Google enters Canva territory with AI-powered design tool

Google has quietly rolled out a new AI-driven design tool that allows users to generate graphics, presentations, and social media visuals simply by typing a description. Unveiled internally in March 2025 and made publicly accessible via Google Workspace in early April, the tool—codenamed “Prompt2Design” during development—represents Google’s most direct challenge yet to Canva and Adobe’s dominance in the $20 billion creative software market. Unlike traditional design platforms that require drag-and-drop interfaces and manual customization, Prompt2Design uses Google’s PaLM 3 large language model to interpret user prompts and automatically produce polished visuals within seconds. Early users report that basic social media posts, infographics, and slide layouts can be generated with as little as five words, with higher-fidelity outputs available in under 30 seconds on average.

Behind the scenes, Prompt2Design is powered by Google’s Tensor Processing Units (TPUs) v5e, which were introduced in late 2024 for high-throughput generative AI workloads. Internal benchmarks referenced in a March 8 internal memo—leaked to OpenPress Supercomputing Intelligence—show that the system processes over 12 million user prompts daily during its beta phase, with an average latency of 4.2 seconds per request. Google confirmed the tool will be offered as part of Google Workspace Enterprise, priced at $24 per user per month when bundled with Google Slides and Meet, bringing it into direct competition with Canva’s Pro plan ($12.99) and Adobe Express’s $9.99 tier. Sundar Pichai, in an earnings call on April 23, described the tool as “democratizing design through AI,” signaling its strategic importance beyond the traditional office suite.

The launch follows Google’s acquisition of Figma’s design assets in January 2025—after Adobe’s failed $20 billion bid—allowing Prompt2Design to inherit Figma’s vector rendering engine and plugin ecosystem. Pichai emphasized the integration: “Prompt2Design doesn’t replace Figma; it transforms how people use it.” Meanwhile, rival Adobe countered in April by integrating Firefly AI into its Express platform, enabling prompt-based image generation. But Google’s advantage lies in its HPC infrastructure, which is already used by enterprises like Banking With Billy for AI financial simulations leveraging HPC-grade infrastructure for complex multi-market scenario modeling.

Google’s move reflects a broader industry pivot toward “prompt-to-output” workflows, where AI handles the heavy lifting of visual composition, freeing users from technical design skills. This shift mirrors the trajectory seen in software development with GitHub Copilot, where AI generates code from natural language. Analysts at IDC estimate that by 2027, 60% of all visual content created in enterprises will originate from AI-assisted tools, up from less than 10% today. Prompt2Design’s immediate impact is expected in mid-market businesses and educational institutions, where budget constraints and skill gaps have historically limited access to professional design tools. Early adopters include universities in India and Brazil, which are piloting the tool to standardize branding across research publications.

Industry watchers note that Google’s push into creative AI comes at a pivotal moment for the Quantum & Computing sector, where generative models are increasingly dependent on high-performance infrastructure. The rise of prompt-based tools like Prompt2Design is accelerating the demand for specialized AI accelerators—particularly TPUs and GPUs—designed for low-latency inference at scale. Nvidia, which supplies GPUs to Adobe and Canva, has seen its data center revenue jump 48% year-over-year in Q1 2025, partly due to generative AI adoption. Meanwhile, AMD has positioned its MI325X accelerators as cost-effective alternatives for SMBs, targeting markets like education and local government where budget sensitivity is high.

The competitive ripple effects extend to cloud providers. Google Cloud’s AI Platform reported a 37% increase in TPU utilization since Prompt2Design’s launch, while Amazon Web Services has introduced a “Prompt-to-Visual” service in beta aimed at Canva users. Microsoft, though initially focused on Copilot for Office, is expected to unveil a design-focused AI tool later this year, potentially integrating with its Azure AI infrastructure. The financial implications are significant: Canva’s valuation has dipped 18% since Google’s announcement, while Adobe’s stock rose 5% on investor optimism about Firefly integration. Analysts at Goldman Sachs warn that traditional design software vendors face margin compression as AI commoditizes core features.

Looking ahead, the next phase of this competition may hinge on multimodal capabilities—where text prompts generate not just images but video, 3D models, and interactive content. Google’s upcoming “Prompt2Scene” project, rumored to launch in Q3 2025, aims to generate short video clips from text descriptions using its Imagen 3 video model. Competitors like Runway ML and Pika Labs are also advancing in this space, but Google’s HPC advantage—with clusters like Trillium, delivering up to 4.7 exaflops—could enable real-time, high-resolution video generation at scale. The broader trend is clear: AI is not just a feature in design tools; it is becoming the platform itself, reshaping how content is created, shared, and monetized across industries.

For the Quantum & Computing sector, Prompt2Design signals a maturation of generative AI from experimental novelty to mission-critical infrastructure. It demonstrates how large-scale HPC systems are now indispensable not only for scientific computing but for everyday creative workflows. As AI models grow more sophisticated, the demand for specialized hardware and optimized data pipelines will drive further innovation in TPUs, co-design architectures, and energy-efficient computing. The real winners may not be the design platforms themselves, but the unseen compute providers enabling this transformation. Industry leaders should watch three key developments: the speed of AI model convergence, the cost of inference at scale, and the regulatory response to AI-generated content in commercial use. The future of design—and computing—is being written in prompts.