What is the Point of the GB Electricity System?
Does it make sense to connect wind farms on the Shetland Islands?
N.B. - this was first posted on the 1st of February 2026. I have edited it in response to reader feedback (8th of February 2026) Briefly, within the transmission charging regime there are some sizeable charges for localised transmission infrastructure - so the original piece likely exaggerated the extent to which system costs aren’t fully considered when projects are built. I have now included some detail around this. The relevant text will be in bold italics.
January has been busy – particularly with the recent results of the AR7 offshore wind round, published on the 14th.
It has been a month reflection for me – I have been thinking about the UK electricity system - wondering who it serves and to what purpose?
Does it serve to provide its end customers with cheap, clean and reliable electricity? Or has it just become a giant, expensive web of bad incentives?
I am a natural optimist but have felt my cynicism rising of late.
I am trying to analyse energy in the UK to help improve policy. My writing reflects my personal views. None of the content should be construed as investment advice. I have done my best to ensure that the content below is accurate – but I am human and will make mistakes – if you spot any, please let me know and I shall update as appropriate.
In early December I read reports from two consultancies that suggested AR7 offshore wind prices around £94/MWh (2024 prices) could be cost neutral for consumers.
I read both and disagreed with the conclusions. I felt they missed important points:
GB electricity prices are already very high – cost neutral isn’t good!
We are signing 20-year CPI linked contracts. Using 2030 or 2035 as your point of cost assessment misses the impact of CPI build up over time
I felt the reports failed to consider the full impact of transmission/balancing costs
When the results of AR7 for offshore wind emerged, (at c. £91/MWh in 2024 terms), the £94/MWh ‘benchmark’ established a month earlier was used by many to frame AR7 as some kind of victory for the GB consumer.
We need to start having a more honest debate. If proponents want to justify the AR7 contracts on decarbonisation or energy security grounds, then make that argument. It is much more honourable than hiding behind opaque reports that were commissioned by the beneficiaries of AR7.
But back to the electricity system – who does it serve and to what purpose?
System level thinking
UK energy policy doesn’t fully consider the wider electricity system when comparing different projects and technologies. Debate has descended to comparing the costs of power generation alone.
This article uses the Viking Wind Farm on the Shetland Islands (top right on the map below) and the Shetland HVDC undersea connection to illustrate the lack of system level thinking.
The Shetland Islands: Plenty of wind
The Shetland islands are an archipelago, lying about 110 miles from Northern Scotland, with a population of around 23,000.1
Prior to 2024, the islands were separate from the Main Interconnected Transmission System of Great Britain. Electricity was provided by three groups of assets:2
Lerwick power station (c. 50%) - a 72.8 MW diesel fired generation station
Sullom Voe Terminal power station (c. 30%) – a gas fired station that primarily powers the Sullom Voe gas terminal
A series of small renewable generation assets (c. 20%)
The Shetland Islands are one of the windiest parts of the UK. This article from the Met Office lists the Shetlands as the county with the highest average wind speed from 1981-2010.
What’s the problem?
The windiest part of the UK sounds like a great place to build an onshore wind farm. The March 2021 accounts for Viking Energy Wind Farm LLP anticipate load factors of around 48% on their 443 MW wind farm project.
48% is a remarkable load factor for an onshore wind firm – the UK onshore average was 25.7% in 2024.3 DESNZ assume 37% for onshore projects commissioning from 2035-2050.4
I added up the capex costs for the 443 MW Viking Wind Farm (from Companies House accounts) and expressed the values in 2024 real terms. I got to £730m or about £1.65m per MW.
The 2025 Generation Cost report from DESNZ assumes capital costs of about £1.42m per MW for an equivalent sized onshore wind project commissioning in 2030 – so this appears to be in the right capital cost ballpark.5
This all looks good – windy location, high assumed load factors and reasonable capex numbers. But there is an issue.
Building a 160 mile undersea cable to Northern Scotland isn’t cheap!
Expensive connections
The 600 MW Shetland HVDC connection was given a capital cost allowance of £641.8m by Ofgem in 2020/21 values, or c. £790m updated to 2024 terms.6
This equates to a little over £1.3m per MW of capacity. If you consider the two projects together, then the total for the wind farm AND interconnector is more like £2.95m per MW.
