue up 41%; manufactures Microsoft's Maia 300.
Contender
AMD
3.40Why: Agreed to acquire World Labs for about $8.2B and Taalas for inference.
Broadcom
3.40Why: AI chip revenue of $16.7B with Q4 guidance up 236%; agreed to finance up to $42B of Anthropic's TPU lease.
Google (TPU)
3.40Why: Committed EUR 13B to AI infrastructure in Finland; signed an 890 MW nuclear power deal.
Challenger
Microsoft (Maia)
2.60Why: Announced the Maia 300 inference chip for 2027; withdrew from a planned UK data center.
CoreWeave
2.60Why: Signed over $25B of new commitments; contracted power rose to 4.2 GW.
Nebius
2.60Why: Secured a $3.5B financing facility from Citi.
Nscale
2.60Why: Compute deal with Anthropic reported at $45B; $3.5B Figure deal; $3.36B pre-IPO financing.
Low confidenceModal
2.00Why: Raised a $750M Series D at $15.75B; run rate passed $1B.
Low confidence
Watch
Cerebras
1.60Why: No material moves since its May IPO.
Leader
No Leader: no company yet owns the cross-vendor view of AI cost.
Contender
ServiceNow
3.00Why: Launched AI Cost Management across model, SaaS and GPU costs; added value intelligence to its control tower; sells Flow standalone.
Google Cloud
3.00Why: Introductory Gemini 4 Argon pricing; per-project AI spend caps in the Gemini agent.
Stripe (incl. OpenRouter)
3.00Why: Agreed to acquire OpenRouter; launched Treasury for agents.
Challenger
Salesforce
2.80Why: Completed the Fin acquisition; new Agentforce editions priced 11-13% higher.
Microsoft
2.60Why: Cut internal Claude spending by about a third (reported by The Information).
- The Information
IBM (Apptio)
2.40Why: Launched Apptio AI Value & ROI, linking AI and token spend to outcomes.
Workday
2.20Why: AI drove over 25% of new contract value, sold through Flex Credits.
Watch
Snowflake
1.80Why: No cost-of-AI moves in the last 90 days.
RAW PEC advises companies in the ServiceNow ecosystem; this placement received a second review. See method.
THE FRAMEWORK
How we score
Each company is assessed on Position (40%), Scale (20%) and Momentum (40%), each scored from 1 to 5. Weighted scores determine tier placement; thin or single-source evidence is marked Low confidence and cannot rank above Challenger.