Adobe and Nvidia partnership builds next-generation Firefly AI models with digital twin workflows

Nvidia, Adobe bring AI-powered digital production to GTC - SiliconANGLE

Nvidia and Adobe team up to bring digital twins into the marketing machine Generative artificial intelligence is transforming content creation, but companies that make physical products need more than just images -- they need pixel-perfect brand fidelity at scale. That tension between creative speed and product accuracy is fueling a new era of AI-powered digital production in marketing. Adobe Inc. and Nvidia Corp. announced a strategic partnership at GTC 2026 to accelerate AI-powered creation, production and personalization, combining Adobe's creative and marketing workflows with Nvidia's open models, libraries and accelerated computing. A new cloud-native 3D digital twin solution enables brands to carry a product's digital identity from the design phase into campaign production, according to Rev Lebaredian (pictured, right), vice president of Omniverse and simulation technology at Nvidia. "We're taking these digital twins that we're using as part of the design cycle of these products and carrying those digital twins forward into the marketing and communications part of the same enterprise," Lebaredian told theCUBE. "It's really important that you use the actual digital twin because as great as AI is at creating content ... it's really important when you represent your product that what you're showing is exactly what it's going to be like when people interact with it in the physical world. You can't really promise that with a pure AI-generated image." Lebaredian and Varun Parmar (left), general manager at Adobe, spoke with theCUBE's John Furrier at the Nvidia GTC AI Conference & Expo, during an exclusive broadcast on theCUBE, SiliconANGLE Media's livestreaming studio. They discussed how AI-powered digital production is changing marketing workflows by helping brands scale content without sacrificing product identity. (* Disclosure below.) The Nvidia-Adobe partnership centers on a workflow that converts computer-aided design files into simulation-ready digital twins, routing them through Adobe's content automation platform. Instead of recreating product assets for each campaign, brands can use a consistent digital representation across design, marketing and commerce workflows, Parmar noted. "The need for content is going to go up by 5x over the next two years," Parmar said. "We know generative AI is the solution for content scaling. However, there are certain industries where identity preservation is really important, where you want to make sure that there are exact pixels of the product that are represented and there's no hallucination." That requirement points to the broader view of digital twins and Nvidia Omniverse as core simulation infrastructure for the physical world. In industries such as automotive, consumer electronics and luxury goods, digital twins are not just useful for marketing content -- they serve as persistent digital representations of products, helping companies maintain accuracy from design and manufacturing through brand and campaign execution, according to Lebaredian. "Being able to represent physical things inside a computer and do iterations on it ... is a superpower," Lebaredian said. "Without it, we actually don't really have a chance at creating the complex products that we're creating today and into the future." The payoff could be substantial. Campaign asset production that once took 28 to 30 days can now be done in minutes, Parmar emphasized. And because a digital twin exists before a physical product ships, brands can begin campaigns months earlier, extending AI-powered digital production from design through launch, Lebaredian added. "In the case of the automotive industry ... it can be a year or more from the point at which they know exactly what the car is supposed to be like, because the design is finalized and they're manufacturing it ... until they have a final production car out," Lebaredian said. "You can get a head start and be ready day zero." Stay tuned for the complete video interview, part of SiliconANGLE's and theCUBE's coverage of the Nvidia GTC AI Conference & Expo.

