Some weeks in tech feel loosely held together, like a set of headlines that were never really meant to be in the same room yet somehow still point in the same direction. This was one of those weeks. AI data centre spending kept climbing, the EU floated changes to environmental checks for strategic projects, and storage analysts reported more pressure on enterprise SSD pricing.[1][2][7] None of that exists in isolation if you are planning infrastructure or advising customers on their next wave of investment.
AI data centre spending keeps climbing
It is still a little odd to see AI infrastructure spending treated almost like a macroeconomic indicator, but that is where we have landed. Recent analysis suggests AI data centre capex is now shaping broader market behaviour rather than just following it.[1][2] Google’s move to put Amin Vahdat at the centre of its AI infrastructure plans underlines the point, signalling an era where physical compute capacity, networks and cooling are every bit as important as the models that sit on top.[3]
For teams scoping server refreshes or new platforms, that shift shows up in very practical ways. Conversations that once focused on CPU counts and basic storage now spill into questions about GPU density, memory bandwidth, cooling headroom and how fabrics will cope as workloads grow. In those moments, drawing on Hammer’s experience across servers and AI technologies helps pull the discussion back into something manageable and real. When a customer needs to balance AI ambition with the constraints of their racks and power, having a clear view of the server and compute options available through that portfolio can make the path forward much less abstract.
Europe rethinks rules for strategic builds
The regulatory side of the story moved as well. The European Commission has proposed easing certain environmental assessments for strategic projects, including AI data centres and gigafactories, in an attempt to speed up deployment while still claiming alignment with longer term climate goals.[4][5] It is a delicate balance and it will not be without controversy, but it shows how seriously policymakers are taking the need for new capacity.
Closer to home, the announcement that part of the Drax coal site is set to become a 100 MW data centre by 2027 carries its own message.[6] There is symbolism in a coal facility being repurposed for digital infrastructure, but the real signal is about power. Grid access and available capacity are increasingly the hard limits on what can be built and how fast it can go live.
For partners and MSPs, this changes how early power and resilience conversations need to happen. Instead of treating power design as a late stage detail, it is becoming part of the first architecture conversations. Being able to anchor that discussion in concrete options from Hammer’s infrastructure portfolio helps. Talking through UPS behaviour, PDU configurations, monitoring and basic data centre hardware feels more productive when there are clear product families and design patterns to point at, rather than leaving it as a vague constraint in the background.
Storage absorbs more of the AI load
Storage rarely gets the headlines, but it has a way of deciding how smooth everything else feels. TrendForce reported another rise in enterprise SSD shipments and pricing in Q3 2025, reflecting a mix of tighter supply and robust demand from AI and cloud workloads.[7] Earlier production cuts have not fully unwound, and analysts note that new inference workloads are stressing storage fabrics in ways that traditional refresh cycles do not always anticipate.[7][8]
On the ground, that often looks like a familiar squeeze. Customers want low latency and strong throughput for AI and analytics while expecting their storage budget to behave predictably. The all flash conversation can be tempting, but it is not always realistic. This is where Hammer’s experience around enterprise data storage solutions and enterprise components becomes useful. Many environments benefit from placing the most demanding workloads on NVMe SSD tiers while shifting long term or low touch data to high capacity HDD layers. The exact split varies by customer, but the principle of matching tiers to use cases is consistent.
Why this matters for 2026 planning
Each of these stories on its own is interesting but not transformative. Put together, they sketch out a pattern. AI workloads are pulling infrastructure design into new territory, regulators are adjusting rules to keep up with strategic demand and storage markets are moving under the pressure of fresh requirements.[1][2][4][7]
For enterprises, MSPs and partners working with Hammer, this is a useful moment to check the assumptions baked into 2026 plans. Density expectations, power availability and storage tiering are no longer static variables. When multiple layers of the stack move at once, having a partner that can connect servers, networking, storage and physical infrastructure in a coherent way becomes a meaningful advantage rather than a simple convenience.
More detail on Hammer’s product families and services is available at www.hammerdistribution.com.
References
- Reuters – “AI masking the economy cuts both ways” (10 Dec 2025)
- S&P Global – “Data center investment moves macro needle”
- Reuters – “Google names Amin Vahdat as new chief of AI infrastructure buildout”
- The Guardian – “EU proposes exempting AI gigafactories from environmental assessments”
- Council of the EU – “AI: Council adopts position on the updated regulation to create AI gigafactories”
- Reuters – “Drax plans to turn coal era infrastructure into data centre by 2027”
- TrendForce – “Enterprise SSD prices and shipments surge in 3Q25”
- TrendForce – “AI demand fuels enterprise SSD growth”