loading
loading
When Vineet Chirania was scaling his train-ticket booking startup, Trainman, a decade ago, he encountered a problem that would eventually inspire his next venture. His team relied on Datadog, an application monitoring tool, to track the performance of its software systems. Then one day, the startup received an unexpectedly high bill after one of its developers started feeding additional custom metrics into Datadog. Over at BharatPe, senior engineering executive Vijay Aggarwal was facing a similar challenge with another observability giant, New Relic. Both Aggarwal and Chirania quickly realised the problem. As apps grow and generate more telemetry data, monitoring costs become increasingly difficult to forecast. Add concerns around data residency and dashboard latency, and the pain points become hard to ignore. These experiences eventually led the duo to build CubeAPM, an observability platform that claims to reduce enterprise monitoring costs by as much as 60-80%, while keeping customer data within their own cloud environments. Today, the bootstrapped startup serves around 50 enterprise customers, including Delhivery, RedBus, PolicyBazaar, Shadowfax and the Ola Group. Simultaneously, CubeAPM claims to be churning an annual recurring revenue (ARR) of nearly $1.5 Mn, up roughly 3-4X over the past year, all while remaining profitable since inception. But CubeAPM's ambitions extend beyond being an affordable alternative to incumbents. The startup is betting that its architecture, designed around self-hosting, data residency and AI-assisted troubleshooting, can help it carve a place in the global observability market, which is estimated to become a $20 Bn opportunity by 2031. From Trainman To CubeAPM The origins of CubeAPM are deeply rooted in the founders' operational experiences. Chirania spent much of his career building Trainman, which eventually scaled to more than 30 Mn app downloads before being acquired by the Adani Group. Aggarwal, meanwhile, built his reputation leading engineering teams at high-scale internet startups. Before founding CubeAPM, he served as the CTO of BharatPe and earlier headed engineering at Grofers (now Blinkit). Both hail from IIT Roorkee. After Chirania exited Trainman, the two reconnected and discovered that they had experienced the same frustrations with observability software. “The common thread was application performance monitoring. We had used different tools, but the problems were remarkably similar: unpredictable pricing, poor visibility into future costs, and challenges around data residency,” said Chirania. Initially, the founders believed modest cost savings would be sufficient to convince enterprises to switch platforms. But most observability purchases are displacement sales. Customers already have a monitoring platform in place, and migrating critical infrastructure monitoring systems requires massive engineering effort. Saving 15% or 20% wasn't compelling enough. "We realised customers were willing to migrate only if the economics were dramatically better, sending us back to the drawing board," Chirania said. Reinventing The Observability Architecture At its core, observability software helps engineering teams understand what is happening inside their apps. When a website slows down or a payment service crashes, observability platforms collect telemetry data, such as logs and metrics, to help teams identify the root cause and resolve issues quickly. However, scaling observability is a key challenge as enterprises generate massive volumes of telemetry data, making real-time storage, indexing and analysis increasingly expensive and resource-intensive. Moreover, traditional observability vendors such as Datadog and New Relic typically operate SaaS-based models, where customer data is sent to the vendor's cloud for processing and storage. CubeAPM takes a different route to reduce costs. Rather than moving customer data outside their environments, the startup installs its platform inside the customer's own cloud infrastructure. The approach allows data to remain within the customer's cloud environment and satisfies compliance requirements more easily than foreign SaaS platforms. According to the founders, this architecture addresses three key enterprise concerns simultaneously: cost, data residency and latency. The architecture also reduces data transfer costs that enterprises often incur when moving large volumes of information to external vendors. More importantly, the startup claims it can deliver significantly faster dashboard performance because the data remains geographically closer to customer workloads. This model, however, created a new challenge: how to manage and support a system running inside a customer's environment. To solve this, CubeAPM developed a framework that sends only minimal operational metadata to its own systems, allowing its engineers to monitor deployments remotely and provide support without accessing sensitive customer data. Tech Behind Cost Advantage Cost reduction is the centrepiece of CubeAPM's value proposition. The startup argues that traditional observability platforms often charge customers across multiple dimensions, including hosts, users, indexed data, custom metrics and support plans. As infrastructure scales, those variables compound and make bills difficult to predict. To solve this, CubeAPM prices its products primarily on data ingestion volume, bundling unlimited users, hosts, retention and support into a single pricing