Next-gen Firefly models to be built on Nvidia's Cosmos open models

Adobe and Nvidia announced a strategic partnership on Monday to accelerate AI-powered creative and marketing workflows by integrating Adobe's Firefly generative AI models with Nvidia's accelerated computing infrastructure. The deal, unveiled at Nvidia's GTC conference in San Jose, targets enterprise-scale content production and represents a significant expansion of a collaboration spanning more than two decades. The partnership aims to reshape how brands and studios generate marketing materials, product imagery, and campaign assets. Adobe will develop its next generation of Firefly models using Nvidia's CUDA-X, NeMo libraries, and Cosmos open models, according to the announcement. The companies described the goal as achieving "best-in-class creative precision and control" across creative and marketing pipelines. Adobe also plans to explore Nvidia's Agent Toolkit software and Nemotron open models to create automated workflows for content and campaign production. "Content creation is exploding, and our partnership with NVIDIA is grounded in a shared vision to reinvent creative and marketing workflows with the power of AI," Adobe chair and CEO Shantanu Narayen said. Nvidia founder and CEO Jensen Huang said the companies are "uniting our research and engineering teams to accelerate Adobe's beloved applications with NVIDIA CUDA and jointly build state-of-the-art world foundation models that reimagine creativity and transform customer experiences." A cloud-native 3D digital twin solution entered public beta as part of the announcement. The platform creates virtual replicas of physical products using Nvidia Omniverse libraries, OpenUSD data interchange, RTX rendering, and real-time cloud streaming. Brands can use the system to generate pack shots, lifestyle imagery, configurable 3D product experiences, and virtual try-ons. Adobe's Firefly Foundry, an enterprise service for custom AI models trained on proprietary brand content, will integrate Nvidia's computing and AI technologies. Nvidia infrastructure will also be incorporated into Adobe Acrobat, Photoshop, Premiere Pro, Frame.io, GenStudio, and Adobe Experience Platform. The partnership includes a joint go-to-market strategy to drive enterprise adoption through Firefly Foundry. The Adobe deal forms part of Nvidia's broader enterprise AI initiative at GTC, which includes the launch of the Agent Toolkit platform with partners including Salesforce and SAP, according to VentureBeat.

Adobe, Nvidia Join Hands to Build the Next Generation of Firefly Models

The two tech giants will collaborate to build agentic creative workflows Adobe and Nvidia announced a strategic partnership at the GPU Technology Conference (GTC) on Monday. This collaboration focuses on artificial intelligence (AI). It spans agentic creativity and marketing workflows, the development of the next-generation Firefly models, and the integration of Nvidia's platforms into Adobe products. The partnership is part of several major announcements made by Nvidia at the event, including the introduction of the NemoClaw stack for OpenClaw agents, DLSS 5 graphics upscaler, and the launch of the Vera Rubin platforms for agentic workloads. Adobe and Nvidia Announce Strategic Partnership In a newsroom post, the GPU maker announced the partnership with Adobe and detailed the key pillars. A key focus area is developing the next generation of Firefly models. The company's latest generation of Firefly models was unveiled in October 2025, including the Firefly Image Model 5. The next iteration of the foundation model will now be built on Nvidia's AI infrastructure. The post highlighted that the future models will be focused on creative precision and control, as well as marketing pipelines. To support the development, Adobe will use Nvidia's computing technology by tapping into its Cuda-X, NeMo libraries, Cosmos open models, and the Agent Toolkit software. "For more than 20 years, Nvidia and Adobe have partnered to push the boundaries of design and creativity. Today, we are taking that partnership to a new level -- uniting our research and engineering teams to accelerate Adobe's beloved applications with Nvidia CUDA and jointly build state-of-the-art world foundation models that reimagine creativity and transform customer experiences," said Jensen Huang, founder and CEO of Nvidia. The two companies will also collaborate on new agentic creative and marketing workflows for content, campaign, and production speed, where Adobe will explore Nvidia's Agent Toolkit software and Nemotron ropn models to power these experiences. Additionally, Adobe will also launch a cloud-native, brand identity-preserving 3D digital twin solution focused on marketing. It will create virtual replicas of physical products aimed at creating experiences for enterprises. It will integrate Nvidia's Omniverse libraries. The two companies will also work together on NemoClaw, Nvidia's security and privacy layer for OpenClaw agents.

Adobe and Nvidia are collaborating to build the next-generation Firefly AI platform

Adobe has announced on their website a strategic partnership to build the next-generation Firefly AI platform. The idea is to expand Adobe's generative AI ecosystem, which covers the production of images, videos, audio and design. Firefly already offers access to several AI models. And in the future, Nvidia will bring its technologies to the platform, such as CUDA-X computing libraries, Cosmos models and agent-based AI tools. Among these innovations is 3D digital twin, which is "the solution [that] creates virtual replicas of physical products that act as permanent digital identities for marketing and commerce experiences". In other words, it allows companies to create virtual copies of their products for marketing use, like the production of advertising materials without a physical description, and realistic images and videos can be generated from the product directly in a digital environment. What's more, Adobe plans to utilise Nvidia's NemoClaw technology, which brings more security and management tools to agent-based AI. These can be used to build systems that can independently produce content based on a given description, from background research to a finished result. Nvidia's technology will be widely integrated into Adobe's products, like Acrobat, Photoshop, Premiere Pro, Frame.io, and Experience Platform, as reported by Muropaketti.

