Market Intelligence

The GPU Market Intelligence platform tracks 2,700+ discrete entities across the value chain: GPU manufacturers, systems integrators, neocloud operators, data centre providers, enterprise customers, adjacent infrastructure players. The database ingests signals from 340+ data sources: job postings, regulatory filings, procurement announcements, patent filings, earnings transcripts, web property analysis.

Company cohort scoring applies 28 discrete metrics across four domains: Capital deployment velocity (capex, balance sheet, financing runway); Customer concentration (top-10 revenue %, churn cohort, contract terms); Operational metrics (fleet size, utilisation, CAC); Strategic positioning (IP ownership, partnership depth, geographic diversification). This yields weekly-updated company profiles enabling rapid competitive intelligence. Customers routinely discover partner/competitor announcements 2-4 weeks before public disclosure.

Real-time pricing capture across 23 major cloud providers updates hourly, tracking on-demand and spot rates for H100, H200, B100, and GB200 accelerators across regional markets. Current snapshot (March 2026): H100 on-demand ranges $2.08-$2.94/hour; Blackwell on-demand $4.80-$7.14/hour reflecting supply scarcity. Spot rates show 58-72% discount to on-demand depending on utilisation volatility and regional supply surplus.

Weekly pricing trends isolate three distinct dynamics: Structural pricing (platform features, customer cohort, support tier) showing 8-18% variance; Supply-driven pricing showing 25-40% swings as new capacity comes online; Demand-driven pricing compressed to 2-6% range as platforms limit variance to retain customers. Contract pricing analysis (30-day, 90-day, annual commitment terms) reveals 12-24% discounts for committed workloads, aligning with operator financing needs. The platform provides real-time pricing intelligence enabling customers to identify arbitrage opportunities and optimal commitment strategies.

GPU cloud market structure analysis maps the operator landscape: Hyperscaler-owned platforms (AWS, Google, Microsoft) control 42-48% of installed capacity; Enterprise-focused operators (Lambda Labs, Crusoe Energy) command 8-12% market share at 15-28% pricing premium; Price-competitive platforms (Vast, Paperspace) drive 15-20% market share through aggressive underpricing; Vertical-integrated operators (research labs, financial firms) serve 12-18% proprietary workloads; Capacity brokers and aggregators hold 8-14% share. Competitive intensity shows pronounced bifurcation: premium tiers (enterprise support, managed services) show stable 12-18% annual churn and stable pricing; commodity tiers show 45-52% annual churn and -8-12% annual pricing decay.

Market leadership emerges from two mechanics: network effects (customers value platform stability and ecosystem breadth) and lock-in via integration (custom orchestration, data pipeline integration).

Weekly analysis synthesis (delivered every Monday, 4 PM UTC) identifies market inflection points, emerging competitors, consolidation signals, and pricing trends. Recent syntheses identified: Spot market volatility expansion (72-hour pricing windows moving from 2-6% variance to 15-22%) signalling supply-demand mismatches; Blackwell constraint emergence as demand outstripped available inventory by 140-200%; Customer defection patterns from mid-tier operators toward hyperscalers despite 15-28% premium; Operator capex pause across smaller players reflecting financing constraints.

Subscriber base includes: NVIDIA strategic planning and product teams (pricing strategy); CoreWeave and Lambda Labs operational and financial teams; Brookfield Data Centre group (capex validation); multiple private equity sponsors evaluating GPU infrastructure; large enterprise AI teams conducting make-vs-buy analysis. Subscribers report platform utilisation for: competitive positioning, capex validation, contract negotiation, and quarterly board reporting on cloud infrastructure economics.

2,700+ entity database tracking capital deployment, customer concentration, utilisation, and strategic positioning across 340+ signal sources

On-demand pricing: H100 $2.08-$2.94/hour, Blackwell $4.80-$7.50/hr; annual commitment discounts of 12-24% available (pricing data current as of Q1 2026;verify current rates with providers directly)

Market structure bifurcated: hyperscalers 42-48% share but stable churn, price-competitive platforms 15-20% share driving -8-12% annual pricing decay

Weekly synthesis identifies inflection points; subscribers include NVIDIA, CoreWeave, Brookfield, and enterprise AI teams for strategic and capex validation