Gartner: Spending On GenAI Models To Total $14.2 Billion In 2025
Specialized GenAI models are trained or fine-tuned on industry or business process-specific data. The way DSLMs are developed and deployed results in an impact that exceeds the direct spending they generate. Gartner predicts that by 2027, more than half of the GenAI models used by enterprises will be domain-specific (that is, specific to an industry or business function), up from 1% in 2024.
'Foundation GenAI models (including LLMs) are trained on vast amounts of data and used for many different tasks. They are the first models supporting GenAI and will continue to represent the largest area of spending by organizations in the coming years,' said Arunasree Cheparthi, Senior Principal Research Analyst at Gartner. 'However, organizations are also turning to more domain-specific or vertical GenAI models because they offer improved performance, cost, reliability and relevance in targeted enterprise use cases over foundation models.'
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
TECHx
18 hours ago
- TECHx
Why Quality Data Matters for AI Success
Home » Expert opinion » Why Quality Data Matters for AI Success Quality data is the foundation of AI success. Yasser Shawky explores why trusted, AI-ready data is essential for innovation, scalability, and long-term business value. I was recently speaking to my mother about what she'd been cooking lately, and was amused to learn that she now uses ChatGPT to help her come up with new recipes. I couldn't help but smile at how pervasive AI has already become. When even our parents are comfortable treating LLMs as their culinary consultants, you know the technology has crossed a noteworthy threshold. It also got me thinking. In this case, the advice she received was useful. But what if the AI hallucinated, as it sometimes does, and suggested swapping sumac for cinnamon in a delicate Middle Eastern dish? That would almost certainly ruin the flavour. And what if the same thing happened in a more serious context like a home remedy, financial advice, or supplement recommendation? The consequences could be dire. That small, charming exchange has stayed with me, because it's a fitting metaphor for what I see happening across the Middle East as organisations rush to capitalise on AI. Like my mother with her recipes, businesses see the promise and want to seize it. But without proper visibility into the inputs AI models are drawing from, and without governance over the data that fuels them, they're exposing themselves to risk. AI's Dot-Com Moment With AI adoption proceeding at breakneck pace, it might feel like we're venturing deep into uncharted territory. And while that's undoubtedly true, the IT industry has stood at a similar inflection point before. Just over two decades ago, the world watched as the dot-com boom of the late 1990s rewrote the rules of business almost overnight. In that exhuberent environment, many crashed and burned, but those that invested in solid infrastructure, user experience, and sustainable business models emerged as titans, still standing strong today. Now, we're living through AI's own dot-com moment and just as with the internet revolution, the winners won't be those who leap in first. Winners will be those who lay the right foundations, especially when it comes to data. Why Data, Not Just AI, Should Be the Starting Point It's easy to be dazzled by the tech itself. Large language models, generative AI, machine learning, have entered daily dialogue. But these are just enablers. The true foundation of any successful AI implementation is a strong data platform. It's a truth the industry has long known but often skirts around, poor data quality destroys business value. According to Gartner, the average cost of bad data is estimated at US$10.8 million per year per organisation. The reasons? Duplicate entries, outdated records, inconsistent formatting, missing fields, lack of data governance and lineage, and fragmented systems. Every Chief Data Officer is all too familiar with these challenges. But flip the equation and the potential is transformative. When data is accurate, complete, consistent, timely and relevant, what we at Informatica call 'AI-ready', then use cases that once seemed overly complex suddenly become achievable and durable. That's when data chaos is turned into business value. It's not just about better models. It's about creating a single, authoritative view of enterprise data that everyone, from executives to frontline staff, can trust. Like a chef who knows exactly what's in their pantry and how fresh it is, with data you can trust, you can cook up something extraordinary, without unpleasant surprises. One Foundation, Many Use Cases Let's explore this in practice by taking the example of an innovative bank. Its marketing team is keen to leverage AI to offer hyper-personalised financial products. They want to know when a customer might be buying a home, starting a business, or sending their child to university so they can provide timely, tailored offers. Meanwhile, their fraud detection team wants to use AI to spot suspicious patterns in real-time, saving customers the need to self-report. Two vastly different use cases. But they share one foundational need: up-to-date, accurate behavioural and transactional data. The marketing team can't personalise without knowing the customer. The fraud team can't protect without understanding what 'normal' looks like. If each team tries to build its own data pipeline, you end up with duplication, inefficiency, and inconsistent insights. Instead, by investing in a common, trusted data layer, like we at Informatica call turning chaos into business value, you empower every department across the organization to innovate, scale and advance. According to McKinsey, as much as 70% of the effort in