Adobe and NVIDIA Partner to Advance Next Generation Firefly AI Models

Adobe and NVIDIA will also work together on NVIDIA NemoClaw -- an open source stack that simplifies running OpenClaw always-on assistants more safely, with a single command. As part of the NVIDIA Agent Toolkit, it installs the NVIDIA OpenShell runtime -- a secure environment for running autonomous agents and open source models like NVIDIA Nemotron. In partnership with NVIDIA, Adobe is launching a cloud-native, brand identity-preserving 3D digital twin solution (public beta). The solution creates virtual replicas of physical products that act as permanent digital identities for marketing and commerce experiences. Integrating NVIDIA Omniverse™ libraries into Adobe technologies, the collaboration expands support for 3D digital twin workflows built on OpenUSD for marketing content automation. With seamless interoperability across tools, brands can generate everything from consistent pack shots and lifestyle imagery to configurable 3D product experiences and immersive virtual try-ons. Adobe will also harness NVIDIA AI infrastructure, AI libraries, services and models to accelerate and optimize every layer of its AI-powered tools across creativity, productivity and customer experience orchestration -- including Adobe Acrobat, Photoshop, Premiere Pro, Frame.io, Adobe Firefly Foundry, Adobe GenStudio and Adobe Experience Platform. With Adobe Firefly Foundry, Firefly's commercially safe AI models are deeply tuned with a company or IP owner's unique, proprietary brand or franchise content, which is critical for media and entertainment studios. Adobe Firefly Foundry will integrate NVIDIA's advanced computing and AI technologies to power enterprise‑grade custom AI that delivers commercially safe content at scale.

Adobe partners with NVIDIA to develop next-gen Firefly AI models and agentic workflows

Adobe has announced a strategic partnership with NVIDIA to accelerate AI-powered creation, production, and personalization. The collaboration includes development of next-generation Adobe Firefly models and expansion of agentic AI workflows. The partnership brings together Adobe's creative and marketing workflows, models and technologies with NVIDIA's open models, libraries, research and accelerated computing, as generative AI continues to reshape content creation and marketing. As part of the collaboration, Adobe and NVIDIA will jointly develop and integrate AI technologies across creative, marketing and enterprise workflows. Key announcements include: These developments focus on enabling high-quality, interactive content creation, automation and scalable marketing workflows. Adobe will develop the next generation of Firefly models using NVIDIA accelerated computing technologies, including CUDA-X, NeMo and Cosmos open models. These models are designed to deliver improved precision and control across creative and marketing pipelines. Adobe Firefly Foundry will integrate NVIDIA AI infrastructure to deliver enterprise-grade models trained on proprietary brand and franchise data. These models support generation of commercially safe, IP-protected content across image, video, audio, vector and 3D formats. Adobe is exploring NVIDIA OpenShell and Nemotron, part of the NVIDIA Agent Toolkit, to support hybrid, long-running agentic workflows in secure and cost-efficient environments. The companies are also working on NVIDIA NemoClaw, an open-source stack that enables safer deployment of always-on assistants with a single command. It includes the OpenShell runtime for running autonomous agents and open models such as Nemotron. These technologies are being evaluated for integration into large-scale workflows within Adobe Experience Platform, with NVIDIA providing engineering support and early software access. Adobe is introducing a cloud-native, brand identity-preserving 3D digital twin solution in public beta. The solution creates virtual replicas of physical products that act as persistent digital identities for marketing and commerce. The platform integrates NVIDIA Omniverse libraries, OpenUSD framework, RTX rendering and Omniverse Kit App Streaming for real-time cloud streaming with Adobe's generative AI and workflow systems. It enables creation of: Adobe will use NVIDIA AI infrastructure, libraries, services and models to optimize its platforms across creativity, productivity and customer experience orchestration. This includes: Frame.io will use NVIDIA CUDA acceleration for media decoding, content management, semantic search and AI-driven insights across image, video and 3D content. The 3D digital twin solution is available in public beta. Other integrations and AI capabilities will roll out across Adobe platforms over time. Speaking on the partnership, Shantanu Narayen, chair and CEO, Adobe, said: Content creation is rapidly expanding, and our partnership with NVIDIA is grounded in a shared vision to reinvent creative and marketing workflows with the power of AI. As AI transforms how marketing teams and media and entertainment studios operate, Adobe and NVIDIA will bring together our Firefly models, CUDA libraries into our applications, 3D digital twins for marketing, and Agent Toolkit and Nemotron into our agentic frameworks to deliver high-quality, controllable and enterprise-grade AI workflows of the future. Commenting on the partnership, Jensen Huang, Founder and CEO, NVIDIA, said:

