ChatGPT said:
Good question — it actually comes down to a mix of economics, geopolitics, and infrastructure. A few key points explain why the US (or US companies like Google) would invest in big data centres in the UK despite high energy costs: 1. Geopolitical stability & trust
Hosting data in the UK means keeping it within a trusted ally’s jurisdiction. For sensitive data, governments and corporations don’t always want it stored in countries with weaker legal or political ties.
It also reassures UK and EU clients who want their data stored closer to home under UK/EU regulations.
2. Connectivity & demand
The UK is a global internet hub with excellent undersea cable connections to the US, Europe, Africa, and beyond. This makes it efficient to serve millions of users in Europe and beyond.
Demand for cloud services is huge in the UK and Europe, and companies like Google, AWS, and Microsoft need local infrastructure to compete.
3. Energy strategy
Yes, the UK’s energy prices are high, but tech giants sign long-term contracts for renewable energy to stabilise costs.
Google, for example, offsets its data centres with direct investment in wind and solar projects in the UK. That way, they “lock in” energy supply and price for decades, insulating themselves from volatile energy markets.
4. Business incentives
The UK government often offers tax breaks, land deals, or infrastructure support to attract big investment. The prestige of hosting major tech infrastructure can outweigh the local cost of energy.
5. Redundancy & resilience
Data centres need to be geographically distributed. You wouldn’t want all of Europe’s data stored in, say, Ireland or the Nordics — one outage, regulatory change, or disaster could be catastrophic. The UK is an important node in that global resilience network.
So, while energy costs are high here, the strategic, regulatory, and market advantages make the UK worth it for Google and others.
👉 Do you want me to dig up the actual Google UK data centre investments (e.g. their recent £800m North London project) and show how they’re structuring the energy side of it?
Yes Please.Here’s a breakdown of what Google is doing in its recent UK data-centre investments — especially around the Waltham Cross site — and how it’s trying to deal with the energy cost / emissions / supply issues. Let me know if you want this in table form or broken down region by region.
What we do know from recent announcements
£5 billion investment over two years
Google is committing about £5bn (≈ US$6.8bn) in the UK over two years for capital expenditure (capex), engineering, R&D, including its AI arm DeepMind.
PR Newswire+2technologymagazine.com+2 Part of this is to open a new data centre in Waltham Cross, Hertfordshire.
Energy / Carbon‐free / Clean energy goals
Google aims for its UK operations to run on ≈95% carbon-free energy by 2026.
They’ve signed a deal with Shell Energy Europe Ltd, to act as their “24/7 Carbon-Free Energy Manager” in the UK. Shell will manage a power portfolio, including battery storage systems, to smooth out the intermittency of clean energy generation.
One long-term power supply piece: Google has a contract (or contracts) with ENGIE for Moray West, a large offshore wind project in Scotland.
Design for efficiency / environmental impact
The Waltham Cross data centre uses advanced air-cooling technology to reduce water usage (i.e. keep it nearer to domestic levels).
It is equipped to recover surplus heat: that is, redirect heat generated by the data centre to local homes, schools or businesses.
Energy supply balancing, resilience
Use of battery energy storage systems (BESS) is central. These allow periods of over-generation (when renewable sources are producing more than needed) to store excess, then release when generation is low. Helps flatten out supply variability.
Shell’s role includes optimising the clean energy portfolio, handling intermittency, and using trading / grid balancing expertise.
What this says about handling the “high energy cost” problem
From what Google has announced, they are trying to mitigate the risks and costs associated with high energy (especially electricity) in these ways:
Long-term renewable energy contracts (e.g. wind farms) to lock in lower, more predictable costs. Renewable energy might have high upfront cost, but long-term PPAs (power purchase agreements) can provide predictability.
Using storage (batteries) to avoid buying expensive electricity at peak times, and to reduce waste when there’s a surplus.
Efficiency measures: cooling technologies, heat reuse, designs that reduce some of the operational costs.
Working with local/regional infrastructure and partners (e.g. Shell, ENGIE) so that supply chain, grid stability, regulatory incentives (tax, subsidies, planning) help reduce risk.
Some of the open / less clear parts
While 95% carbon-free is ambitious, “carbon-free” doesn’t always mean “cheap” or “stable cost” — sometimes renewables + storage still cost more, depending on grid pricing, taxes, etc.
The remaining ~5% (or more, depending on hour by hour) may require procurement from non-renewable or more volatile sources, or use grid power when renewables aren’t producing enough. That carries both cost and emissions implications.
Grid capacity & infrastructure: high electricity demand from large data centres can stress local grids. If upgrades or reinforcements are needed, that adds cost (both to the operator, and possibly passed on via tariffs). There may also be political / regulatory / community pushback (e.g. concerns about emissions, land use).
Timing risk: some technologies (e.g. storage, heat recovery) may take time to build, regulatory approvals to get. Energy markets (wholesale, taxes, carbon pricing) can shift in the meantime, altering cost calculations.
If you like, I can try estimating whether, under current UK electricity prices (wholesale + transmission + taxes etc.), Google’s setup is likely to profit vs be forced to absorb high energy costs. Do you want me to pull that together?
OK
It’s hard to be super precise (because Google doesn’t publicly break out all UK costs, and prices & operating conditions vary by site), but we can sketch a rough estimate to see whether Google’s UK data centres are likely to be profitable despite high energy costs. I’ll lay out the assumptions, ballpark numbers, and where the risk lies. Estimate of cost vs revenue (very rough model)
Here’s one way to think about it: pick a hypothetical large data centre, estimate its power demand & energy cost, then compare that against what revenue or savings it might generate (or cost avoidances).
