This week, the cloud story was less about new services and more about the pipes underneath everything.
In just a few days, we saw:
AWS and Google Cloud quietly launch a jointly engineered multicloud network fabric.
OpenAI and NEXTDC sign a deal to build a 650 MW AI campus in Sydney, triggering warnings about national power supply.
BT roll out a UK sovereign data platform for business and public sector.
If you own cloud strategy, this isn't background noise. It's a preview of how your traffic will flow, where your AI runs, and who controls your data rails.
🌩 This Week's 3 Signals (All From This Week Only)
1. AWS & Google Launch Joint Multicloud Networking
What happened
AWS and Google Cloud announced a jointly engineered multicloud networking solution, combining AWS Interconnect-multicloud with Google Cloud Cross-Cloud Interconnect. You can now spin up private, high-speed links between AWS and GCP in minutes, via console or API, with bandwidth from 1 to 100 Gbps and quad-redundant, MACsec-encrypted paths.
Why it matters
This is multicloud moving from DIY wiring to first-class product. Instead of bespoke VPNs and hand-built routing, you get opinionated, managed connectivity between two hyperscalers your teams probably already use. That changes how you think about failover, DR, data gravity, and vendor lock-in across clouds.
Action to be taken
In Cloudshot, you'll want to treat AWS↔GCP links as explicit first-class edges in your map, not just "oh, there's a VPN somewhere." Model which workloads could benefit from this: cross-cloud analytics, DR, AI training in one cloud with serving in another. Then simulate what happens if that new backbone is degraded or misconfigured — so multicloud reduces risk instead of quietly adding new failure domains.
2. OpenAI + NEXTDC Plan a 650 MW AI Campus in Sydney
What happened
Australian data center operator NEXTDC signed an MoU with OpenAI to build a hyperscale AI campus and GPU supercluster at its S7 site in Eastern Creek, Sydney — projected capacity around 550–650 MW, making it the largest such facility in the Southern Hemisphere. Local experts immediately raised concerns that the campus could consume as much power as hundreds of thousands of homes, warning that Australia's current grid and upgrade timelines are already lagging AI demand.
Why it matters
This is what AI looks like when it hits physical limits: power, transmission, and policy, not just GPU availability. If you're in APAC, AI-heavy regions like Sydney will become magnets for latency-sensitive workloads — but also hotspots for energy constraints, pricing volatility, and political scrutiny around who gets priority when the grid is tight.
Action to be taken
In Cloudshot, mark AI-intensive regions and providers as a distinct risk layer. Tag workloads that depend on AI campuses (now or in the near roadmap), and run scenarios where capacity is throttled, prices jump, or energy policy forces constraints. The goal: know exactly which services are exposed if "AI power hubs" hit a wall — before that shows up as surprise quota errors and doubled bills.
3. BT Launches a UK Sovereign Data Platform
What happened
BT launched a sovereign data platform for UK public-sector and business customers, explicitly focused on keeping data stored and processed within the UK. The platform is pitched as a way to handle AI, voice, and cloud workloads under tighter regulatory and security expectations, with BT promising true UK data sovereignty, secure operations, and resilience — and plans to extend sovereign options across more products in 2026.
Why it matters
Sovereign cloud isn't just an EU Brussels story anymore. For any organization with UK presence, this raises the bar on where data is allowed to live and which providers can touch it. As AI usage ramps up, regulators are going to ask "where is this model trained, run, logged?" — and answers like "somewhere in Western Europe" will age badly.
Action to be taken
In Cloudshot, separate UK-sovereign workloads from generic "EU/Global" ones on your map. Draw clear boundaries: which datasets must never leave the UK, which services cross those borders, and where third-party SaaS or AI APIs punch unexpected holes. That way, when legal or procurement asks for a residency view, you're not reverse-engineering it from tickets and tribal knowledge.
💡 Cloudshot Tip of the Week
"Draw the pipes, not just the boxes."
This week's stories are really about connectivity and constraints:
New managed pipes between AWS and GCP.
Massive power pipes feeding AI campuses like Eastern Creek.
Legal and regulatory pipes defining what counts as "sovereign" in the UK.
In Cloudshot, the real leverage comes when your diagram isn't just resources, but who talks to whom, over what, and under which rules. Make sure your command map shows:
Cross-cloud links and bandwidth tiers
Region-level dependencies and energy-sensitive hotspots
Sovereign boundaries and data flow crossings
If the pipes aren't visible, the risks are invisible too.
🗓 What We Published This Week
🧠 Dec 1 (Mon)
The cloud isn't failing — your teams are making decisions with incomplete truth
Why outages and cost overruns often come from decision blind spots, not pure technical failure.
Full Article💡 Dec 2 (Tue)
Why engineers resolve symptoms instead of causes: the missing "context gap"
How siloed dashboards force teams to fix what's noisy, not what's actually broken.
Full Article🎥 Dec 3 (Wed)
Demo: Live dependency tracing that reveals hidden choke points before they break
A walkthrough of how Cloudshot's live path tracing surfaces fragile links long before incidents.
Full Article📘 Dec 4 (Thu)
See the financial impact of every change the moment it happens
Linking infra changes to cost in near real-time, so FinOps and engineering finally work off the same picture.
Full Article🔭 Strategic Signal
Put the week together and a pattern emerges:
Hyperscalers are cooperating on multicloud plumbing.
AI vendors are helping build mega-campuses that push local grids to their limits.
Telcos like BT are turning sovereign data platforms into core products, not nice-to-have add-ons.
Cloud isn't flattening.
It's stratifying into pipes, power, and policy — each with its own constraints.
Your job isn't to pretend it's simple.
Your job is to give your org one clear map of that complexity so decisions aren't made blind.
That's what Cloudshot is built to be: the command map for this new plumbing.
⚠️ Before It Happens to You…
One misconfigured multicloud link.
One AI campus bumping into power limits.
One new sovereign requirement in a key market.
You won't get to schedule those moments.
You do get to decide whether your team walks into them with a shared picture — or a collection of disconnected dashboards.
The Cloud Today compresses a week of cloud and AI infra shifts into a 2-minute briefing so your leadership doesn't waste hours guessing what actually mattered.
🕒 Two minutes now. Months of misaligned roadmaps avoided later.