9 public cloud companies where AI costs are outpacing revenue and compressing margins
Across 9 public cloud companies, AI-driven cost of revenue is growing faster than revenue itself — with divergence of +300 to +3,700 bps and early margin compression. In several cases, cost growth is 2–5x revenue growth.
The AI Cost-Revenue Divergence (ACRD) detection framework compares revenue growth, cost-of-revenue growth, and gross-margin change to identify early signs of AI workload margin pressure.
ACRD trailing signal from past 10-Q: Active Cost of revenue growth exceeded revenue growth by a wide margin (+1,090 bps spread), producing a meaningful gross-margin decline (−262 bps).
ACRD current Signal from most recent earnings call: Active Management stated that gross margin pressure was driven primarily by higher premium messaging volume and hosting costs associated with expanded AI-driven customer engagement workloads.
The Bottom Line: Cost-of-revenue growth substantially exceeded revenue growth, compressing gross margin by 262 bps in the quarter. Management explicitly ties higher infrastructure and messaging costs to increased usage of AI-driven engagement capabilities; classification: AI-Linked.
C3.ai
-29.1%
18.4%
+4,750 bps
−2,710 bps
Active
ACRD trailing signal from past 10-Q: Active Cost-to-serve increased while revenue declined, and gross margin compressed materially year over year.
ACRD current Signal from most recent earnings call: Active Quarterly results show further revenue contraction and gross margin collapse, with cost growth outpacing revenue decline; management commentary centers on restructuring and cost reduction to stabilize operations.
The Bottom Line: Financial divergence is present and accelerating; however, management disclosures indicate restructuring and revenue contraction as primary drivers rather than rising AI inference costs; classification: AI-Linked
Cloudflare
30.7%
52.5%
+2,179 bps
-372 bps
Active
ACRD trailing signal from past 10-Q: Active Cost growth materially exceeded revenue growth by more than 500 basis points and gross margin declined year over year, confirming financial divergence.
ACRD current Signal from most recent earnings call: Active Management highlighted accelerating AI-driven traffic, increased network cost allocations, and higher CapEx intensity tied to AI and pool-of-funds variability, reinforcing visible cost-revenue slope imbalance.
The Bottom Line: Confirmed financial divergence in the most recent quarter, with cost growth significantly outpacing revenue growth and clear gross margin compression. AI agent-driven workload expansion is increasing network cost intensity; classification: AI-Linked
Fastly
7.8%
14.6%
+680 bps
−220 bps
Active
ACRD trailing signal from past 10-Q: Active Cost-to-serve is growing materially faster than revenue, and gross margin declined year over year.
ACRD current Signal from most recent earnings call: Active Quarterly results show continued cost growth outpacing revenue growth, with margin compression visible; management disclosures reference ongoing platform investment and infrastructure intensity without unit cost reconciliation.
The Bottom Line: Financial divergence is present, with cost growth outpacing revenue growth and margin compression visible. Management has not explicitly attributed this to AI workloads; however, as an edge compute provider, Fastly is structurally exposed to rising compute intensity. AI-exposed
GitLab
24.6%
45.6%
+2,099 bps
−190 bps
Active
ACRD trailing signal from past 10-Q: Active Cost growth materially exceeded revenue growth by more than 500 basis points and gross margin declined year over year, confirming financial divergence.
ACRD current Signal from most recent earnings call: Active Management emphasized increased investment in AI capabilities and platform expansion, while margins reflected elevated cost structure relative to revenue growth, reinforcing visible slope imbalance.
The Bottom Line: Confirmed financial divergence in the most recent quarter, with cost growth significantly outpacing revenue growth and clear gross margin compression. AI initiatives are strategically central but not yet demonstrating operating leverage; Classification: AI-Exposed
Oracle
9.0%
14.2%
+520 bps
−240 bps
Active
ACRD trailing signal from past 10-Q: Active Cost of revenue growth exceeded revenue growth by roughly 520 bps, producing a meaningful gross-margin decline (−240 bps).
ACRD current Signal from most recent earnings call: Active Management repeatedly attributed accelerating infrastructure demand to AI workloads, including AI infrastructure revenue growing 243% year over year and multicloud database revenue expanding 531%.
The Bottom Line: Cost-of-revenue growth substantially exceeded revenue growth, compressing gross margin by roughly 240 bps in the quarter. Management directly attributes surging infrastructure demand to AI infrastructure and multicloud database workloads; classification: AI-Linked.
Sprinklr
9.2%
27.2%
+1,800 bps
−490 bps
Active
ACRD trailing signal from past 10-Q: Active Revenue growth of roughly 9% was substantially outpaced by 27% growth in cost of revenue, producing a spread of about +1,800 bps and a material gross-margin decline in the quarter. The divergence indicates meaningful pressure at the unit-economics level.
ACRD current Signal from most recent earnings call: Active Management continues to position the platform as AI-native unified customer experience management, emphasizing AI-driven automation across marketing, service, and contact center workflows. However, there was no explicit reconciliation of AI-related workload costs versus revenue contribution, leaving the cost-intensity implied but not quantified.
The Bottom Line: Costs are rising materially faster than revenue, producing clear gross-margin compression in the quarter. While the company emphasizes an AI-native CXM platform and AI-driven automation, management does not reconcile AI workload intensity with revenue gains; classification: AI-Linked with active cost pressure.
Varonis
9.1%
46.4%
+3,730 bps
−560 bps
Active
ACRD trailing signal from past 10-Q: Active Cost of revenue grew dramatically faster than revenue (46% vs 9%), producing a 3.7-point decline in gross margin, indicating clear margin pressure in the quarter.
ACRD current Signal from most recent earnings call: Active Management emphasized continued transition toward SaaS delivery of the Varonis Data Security Platform and increased cloud-hosted infrastructure supporting the SaaS model, implying structurally higher service delivery costs relative to the historical on-premise model.
The Bottom Line: Cost-of-revenue growth far outpaced revenue growth, producing a meaningful gross-margin decline in the quarter as the company continues migrating customers to its SaaS platform. The cost structure reflects the economics of delivering hosted security analytics and monitoring infrastructure rather than traditional license software; classification: AI-Linked.
Wix
14.0%
15.7%
+170 bps
−100 bps
Active
ACRD trailing signal from past 10-Q: Emerging Cost of revenue increased modestly faster than revenue (+170 bps spread), producing a measurable but not severe gross-margin compression (−100 bps).
ACRD current Signal from most recent earnings call: Active Management explicitly attributed the margin decline to AI compute and inference costs associated with Base44 growth and AI product scaling, indicating elevated workload intensity.
The Bottom Line: Cost-of-revenue growth outpaced revenue growth and produced measurable gross-margin compression in the quarter. Management directly attributes margin pressure to AI inference and compute investment supporting Base44 growth; classification: AI-Linked.
ACRD shows costs rising faster than revenue. Who wins AI? The first to cut its cost
Source: RevenueModel.ai
Signals are classified by cost-revenue divergence measured in basis points:
Positive — revenue growth exceeds cost growth Neutral — within ±300 bps Emerging — 300–500 bps divergence with margin compression Active — >500 bps divergence with visible gross-margin pressure
Financial data derived from the most recent company 10-Q filings and corresponding earnings-call transcripts. The dataset is continuously updated as new filings appear.
