Graph AI, a two-year-old software company working to modernize how pharmaceutical companies track drug side effects, announced Tuesday it has raised $13.3 million in a Series A round led by Insight Partners. Bessemer Venture Partners, which led the startup's seed round, joined as an investor again.
The Pleasanton, California-based company wouldn't say how much it's now worth. Including the earlier seed investment, Graph AI has pulled in $16.3 million since its founding—a modest sum in venture terms, but one the company says is enough to challenge a market long dominated by Oracle and other legacy software giants.
The pitch centers on speed and cost. Graph AI's platform, the startup claims, can process adverse event reports in under 10 minutes, down from more than three hours with traditional systems. Operating costs, the company says, can drop by as much as two-thirds. Those are striking promises in an industry where regulatory missteps can derail approvals or spark recalls.
Pharmacovigilance—the somewhat ungainly term for monitoring drug safety after products reach the market—is a global business estimated at $8 billion to $9 billion. It's also one that still leans heavily on fragmented software, manual data entry, and handoffs between teams. An adverse event report might arrive as a PDF, an email, or even a phone call. Someone has to read it, code it, validate it, and submit it to regulators. The workflow hasn't changed much in 20 years.
Graph AI built what it calls an AI-native operating system to handle the entire process on a single platform. The company's initial product, called /intake, ingests reports from email, PDFs, XML files, APIs, and medical literature. It extracts patient information, drug details, event descriptions, and timelines, then applies MedDRA classification—the medical dictionary regulators require. A second module, /nucleus, manages case processing and submission. A third, /report, is slated for release in September for aggregate reporting, though the company offered few specifics on timing or features.
The startup declined to name customers but said it's working with pharmaceutical and biotech companies in North America and elsewhere. In June, the platform became available through Google Cloud Marketplace, following Graph AI's participation in Google's ISV Startup Springboard program.
Four founders launched the company: CEO Raghavendra Parvataraju, CTO Vijay Ponukumati, Chief Product Officer Mohan Konyala, and CFO and COO Ashutosh Bordekar. According to Bessemer, the team brings experience from LTI Mindtree, Infosys, ServiceNow, Google, and Cisco. LinkedIn lists 18 employees, though the company's official profile suggests a broader range of 11 to 50. Graph AI also operates an office in Hyderabad, India.
The new capital will fund expansion in the United States and Europe. The startup has been building its presence at industry conferences, including events in Philadelphia and Boston, though precise attendance details weren't provided.
A fourth module, /signal, is in development for predictive safety surveillance. The company said it's working with unnamed design partners on the product but offered no timeline for launch.
Graph AI positions itself squarely against incumbents like Oracle Argus, which held close to 60 percent of the pharmacovigilance software market as of 2025, according to an IDC MarketScape analysis. Other competitors include ArisGlobal LifeSphere Safety, Veeva Vault Safety, Ennov, AB Cube, and IQVIA's Vigilance Platform.
"Pharma companies must strive to get patient safety exactly right," Richard Matus, a principal at Insight Partners, said in a statement. Companies "haven't kept up with what AI can now do."
Nithin Kaimal, a partner and India chief operating officer at Bessemer Venture Partners, added that "patient safety should run on an intelligent, integrated platform—not fragmented tools and manual handoffs."
The regulatory landscape is shifting beneath Graph AI's feet. The FDA released draft guidance earlier this year on AI credibility frameworks. The EU AI Act now classifies certain medical AI systems as high-risk, triggering stricter requirements. And an industry working group published guidance on using AI in pharmacovigilance late last year. Graph AI says its system was designed with these frameworks in mind, incorporating audit trails, traceability, and human oversight.
"We are not replacing human judgment—we are building technology that allows experts to apply it where it matters most," Parvataraju said.
Whether the startup can convince an industry accustomed to Oracle's dominance remains an open question. Pharmacovigilance teams are risk-averse by nature—hardly surprising, given that a missed signal can mean patient harm. Switching software systems in this environment requires more than a demo and a pitch deck. It requires trust, regulatory comfort, and proof that the AI won't stumble when it counts.
For now, Graph AI is betting that the old systems are slow and expensive enough that companies will take the risk.