model. While this simplicity resonates with enterprises that struggle to forecast monitoring expenses, the real differentiator comes from its underlying infrastructure efficiency. According to Chirania, the startup has developed proprietary compression and storage technologies that dramatically reduce the amount of infrastructure required to store telemetry data. He claims CubeAPM can compress 100 GB of incoming data into roughly 4 GB of stored data, reducing storage, compute and memory requirements. This efficiency becomes important because customers host the platform within their cloud environments. The startup also emphasises compatibility with existing observability ecosystems. Rather than forcing customers to replace their existing stacks, CubeAPM works with popular telemetry agents and standards, including OpenTelemetry, Datadog, New Relic and Elastic. This interoperability allows migrations to happen relatively quickly, reducing one of the biggest barriers to adoption. “We wanted switching to feel like hours rather than weeks," Chirania said. Building A Full-Stack AI Observability Platform What began as an application performance monitoring tool has gradually expanded into a broader observability suite. CubeAPM now offers six major modules: Application Performance Monitoring (APM) Log Management Infrastructure Monitoring Real User Monitoring (RUM) Synthetic Monitoring Error Tracking The startup’s typical users are engineering, DevOps and site reliability engineering (SRE) teams. Most begin usage with APM before expanding into additional modules. This “land-and-expand” motion has become a major growth driver for CubeAPM. According to Chirania, roughly half the company's customers currently use multiple products within the CubeAPM suite. Now, as AI reshapes software development, CubeAPM sees observability becoming even more critical as AI-generated code may eventually create systems that developers understand less intuitively than the software they write themselves. As a result, identifying root causes could become more difficult, increasing reliance on observability platforms. To prepare for that shift, CubeAPM has begun integrating AI capabilities directly into its workflow. One of its recent launches is an MCP server that allows AI coding assistants, including Cursor, Claude Code, OpenAI Codex and Gemini, to access telemetry data directly from CubeAPM. This allows engineers to skip dashboards altogether, query logs through AI agents and troubleshoot issues through conversational root-cause analysis. The startup is also using AI internally. Over the next 12-24 months, AI-powered troubleshooting is expected to become one of CubeAPM's primary product priorities. Chasing The Enterprise Stack Unlike many enterprise software startups chasing growth through venture capital, CubeAPM has remained entirely bootstrapped. The startup says that its GTM strategy has largely relied on founder networks, particularly within India's tech ecosystem, where both founders have spent years building relationships. This approach has helped the startup, which operates with a team of 25 employees, acquire customers without large marketing budgets. Moreover, CubeAPM claims to report renewal rates above 95%, which management attributes to its support model. Rather than relying exclusively on ticketing systems, it creates direct communication channels with clients through Slack, WhatsApp, Microsoft Teams or Google Chat, allowing engineers to respond quickly during incidents. For now, the company remains focused on enterprise and mid-market customers. Ironically, despite being founded by startup veterans, CubeAPM has intentionally avoided targeting startups due to purely economic reasons. Migration efforts make the most sense for organisations already spending substantial amounts on observability software. For smaller startups, the savings generated by switching platforms may not justify the integration effort. This could change in the future, with CubeAPM now exploring a dedicated startup offering. CubeAPM says its platform now powers more than 1 Bn requests a month across its customer base of 50 clients. It currently claims an ARR of nearly $1.5 Mn while remaining profitable. The Road Ahead CubeAPM's biggest challenge may not be technology but market penetration. The observability category is dominated by well-capitalised global incumbents that have spent decades building products, ecosystems and distribution channels. Convincing enterprises to replace mission-critical infrastructure monitoring tools may not be easy. The startup's answer is to focus relentlessly on the areas where it believes incumbents remain vulnerable: predictable pricing, data residency, support quality and infrastructure efficiency. International expansion is also emerging as a priority. While most customers are currently based in India, CubeAPM sees significant opportunities in the US, Europe, Southeast Asia and the Middle East, particularly among enterprises with stringent compliance requirements. Despite investor interest in infrastructure software and AI, the founders say they are in no rush to raise capital. Their goal remains straightforward: build a globally trusted observability platform that helps enterprises monitor complex software systems at a fraction of the cost of traditional vendors.
document fetched 2026-09-21 02:47 · canonical https://inc42.com/startups/how-cubeapm-plans-to-take-on-datadog-and-new-relic-with-its-80-cheaper-ai-observability-stack