Google's Antigravity, an AI coding assistant that consolidates the company's fragmented coding tools, is facing significant security scrutiny just as it attempts to compete with Claude Code and other rivals. Security researchers at Mindgard, an AI security firm, have publicly disclosed a trusted workspace vulnerability in Antigravity, marking one of over 150 high-impact security issues the firm has found across major AI products.

What Security Vulnerabilities Does Antigravity Have?

Mindgard, a Boston and London-based AI security company spun out of Lancaster University research, identified a trusted workspace flaw in Google Antigravity as part of its broader work testing and protecting AI systems against attacks. This vulnerability joins a growing list of security concerns discovered in widely used AI coding tools, including zero-day code execution flaws in Cursor IDE and guardrail failures in ChatGPT's image generation capabilities.

The disclosure comes at a critical moment for Antigravity. According to internal reports, Google has been consolidating its competing AI coding tools into Antigravity to stop internal product cannibalization between DeepMind, Google Cloud, and Android teams. The tool represents Google's attempt to centralize its fragmented approach to AI coding assistance, but the security vulnerability raises questions about the maturity of this consolidation effort.

How Is Antigravity Performing in the Market?

Despite the security concerns, Antigravity maintains a modest but stable presence in the developer market. According to JetBrains' Developer Ecosystem Survey 2026, which gathered responses from more than 15,000 professional developers worldwide, Antigravity holds 6% adoption among developers using AI coding agents at work. However, the tool has experienced a significant leap in awareness, rising from 29% in January 2026 to 47% by May through July 2026.

India represents Antigravity's strongest market, where it is practically tied as the third most-popular AI coding tool alongside Cursor. In India, 15% of developers use Antigravity at work, up from 10% in January 2026. This regional strength suggests Google has found some traction in specific markets, even as global adoption remains limited compared to competitors.

The broader AI coding agent market shows intense competition. Claude Code has emerged as the dominant player, used by 39% of professional developers worldwide as of May through July 2026, up from just 18% in January. OpenAI's Codex has experienced rapid growth, jumping from 3% adoption to 16% in the same period. GitHub Copilot, which pioneered AI-assisted coding in 2023, has declined from 29% adoption a year prior to 21% in the latest survey.

Why Does Google's Internal Fragmentation Matter?

The security vulnerability in Antigravity reflects deeper organizational challenges at Google. Multiple teams within the company have been building overlapping AI coding tools, creating redundancy and slowing innovation. DeepMind, Google Cloud, and the Android team each spent the past year developing competing AI coding assistants before being forced to consolidate into Antigravity.

This fragmentation has real consequences. While competitors like OpenAI have maintained consistent product development across multiple interfaces, Google's approach has created confusion and diluted its market presence. The Gemini CLI and Code Assist tools were merged into Antigravity specifically to address internal cannibalization, suggesting the company struggled to coordinate its AI coding strategy.

Steps to Understanding AI Security in Enterprise Tools

• Vulnerability Disclosure: Security researchers now regularly test AI coding tools for flaws, with Mindgard alone disclosing over 150 high-impact vulnerabilities across products from OpenAI, Google, and Cursor, creating a new category of AI-specific security risks.
• Enterprise Adoption Concerns: Businesses are deploying AI tools deeper into day-to-day operations while facing greater scrutiny over security risks, with Fortune 2000 companies and AI-focused businesses in financial services, pharmaceuticals, gaming, and healthcare now demanding specialized AI security platforms.
• Market Consolidation Pressure: Companies like Google are consolidating fragmented internal tools to improve product coherence and security, though this process can expose vulnerabilities during integration and transition periods.

The security disclosure in Antigravity arrives as Mindgard raised $30 million in Series A funding to expand its AI security platform. The investment reflects growing corporate demand for specialized tools that can identify and defend against attacks targeting AI systems. Mindgard's platform is designed to identify shadow AI, conduct AI red teaming, and provide runtime protection, addressing what the company describes as an entirely new attack surface created by AI systems.

"AI is creating an entirely new attack surface and organisations need a fundamentally different approach to securing it," stated James Brear, Chief Executive Officer at Mindgard.

James Brear, Chief Executive Officer, Mindgard

For developers and enterprises evaluating AI coding tools, the Antigravity vulnerability underscores a broader trend: as AI coding agents become standard in developer workflows, security testing and disclosure are becoming critical differentiators. The fact that Antigravity's flaw was discovered and publicly disclosed suggests the security research community is actively scrutinizing these tools, which may ultimately improve their safety over time.

Google's challenge now extends beyond market share. The company must demonstrate that Antigravity, despite its security vulnerabilities, can evolve into a trustworthy enterprise tool while competing against Claude Code's momentum and OpenAI's rapid product development. The consolidation of Google's fragmented coding tools into Antigravity represents a strategic bet that centralization will improve both security and user experience, but the early security findings suggest this integration process is still maturing.