developing AI solutions is spent on wrangling and harmonising data. Solve that once, and your entire organisation becomes more agile, more aligned and more AI-ready. Getting Your Data AI-Ready So how do you get there? It starts with acknowledging that data is not just an IT problem, it's a strategic asset belonging to the entire organisation. Preparing your data for AI involves more than cleansing spreadsheets. It requires a robust data governance framework that ensures integrity, compliance and accessibility. It requires scalable infrastructure, ideally cloud-based, that can grow with your ambitions. It requires investment in data literacy, so your people, across all functions, understand how to question, interpret and act on data. Most of all, it requires trust. As Robin Miller, Group Data Manager at Lowell so perfectly put it, 'The vision I describe to my colleagues is that they'll be able to implicitly trust the data that informs them, no matter where in our organisation it comes from.' That trust is what enables action. And in a world moving as fast as ours, trusted action is everything. Long-Term Value over Short-Term Buzz There's no denying the momentum behind AI. But there's a real risk in chasing short-term wins — pilots that can't scale, models that aren't reproducible, projects that promise much but deliver little. What's needed is a long-term mindset. That begins with data. By prioritising data quality and investing in AI-readiness from the ground up, organisations position themselves to move faster, not slower. They create the conditions for innovation to thrive. They accelerate business value by enabling data to flow freely, reliably and securely across systems, departments and applications. And most importantly, they ensure that when the hype cycle inevitably plateaus, their AI investments continue to pay dividends. In the race for AI success, it's not the flashiest use cases that win, it's the strongest data foundation. By Yasser Shawky, Vice President, Emerging Markets (MEA) at Informatica
Channel Post MEA
a day ago
- Channel Post MEA
Gartner: Spending On GenAI Models To Total $14.2 Billion In 2025
Worldwide end-user spending on generative AI (GenAI) models is projected to total $14.2 billion in 2025, according to Gartner, Inc. End-user spending on specialized GenAI models, which include domain-specific language models (DSLMs), is estimated to total $1.1 billion in 2025 (see Table 1). Specialized GenAI models are trained or fine-tuned on industry or business process-specific data. The way DSLMs are developed and deployed results in an impact that exceeds the direct spending they generate. Gartner predicts that by 2027, more than half of the GenAI models used by enterprises will be domain-specific (that is, specific to an industry or business function), up from 1% in 2024. 'Foundation GenAI models (including LLMs) are trained on vast amounts of data and used for many different tasks. They are the first models supporting GenAI and will continue to represent the largest area of spending by organizations in the coming years,' said Arunasree Cheparthi, Senior Principal Research Analyst at Gartner. 'However, organizations are also turning to more domain-specific or vertical GenAI models because they offer improved performance, cost, reliability and relevance in targeted enterprise use cases over foundation models.'
Tahawul Tech
2 days ago
- Tahawul Tech
Vectra AI named ‘Leader' in first-ever Gartner Magic Quadrant for NDR
Recognised for the strength of its AI-driven NDR platform, Vectra AI is positioned furthest for completeness of vision and highest for ability to execute in the inaugural report. Vectra AI, Inc., the cybersecurity AI company that protects modern networks from modern attacks, announced it was named a Leader in the 2025 Gartner® Magic Quadrant for Network Detection and Response (NDR). Vectra AI is positioned highest for ability to execute and furthest for completeness of vision, and is the only vendor in the report to be named a leader in both the Gartner Magic Quadrant for NDR and a Customer Choice Winner for NDR in the 2024 Gartner Peer Insights Voice of the Customer. The Vectra AI Platform is purpose-built to defend modern hybrid environments from identity and network-based attacks. As threats accelerate across cloud, data center, remote, and OT domains, Vectra AI provides comprehensive coverage to reduce attack exposure. Its AI agents continuously triage, correlate, and prioritise real threats in real time – eliminating alert fatigue and accelerating response. With Vectra AI, defenders gain control to detect, hunt, investigate, and respond to attacks across the full threat landscape, enabling security teams to focus on maturing their security posture. 'Gartner's decision to publish a Magic Quadrant for NDR reflects just how essential this market has become in modern cyber defense,' said Hitesh Sheth, founder and CEO of Vectra AI. 'Being recognised as a Leader in this inaugural report reinforces Vectra AI's position at the forefront of this critical space. As organisations grapple with growing complexity, identity-based attacks, and AI-driven threats, the Vectra AI Platform delivers what modern defenders need – coverage that reduces exposure, clarity that cuts through the noise, and control to act with speed and confidence,' added Sheth. Vectra AI has been recognized by customers for outstanding product performance and support, earning the distinction of Customers' Choice in the 2024 Gartner Peer Insights Voice of the Customer for Network Detection and Response. As of January 2024, Vectra AI holds a 4.8 out of 5 rating based on 96 reviews, with 96% of customers saying they would recommend the platform. This recognition reflects Vectra AI's deep commitment to customer success and ongoing innovation in protecting modern networks from modern attacks.
