The corporate fear of missing out is palpable, and it’s fuelling an astonishing spending spree. We’re not talking about pocket money here. As recently reported by Solutions Review, a data infrastructure company called Vast Data just inked a staggering $1.17 billion in contracts for its AI cloud platform. That’s billion with a ‘B’. This isn’t just about buying some fancy software; it’s about a fundamental re-tooling of corporate infrastructure, a bet that AI is not just another feature but the future of the entire business engine. And when that kind of money is on the table, you know the stakes are incredibly high.
Hitachi Vantara’s IQ Studio: Are We Building Agents of Fortune or Just Fancier Chatbots?
Amidst the noise, Hitachi Vantara made a move that’s worth paying attention to with the launch of its IQ Studio. Their big pitch? It’s a platform for building “agentic AI”. Now, that sounds like a term cooked up in a marketing department, so what does it actually mean? For years, we’ve interacted with AI that can answer questions. You ask it for the capital of Mongolia, and it tells you. Agentic AI is designed to do things. It’s not just a conversationalist; it’s an autonomous worker.
Think of it like this: a standard AI is like a brilliant research assistant who can find any information you ask for instantly. An agentic AI, on the other hand, is like a seasoned executive assistant. You don’t just ask them for a phone number; you say, “Book me a flight to the London conference next Tuesday, find a hotel near the venue, and add it all to my calendar.” The assistant then performs a sequence of actions—browsing flights, comparing prices, booking a room, checking your schedule—to achieve the goal. That’s the promise of agentic AI for businesses: intelligent systems that can execute complex, multi-step workflows, from processing insurance claims to managing supply chain logistics, with minimal human intervention.
The question for Hitachi Vantara and its clients is whether this will translate into truly autonomous, value-adding agents or just more sophisticated, and potentially more complicated, automation scripts. Will these AI agents become indispensable digital colleagues, or will they be the 21st-century version of Clippy, popping up to “help” in all the wrong ways? The potential is enormous, but the path from a development studio to a flawlessly operating AI workforce is fraught with peril.
The Unsexy Plumbing: Why SnapLogic and Informatica AI Matter More Than You Think
Whilst “agentic AI” grabs headlines, the real work, the stuff that makes or breaks any of these grand projects, is happening in the data trenches. This is where companies like SnapLogic and Informatica AI come in, and their work is arguably more critical than the flashy models running on top. An AI is only as smart as the data it’s trained on. If you feed it garbage, you’ll get very confident, very fast, and very expensive garbage in return.
SnapLogic recently introduced new governance tools for what it calls the MCP, or Multi-Cloud Platform. Let’s be frank, most large organisations don’t have a neat, tidy data warehouse. Their data is a chaotic sprawl across Amazon Web Services, Microsoft Azure, Google Cloud, and their own on-premise servers. It’s a mess. SnapLogic’s play is to provide the discipline, the framework to govern this chaos. Their tools are designed to ensure data is consistent, secure, and accessible across these different environments. It’s the digital equivalent of hiring a team of librarians to organise a library after a hurricane. It’s not glamorous work, but without it, you have no library—just a pile of books.
Similarly, Informatica AI has been a stalwart in the data management world for decades, and its role is becoming ever more central. It offers the foundational tools for data integration, quality, and governance that are absolute prerequisites for any serious AI initiative. These companies are providing the essential plumbing. Without clean, well-managed, and properly governed data pipelines, your multi-million-pound AI strategy will collapse under the weight of its own inconsistencies. The biggest challenge in enterprise AI isn’t a lack of clever algorithms; it’s a lack of clean data.
The Scalability Headache: Can Your Building Handle the AI Brain?
The AI arms race has a very real, physical constraint: power and heat. The specialised chips that power today’s AI models, particularly from the likes of Nvidia, are incredibly powerful, but they also run incredibly hot and consume vast amounts of electricity. Your standard air-cooled data centre is starting to look like a relic from a bygone era.
This is why industry voices, like those from Lenovo, are increasingly stressing the need for liquid-cooled infrastructure. As reported in the same weekly roundup from Solutions Review, high-density AI workloads are pushing traditional cooling systems to their limits. It’s a simple physics problem: you can’t cram that much computational power into a small space without finding a more efficient way to get rid of the heat. Trying to run a massive AI training cluster on air cooling is like trying to cool a blast furnace with a desk fan. It’s simply not going to work.
This physical reality puts that $1.17 billion Vast Data deal into perspective. A significant chunk of that money isn’t just for software licences; it’s for building the physical infrastructure that can handle these extreme demands. Enterprises are realising that becoming an “AI company” isn’t just a software upgrade. It might require a complete overhaul of their physical data centres, a capital-intensive undertaking that separates the truly committed from those just paying lip service to the trend.
Governance: The Adults Finally Enter the Room
With all this power, data, and autonomy, a single, terrifying question arises: who is in control? The rush to deploy AI has often outpaced the development of robust governance frameworks to manage it. We’re beyond simply worrying about data privacy; we’re now talking about cost control for runaway AI processes, ensuring model outputs aren’t biased or discriminatory, and guaranteeing the AI’s actions align with the organisation’s strategic goals and ethical standards.
The introduction of dedicated governance tools, like those from SnapLogic, is a sign of a maturing market. The initial “move fast and break things” phase is giving way to a more sober realisation that ungoverned AI is a massive liability. Organisations are now scrambling to establish best practices, create oversight committees, and implement frameworks that provide guardrails for their AI initiatives. This is the necessary, if sometimes tedious, bureaucracy that follows every technological explosion. The Wild West always gets a sheriff, and the untamed frontier of enterprise AI is finally seeing law and order beginning to take shape.
So, where does this leave us? The latest wave of enterprise AI updates points to a few clear trends. Firstly, the ambition is shifting from passive AI that analyses the past to active, agentic AI that shapes the future. Secondly, the unglamorous-but-vital work of data management and governance is finally getting the attention it deserves. And thirdly, the immense physical and financial costs of building true AI capability are becoming starkly clear.
The race is well and truly on, and the spending is astronomical. The real question we should all be asking is this: a year from now, how many of these grand AI projects will be delivering tangible, transformative value, and how many will simply be expensive monuments to a fear of being left behind? What do you think?