Adobe and NVIDIA Partner to Scale Firefly and Agentic Workflows

Through a strategic collaboration, Adobe and NVIDIA are integrating advanced computing and AI technologies to redefine the creative and marketing landscape. By leveraging NVIDIA's high-performance libraries, Adobe is developing the next generation of foundational Firefly models, designed to offer unparalleled precision and control across creative pipelines. This partnership extends into the realm of agentic workflows, streamlining content production and campaign speed through intelligent, automated systems. Furthermore, Adobe is building a cloud-native 3D digital twin solution on NVIDIA Omniverse, specifically engineered to help brands maintain visual identity while scaling marketing assets. These innovations are anchored by the Adobe Firefly Foundry, which utilizes NVIDIA's AI infrastructure to power enterprise-grade, custom AI solutions that generate commercially safe content at a global scale. Adobe and NVIDIA today announced a strategic partnership to accelerate AI-powered creation, production and personalization, including delivering the next generation of foundational Adobe Firefly models and agentic workflows. The partnership will bring together Adobe's creative and marketing workflows, models and technology and NVIDIA's open models, libraries, research and accelerated computing, as demand for content continues to surge and generative AI reshapes creative and marketing workflows. Through this partnership, Adobe and NVIDIA will advance the creative industry by developing next-generation Firefly models that will deliver best-in-class creative precision and control for creativity and marketing pipelines. The models will be built on NVIDIA's advanced computing technology and tap into NVIDIA CUDA-X™,NVIDIA NeMo™ libraries, NVIDIA Cosmos™ open models and NVIDIA Agent Toolkit software to enable the interactive, high-quality creation customers expect. "Content creation is exploding, and our partnership with NVIDIA is grounded in a shared vision to reinvent creative and marketing workflows with the power of AI," said Shantanu Narayen, chair and CEO, Adobe. "As AI transforms how marketing teams and media and entertainment studios work, Adobe and NVIDIA will bring together our Firefly models, CUDA libraries into our applications, 3D digital twins for marketing, and Agent Toolkit and Nemotron to our agentic frameworks to deliver high-quality, controllable and enterprise-grade AI workflows of the future." "AI is giving every industry the ability to redefine what's possible," said Jensen Huang, founder and CEO, NVIDIA. "For more than 20 years, NVIDIA and Adobe have partnered to push the boundaries of design and creativity. Today, we are taking that partnership to a new level -- uniting our research and engineering teams to accelerate Adobe's beloved applications with NVIDIA CUDA and jointly build state-of-the-art world foundation models that reimagine creativity and transform customer experiences." Adobe and NVIDIA will collaborate to deliver breakthrough agentic creative and marketing workflows for content, campaign and production speed. Adobe will explore NVIDIA Agent Toolkit software and NVIDIA Nemotron™ open models to power these agentic workflows. Adobe and NVIDIA will also work together on NVIDIA NemoClaw -- an open source stack that simplifies running OpenClaw always-on assistants more safely, with a single command. As part of the NVIDIA Agent Toolkit, it installs the NVIDIA OpenShell runtime -- a secure environment for running autonomous agents and open source models like NVIDIA Nemotron. In partnership with NVIDIA, Adobe is launching a cloud-native, brand identity-preserving 3D digital twin solution (public beta). The solution creates virtual replicas of physical products that act as permanent digital identities for marketing and commerce experiences. Integrating NVIDIA Omniverse™ libraries into Adobe technologies, the collaboration expands support for 3D digital twin workflows built on OpenUSD for marketing content automation. With seamless interoperability across tools, brands can generate everything from consistent pack shots and lifestyle imagery to configurable 3D product experiences and immersive virtual try-ons. Adobe will also harness NVIDIA AI infrastructure, AI libraries, services and models to accelerate and optimize every layer of its AI-powered tools across creativity, productivity and customer experience orchestration -- including Adobe Acrobat, Photoshop, Premiere Pro, Frame.io, Adobe Firefly Foundry, Adobe GenStudio and Adobe Experience Platform. With Adobe Firefly Foundry, Firefly's commercially safe AI models are deeply tuned with a company or IP owner's unique, proprietary brand or franchise content, which is critical for media and entertainment studios. Adobe Firefly Foundry will integrate NVIDIA's advanced computing and AI technologies to power enterprise‑grade custom AI that delivers commercially safe content at scale. Adobe and NVIDIA are deepening their strategic partnership to deliver a comprehensive suite of AI-driven innovations across the creative and enterprise landscapes. At the core of this collaboration is the development of next-generation Adobe Firefly models, which leverage NVIDIA's CUDA-X® and NeMo libraries to provide industry-leading precision for marketing and productivity. This evolution extends into agentic AI, with Adobe exploring NVIDIA OpenShell™ and Nemotron to create secure, high-performance "agentic loops" within the Adobe Experience Platform. By unifying NVIDIA's accelerated computing and Omniverse libraries with Adobe's workflow platforms, the partnership is also pioneering 3D digital twins -- a cloud-native solution that allows brands to automate high-fidelity marketing content while strictly preserving their visual identity. Beyond content creation, the alliance is set to transform business productivity and media management. Adobe Acrobat will integrate NVIDIA Nemotron capabilities to enhance AI-driven document intelligence, while Frame.io will utilize NVIDIA CUDA® to accelerate cloud workflows, enabling lightning-fast semantic search and generative insights across video and 3D media. To ensure these proprietary, IP-protected models reach global franchises at scale, Adobe is launching Firefly Foundry, supported by a joint go-to-market strategy with NVIDIA. This unified approach ensures that enterprises can move from experimentation to full-scale production with faster, more flexible, and commercially safe AI tools.