ParameterTypical / Assumed ValueExplanation / SourcePower demand (IT + overhead) ~ 100 MW (megawatts) Large hyperscale data centres are often in this ballpark.
Power usage effectiveness (PUE) ~ 1.2–1.5 PUE is a measure of how much extra energy (cooling, power losses, etc.) is needed beyond just the servers. Google claims its PUE is better than industry average, reducing overhead.
Effective total consumption ~ 120–150 MW If overhead is ~20-50% above IT load. So 100 MW of computing might mean 120-150 MW drawn from grid.
Hours of operation per year ~ 8,760 hours Running continuously.
Electricity cost ~ £0.25-£0.33 per kWh (or more) for industrial users UK non-residential / industrial electricity prices are often in the region of 24 p/kWh or higher, up to ~33-p p/kWh for some data centre operators.
Calculations
Annual energy usage = ~ demand × hours
If total draw is 130 MW average, then per year:
130 MW × 8,760 h ≈ 1,139,000 MWh (≈1.14 TWh).
Cost of electricity per year = energy usage × electricity price
Using e.g. £0.30/kWh (30 p/kWh) as a rough UK data centre industrial rate in high cost areas:
1.14 million MWh × 1,000 kWh/MWh × £0.30 = £342 million/year just for power input (grid cost, before overheads, maintenance, staff, etc.).
Add other costs: cooling infrastructure, staff, land, taxes, compliance, etc. These aren’t negligible. But for now, focus on power, since that’s the big variable.
Revenue / Benefit Side
What do you get in return? Some potential revenue / cost avoidance streams for a company like Google:
Selling cloud / compute services to businesses, public sector, etc. (servers, storage, AI compute).
Having local data-sovereignty, lower latency for UK/EU customers => premium charge / greater adoption.
Efficiency gains, lower transport costs / network latency could translate into competitive advantage.
If Google can charge, say, £0.10-£0.20+ per compute unit (or equivalent of usage, storage etc.), the revenue from customers may exceed the electricity & operational costs significantly — especially because these big firms often scale well.
Is it Profitable under UK conditions?
Putting those together:
If electricity cost ~ £300-£350 million/year for power, plus perhaps another modest chunk for cooling / overheads, then total operating cost might be, say, £400-£500 million/year for a large site (for simplicity).
If the revenue from lease / compute usage / services is, say, £800 million/year or more (which is plausible for a large hyperscale / AI support facility), there’s a sizable margin.
Key Risks / Variables that Could Tilt It Either Direction
Electricity price volatility
Even a few pence/kWh variation (say from 25p to 35p) shifts cost by tens of millions per year. That’s material.
Contract / renewable agreements
If Google can lock in long-term PPAs (power purchase agreements) or get favorable tariffs, that lowers avg cost. The higher the fraction of energy from renewables, wind, etc., the more stable (and sometimes cheaper) the cost becomes over time.
Grid constraints and ancillary costs / taxes
In the UK, energy isn't just the wholesale cost: there are network charges, transmission, levies, environmental taxes, and sometimes peak usage tariffs. These add on.
Efficiency (PUE etc.)
If the overhead (cooling etc.) is low, operations are efficient, then less energy is wasted; every percentage point of inefficiency avoided is saving money.
Revenue per compute unit / usage intensity
If demand is strong, margins good, pricing favorable, the facility can generate more revenue per MW served. If demand lags, or pricing is depressed (due to competition), revenue may be less.
Regulation and policy risk
Carbon taxes, net-zero requirements, planning constraints, local opposition etc. If new regulation inflicts additional cost, that reduces profit.
Conclusion
Given all that, here’s a summary:
Even with high UK electricity prices, a large, efficient data centre operated by a major company like Google can be profitable — especially if they secure favourable energy contracts, use renewable energy, push efficiency, and have strong demand.
The margins are more sensitive to electricity price than many other sectors: a rise in energy cost can eat significantly into profits.
So Google’s strategy of investing in renewables, battery storage, local generation, and efficiency isn’t just green optics — it’s economically necessary.
THX AI, nice to chat!
Scribble Painting no5 by iServalan Exclusive to Tale Teller Club Records Throw Pillow
From £15.68
Scribble Painting no 1 by iServalan Exclusive to Tale Teller Club Records Classic Mug
From £10.70
Scribble Painting no 2 by iServalan Exclusive to Tale Teller Club Records Classic Mug
From £10.70
Scribble Painting no 3 by iServalan Exclusive to Tale Teller Club Records Magnet
£8.20
Copy of Scribble Painting no 4 by iServalan Exclusive to Tale Teller Club Records Throw Blanket
£36.37
Scribble Painting no 6 by iServalan Exclusive to Tale Teller Club Records Apron
£22.88
Scribble Painting no 7 by iServalan Exclusive to Tale Teller Club Records Classic Mug
From £10.70
Sunflower Shining on a Sunny Day by Sarnia de la Mare Clock
£30.09
Blue Jesus by Scribble Artist illustrator Sarnia de la Mare Throw Pillow
From £15.68
Purple Rain Scribble Painting by Sarnia Throw Blanket
£36.37
An Afternoon in the Park, scribble painting by Sarnia Throw Pillow
From £15.68
After Titian’s by Sarnia, scribble painting mandala portrait Throw Pillow
From £15.68