******************************** (8th Feb 2026 Addition) ********************************
To some extent, the Viking project will have internalised some of the costs of this localised transmission infrastructure. Transmission Network Use of System (TNUoS) charges for generators already vary by region (and tend to be highest in North Scotland) but some projects pay specific charges for their localised infrastructure.
The charges for Kergord (The converter station where Viking connects) are substantial - and should amount to around £26m for 2024/5 on an annualised basis assuming 443 MW of transmission entry capacity. If the wind farm was delivering a 48% load factor, then this would amount to around £14/MWh of transmission charges for the local circuit.
****************************************************************************************************
The interconnector will also allow the islands to have cleaner power via imports when wind output is low. The diesel fired power station will move to a backup role.
On the other hand, it is important to note that the Shetland HVDC cable delivers the power to the North of Scotland, which is already a very constrained part of the transmission network.
The excellent WindTable website suggests that Viking is one of the most curtailed wind farms in Great Britain despite the Shetland HVDC link, with over 60% of its cumulative potential output curtailed by the end of December 2025.
To explain curtailment briefly: If the electricity generated in a certain part of the country (e.g. Northern Scotland) can’t reach demand centres further south, due to a transmission bottleneck, the system operator uses the balancing mechanism to make sure that power is delivered where it is needed.
Generators north of the bottleneck can bid to generate less electricity, and generators to the south will offer to generate more. Crucially the costs of balancing are socialised - electricity consumers pick up the tab via Balancing Service Use of System (BSUoS) charges.
I should explain that it isn’t just its location that causes Viking to curtail so much of its volume. Viking doesn’t appear to have started it’s CfD contract yet, and may have a lower opportunity cost to participating in the balancing mechanism than other generators. Viking was awarded a CfD in two chunks in AR4 and AR5, with a weighted average strike price of c. £68.80 per MWh (2024 prices) across the two rounds.
Returning to the original quandary – if building transmission was free, then Viking looks like a great wind farm - reasonable construction costs and built in a windy location.
But in the real world, laying transmission cables is expensive. We can’t just build renewable projects wherever it is sunniest or windiest. The costs of moving the power must be part of the equation.
(8th of Feb addition) - To some extent, the high regional and localised transmission charges are doing their job in pricing some of the costs of transmission build out. Whether the charges are large enough or predictable enough is beyond the scope of this post. However, the socialised nature of balancing costs looks to be a bigger issue.
The Cost Benefit Analysis:
There are three major transmission owners in the GB market. They are regulated monopolies and recoup their costs from billpayers. They can’t just go out and spend hundreds of millions on a new project without prior approval.
The Shetland HVDC transmission project was brought forward under the “Strategic Wider Works” framework to handle the delivery of large network investments. This process of regulatory oversight is particularly important in this instance - as both the Shetland HVDC link and Viking Wind Farm are ultimately owned by SSE.
As part of a long running approval process, there were multiple pieces of evidence associated with the Shetland HVDC project. Some were fairly positive - for instance this heavily redacted 2020 levelised cost analysis from National Grid ESO (the system operator), which concluded that the net cost of building the transmission project and associated generation compared favourably to an “equivalent offshore wind farm”.7
“when net HVDC capital costs are included within the onshore windfarm options, the 600 and 800MW HVDC link options, combined with the 638, 705 or 810MW generation capacity, produce lower LCOEs than the equivalent offshore windfarm.”8
However, National Grid ESO also produced a cost benefit analysis for the transmission project here. Whilst the numbers were heavily redacted, the descriptive text wasn’t.
There are two elements I want to draw out from the report that illustrate some of the dynamics at play in considering full cost system costs.
The analytical framework that was used
The scenarios where the project had a negative Net Present Value (NPV)
I should mention that the analysis was designed to find the optimal connection capacity and start date, as well as presenting evidence on value for money for consumers.
Cost Benefit Analysis: The Framework
The analysis used a ‘savings approach’ to compare the costs of building the project vs a counterfactual or base case scenario. In this case, the counterfactual is as follows:
“No new link to the mainland is built, and any excess generation on the island, is constrained off by the System Operator”9
Greater detail is offered in the screenshot below:
Essentially, the report assumes that a load of generation capacity is built on the Shetlands. It then analyses whether it is better to build a cable to connect the Shetlands to the mainland, or just pay up and constrain all of the excess generation.
Firstly, its seems kind of wild that this question needs asking - I think it offends most people’s intuition. Constraining lots of electricity feels like a waste - but then again, spending hundreds of millions building a transmission cable to deliver power to the very North of Scotland doesn’t sound hyper logical either.