Adobe, Nvidia expand partnership for AI-powered creative tools By Investing.com

SAN JOSE, Calif. - Adobe (NASDAQ:ADBE) and NVIDIA (NASDAQ:NVDA) announced Monday a strategic partnership to accelerate AI-powered content creation and marketing workflows, according to a press release statement.The announcement comes as Adobe shares trade near their 52-week low of $244.28, down nearly 30% over the past six months despite the company's market capitalization of $102.8 billion. According to InvestingPro analysis, Adobe appears significantly undervalued at current levels, with the stock featured on the platform's Most Undervalued list. The collaboration will focus on developing the next generation of Adobe Firefly models using NVIDIA's computing technology and libraries, including CUDA-X, NeMo libraries, Cosmos open models and Agent Toolkit software. "Content creation is exploding, and our partnership with NVIDIA is grounded in a shared vision to reinvent creative and marketing workflows with the power of AI," said Shantanu Narayen, chair and CEO of Adobe.The software giant, which generated $24.45 billion in revenue over the last twelve months, maintains an impressive gross profit margin of 89.4%. InvestingPro Tips highlight Adobe as a prominent player in the Software industry with exceptional profitability metrics. Notably, 26 analysts have revised their earnings upwards for the upcoming period, signaling growing confidence in the company's AI strategy. Investors can access 15 additional ProTips and comprehensive Pro Research Reports covering Adobe and 1,400+ other US equities on the platform. The companies will work together on agentic creative and marketing workflows, with Adobe exploring NVIDIA Agent Toolkit software and Nemotron open models. Adobe will also integrate NVIDIA Omniverse libraries to create a cloud-native 3D digital twin solution for marketing and commerce experiences, currently in public beta. Adobe plans to use NVIDIA AI infrastructure across its product suite, including Acrobat, Photoshop, Premiere Pro, Frame.io, Adobe Firefly Foundry, Adobe GenStudio and Adobe Experience Platform. Adobe Firefly Foundry will integrate NVIDIA's computing and AI technologies to deliver custom AI models for enterprise clients. The service allows companies to tune Firefly's AI models with proprietary brand content. Frame.io, Adobe's content management platform, will incorporate NVIDIA CUDA to accelerate media decoding and enable semantic search across image, video and 3D content. Adobe Acrobat will integrate NVIDIA Nemotron capabilities to enhance AI output quality for business users. The partnership builds on a collaboration between the two companies spanning more than 20 years, according to Jensen Huang, founder and CEO of NVIDIA. The companies will develop a joint go-to-market strategy to drive enterprise adoption of these AI innovations through Adobe Firefly Foundry. In other recent news, Adobe's first-quarter fiscal 2026 results revealed a 12% growth in both non-GAAP operating income and revenue. The company reported total revenue of $6.4 billion for the quarter, surpassing BofA Securities' estimates by $120 million. However, the announcement of CEO Shantanu Narayen's departure has led to mixed reactions among analysts. Argus downgraded Adobe's stock rating to Hold, citing the leadership transition as a significant factor. Deutsche Bank also lowered its price target to $310, maintaining a Hold rating due to uncertainties surrounding the CEO transition. UBS adjusted its price target to $290, highlighting potential AI-related challenges despite Adobe's strong revenue performance. BofA Securities, however, reiterated a Buy rating with a $460 price target, emphasizing the company's robust revenue and annual recurring revenue growth. Meanwhile, BMO Capital reduced its price target to $285, expressing concerns over Adobe's annual recurring revenue figures. These developments underscore the varied analyst perspectives on Adobe's future amid its leadership changes. This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.

