Editorial illustration for 43% of Indian Enterprises Struggle to Control AI Costs: Splunk Report
43% of Indian Firms Can't Control AI Costs
43% of Indian Enterprises Struggle to Control AI Costs: Splunk Report
Nearly half of India's enterprises can't put a reliable number on what their AI systems actually cost to run. Cisco's Splunk State of Observability 2026 report, drawn from a survey of 2,050 IT operations and engineering professionals across 10 countries, found that 43% of Indian organisations struggle to control AI costs, while 46% can't effectively measure AI workload metrics at all. That's a notable gap given that 85% of Indian organisations say monitoring AI workloads is a priority.
The report points to a specific culprit: AI agents. As companies move from experimenting with AI to running agents at scale, the systems meant to track what those agents are doing, whether they're working correctly, and whether they've been compromised, are struggling to keep pace. Billing spikes are proving more disruptive than steady high costs for many organisations, and telemetry itself is becoming expensive enough to rival the infrastructure it's supposed to watch.
The findings suggest a widening mismatch between how fast agentic AI is being deployed and how prepared enterprises are to watch over it, a tension Cisco's leadership frames as a question of security and visibility becoming impossible to separate.
While 43% of Indian respondents said they could not accurately control AI costs, the global figure stood at 55%. Unexpected billing spikes caused greater operational strain than high baseline costs for 54% of Indian respondents, compared with 60% globally.
Why this matters
The numbers here are a warning light, not a footnote. Indian enterprises are racing to deploy AI agents while 43% admit they can't control what that spending looks like, and 46% can't even measure the workloads they're running. That's not a monitoring gap, it's a blind spot at the exact moment agentic systems start making decisions at machine speed, as Splunk's own framing puts it.
For founders building agentic products, this should read as a market signal: observability tooling built for static applications doesn't translate cleanly to systems that act autonomously and compound errors quickly. The 39% of Indian respondents without clear business outcomes tied to AI monitoring, against 54% globally, suggests buyers are moving faster than their own measurement frameworks. Meanwhile 87% flagging hallucination visibility as important tells us customers know the risk even if their tooling hasn't caught up.
Developers shipping agents into enterprise environments should expect procurement conversations to shift toward cost attribution and outcome tracking, not just uptime. The gap between deployment enthusiasm and operational control is where the next wave of tooling, and probably a few expensive mistakes, will come from.
Related Reading
Open Source D&B rebuilds 642 million‑business database after AI agents hit limits
Open Source Box CEO Aaron Levie Says AI Will Drive "Biggest Cybersecurity Upgrade
Open Source Study: 6.7% of Deepfake Requests on Hugging Face Targeted Children
From Archives Google Cloud Ships llm-d 1.0 With Cache-Aware TPU and GPU Routing
From Archives Deepseek open-sources inference engine powering its fastest AI model
Common Questions Answered
What percentage of Indian enterprises struggle to control AI costs according to the Splunk report?
According to Cisco's Splunk State of Observability 2026 report, 43% of Indian organisations struggle to control AI costs. This figure is notably lower than the global average of 55%, indicating that Indian enterprises face relatively better cost control compared to their international counterparts.
How many Indian organisations cannot effectively measure AI workload metrics?
The Splunk report found that 46% of Indian organisations cannot effectively measure AI workload metrics. This represents a significant gap considering that 85% of Indian organisations say monitoring AI workloads is a priority, creating a disconnect between stated priorities and actual measurement capabilities.
What causes greater operational strain for Indian enterprises: unexpected billing spikes or high baseline costs?
Unexpected billing spikes cause greater operational strain than high baseline costs for 54% of Indian respondents, compared with 60% globally. This indicates that the unpredictability of AI costs is more problematic for Indian enterprises than consistently high expenses.
How many IT professionals were surveyed across how many countries in the Splunk State of Observability 2026 report?
The Splunk State of Observability 2026 report surveyed 2,050 IT operations and engineering professionals across 10 countries. This comprehensive survey provided the data foundation for understanding AI cost management challenges globally and specifically in India.
Why is the inability to measure AI workloads considered a blind spot for Indian enterprises deploying AI agents?
The inability to measure AI workloads is a blind spot because 43% of Indian enterprises cannot control AI costs while 46% cannot measure workloads, yet they are racing to deploy AI agents that make decisions at machine speed. This combination of poor visibility and rapid AI deployment creates significant operational and financial risks that enterprises cannot properly monitor or control.
Further Reading
- Papers with Code - Latest NLP Research - Papers with Code
- Hugging Face Daily Papers - Hugging Face
- ArXiv CS.CL (Computation and Language) - ArXiv