Secondly, the framework isn’t assessing whether the generation and transmission projects in combination make sense for the consumer. The report is written in April 2020, months before construction work started on the wind farm. The analysis assumes the generation capacity already exists - we are debating whether it is worth connecting or not!
Cost Benefit Analysis: Net Present Value
The report ultimately weighs up the costs of building the transmission cable vs. the costs of constraining all of the excess power. The costs of the former are somewhat known, but there are multiple drivers for the constraint costs.
The cost of constraining each MWh: More expensive = more constraint spend
The level of electricity consumption on the Shetlands: More on-island consumption = less constraint spend
How much generation capacity is built: More generation = more constraint spend
I have tried to display all the variables and cases from the report in a single diagram below. In all cases I have assumed the mid sized 600MW interconnector - the size that ended up being built.
Cases 1 and 2 used ROC based constraint cost assumptions, whereas cases 3 and 4 used CfD.10 The report states that the CfD assumptions will result in lower constraint costs relative to ROCs.
I don’t know why ROCs were used in any of the cases - the ROC scheme closed to new entrants three years before this report was published.11 I would focus on cases 3 and 4 which used CfD based estimates for constraint costs.
The two final variables are:
Whether new demand is created on the islands by electrifying oil and gas infrastructure. More on-island electricity consumption means less constraint, and thus favours leaving the Shetlands unconnected.
The amount of new generation capacity added - more generation capacity means more constraint and thus favours building the transmission.
The NPV numbers themselves are redacted, but the text describing each case isn’t. I have coloured the boxes red where the text logically states that the NPV is negative, and green in all other cases. For example, the language for Case 4 states that NPV’s are negative for all scenarios.
This is already a long post, but focussing on case 3 is instructive.
To my mind, the analysis is saying that the Viking Wind Farm alone (443 MW) isn’t enough to justify building the transmission - you have to assume much more generation gets added in future to make building the cable worthwhile.
Ultimately, many factors will have gone into the decision to approve the Shetland projects that I can’t cover in a post - e.g. improved energy security for the islands and possible savings in building out the islands distribution network. But hopefully this case study illustrates some of the costs in building transmission to very remote parts of Great Britain. I have summarised five lessons below.
Building wind farms in the windiest places isn’t always sensible - it might be great for the project owner but bad for consumers
We have to consider the full system costs when comparing different technologies and projects
It is rational to waste some wind - electricity generation increases at a cube function of wind speed. Building the grid to capture every MWh isn’t sensible
We shouldn’t socialise the system costs of antisocial projects - zonal wholesale pricing would help with this, as would changes to the apportionment of balancing costs.
Be aware of incentives - if regulated returns are attractive, participants will look to deploy capital. Don’t assume they care about consumers or the wider system.
Our electricity system should be about providing cheap, clean and reliable power.
We have a long way to go.
ssen.co.uk website - accessed 30th Jan 2026
2025 Generation Costs (DESNZ) - Annex A: additional estimates and key assumptions. 2030 medium scenario, also expressed in 2024 values. Includes pre development, construction and infrastructure costs.
Ofgem: Shetland HVDC Link - Project Assessment Decision - 30th November 2021. CPI adjustment factor of 1.23 to adjust from calendar 2020 values to calendar 2024 values.
“Net” cost of building the transmission project is used to account for savings on capital spending at the local distribution level that would have been incurred without the buildout of the Shetland HVDC scheme.
ROC = Renewable Obligation Certificate. CfD = Contract for Difference.
As a yank trying to understand your situation - under curtailment the project operator gets paid something for the power not delivered? So it sounds like there is a bottleneck moving power out of northern Scotland where the majority of the wind production potential is to the south where the demand is located. Any idea of the total annual curtailment costs imposed by this bottleneck, and what would be the payback period for adding transmission to break the bottleneck? And a final dumb question - why was this much capacity built in the Shetlands if there was no means to get it to market?
Interesting article! Couple of points though:
Generation doesn't rise with the cube of the wind speed: it will at some point but then plateaus at the turbines "rated wind speed".
I also think its slightly unfair to ascribe all the costs of interconnector to Viking when there's some unused headroom on the interconnector: if you account for that the capex looks more similar to an offshore wind farm (approx £2.5m/MW). Bear in mind Viking is capable of getting offshore wind-esque load factors (assumed to be 48%).