Adobe and NVIDIA push Firefly beyond image generation into full-scale enterprise workflows

Adobe and NVIDIA have announced a broader strategic partnership that points to where enterprise AI in creative and marketing software is heading. Rather than focusing only on faster image generation or isolated AI features, the collaboration is aimed at building a deeper technology stack that touches model development, workflow automation, 3D content creation and enterprise-grade custom AI. At the centre of the announcement is Adobe's plan to use NVIDIA's computing technology and software libraries to develop the next generation of Adobe Firefly foundational models. According to the companies, these future models are being designed to offer stronger precision and control across creative and marketing pipelines. Adobe says the upcoming Firefly models will be built using NVIDIA's accelerated computing technologies and will tap into software and libraries such as CUDA-X, NeMo, Cosmos open models and the NVIDIA Agent Toolkit. The intended result is better quality and more interactive output across content production workflows. This suggests Adobe is trying to sharpen Firefly into a more dependable platform for serious production environments, not merely a generative add-on inside creative software. The second major part of the partnership is agentic AI. Adobe and NVIDIA say they will collaborate on agentic creative and marketing workflows designed to improve the speed of content creation, campaign execution and production processes. Adobe is exploring NVIDIA OpenShell, Nemotron and the broader Agent Toolkit as foundations for long-running AI agents that can operate in secure and cost-efficient environments. The industry conversation is moving from single-prompt generation toward systems that can handle multi-step tasks across connected software environments. For Adobe, that could mean AI agents helping teams manage campaign assets, automate repetitive production work, or coordinate workflows across products such as Adobe Experience Platform, GenStudio and Firefly Foundry. The language from both companies suggests they see agentic AI as a full-fledged workflow layer. Another notable part of the announcement is Adobe's new cloud-native 3D digital twin solution, which is entering public beta. Built with NVIDIA Omniverse libraries, the platform is designed to create virtual replicas of physical products while preserving brand identity. Adobe is positioning these 3D digital twins as long-term digital assets for marketing and commerce use cases. That opens the door to more automated creation of pack shots, lifestyle imagery, configurable 3D product views and even virtual try-ons. By building the workflow around OpenUSD and NVIDIA's Omniverse stack, Adobe is also signalling that 3D marketing content will become more connected, reusable and scalable across tools and channels. For brands dealing with large product catalogues, that may prove more valuable than one-off AI image generation. The partnership also stretches across Adobe's wider software portfolio. The companies say Adobe will use NVIDIA AI infrastructure, services and models across products including Acrobat, Photoshop, Premiere Pro, Frame.io, Firefly Foundry, GenStudio and Adobe Experience Platform. In Acrobat, NVIDIA Nemotron capabilities are expected to improve AI document intelligence. In Frame.io, CUDA acceleration is expected to help with cloud workflows, semantic search and media intelligence across image, video and 3D content. For enterprise customers, Firefly Foundry may be one of the more significant pieces of the deal. Adobe says the platform will integrate NVIDIA technologies to help companies build custom, commercially safe AI models tuned to proprietary brand and franchise content. That is especially relevant for media companies, studios and global brands that want AI tools trained around their own assets rather than generic public models.