Coming over all a bit Enron…
“Nobody ever questioned whether Enron would go down the pan 20 years ago.” So says one source this week when talking about a market juxtaposition that’s got a few lenders twitching about the aggregation of risk across the data centre and renewable energy sectors.
As another lender of some considerable renown (if dubious turn of phrase) puts it: “In the midst of the real estate recession, risk departments are waking up to the infra bubble and buttocks are starting to clench.”
Meanwhile, one with a leaning towards the flippant adds: “ESG reporting will ensure sufficient demand for data centres to keep the demand for power so high that we’ll be switching coal back on when the sun isn’t shining or the wind blowing!”
So, what’s got the lending community’s goat right now?
Well, talking to folk around the industry this week, an interesting concern has risen to the surface as lenders have been supporting projects with hyperscalers as the tenants / offtakers while also – awks – financing renewable energy projects with these same names as offtakers.
While not everyone agrees that this is an “Enron moment” and some are of the view that it’s not an issue in the slightest… there’s a school of thought that the stack of cards is looking a whole lot less stable than it did… prior to noticing this aggregation of risk.
Now, in an ideal world, the left hand would always know what the right hand’s doing. In this instance, that equates to the infra team knowing what their colleagues on the energy team are up to.
That seems fair. But if you think back to the sub-prime crisis, it’s clear this frequently is not the case as – back in 2008 – it took banks months to work out their exposure to this shonky real estate play… that the real estate team had been told to shun – a message that never quite made it to other departments.
And – from what was being said this week – we’re back here again as lenders follow the thread back from blue chip clients like Microsoft, Google and Amazon to a bearded hippie in a field with a windmill (OK – so that’s a slight exaggeration).
Capital concentration
As to what causes this concentration of capital, one old chum puts it rather well: “Infra banks in general have been quite bad in last few years at looking at concentration risk. This is caused by the lack of project finance deal flow where risks are standalone save for counterparty risk. Along with low margins, these have led to big tickets in correlated asset classes.”
So, the bottom line is that it’s a problem of the lenders’ own making in a market where pricing has been driven down and there’s a lack of communication across teams.
“Some lenders are just too siloed and deal teams don’t necessarily look beyond their sector remit. Particularly so if individual team budget targets have to be met,” says another lending luminary.
“That’s when lenders start burning the candle at multiple ends which then causes many red faces when a key link of the chain turns out to be much weaker than what has been presented to credit.”
The luminary continues to illuminate: “In principle, there’s nothing wrong with the combo of renewable energy generation and data centres. They form a great and important symbiosis. Just be aware of the credit risk of the counterparties.
“As a power generator you would need to ask yourself whether you would want to enter into a power offtake contract with a data centre SPV which has a very limited balance sheet. Mitigating factors could be appropriate security support or an offtake contract with an investment grade entity would generally be more suitable for project finance.”
Yet another off-record commentator points out that correlated, one directional risk across infra and energy goes even deeper than just this instance, but says it serves as a good case study of a broader issue.
“One thing I’ve noticed is that consolidation across pools of capital within the space with the rise of private credit behemoths buying insurance and LifeCo businesses is now creating a dynamic where a previously disaggregated market for lending is becoming much more concentrated,” they say.
“The bank market is one example but even the USPP universe where I used to sniff around for tickets like a truffle pig, looking for cheques of $25-30 million – maybe max $75 million – now is dominated by 10 accounts that won’t get out of bed for less than $100-125 million.
“On the portfolio side there, they might be syndicating some of that risk down, but the model for a lot of them is aggregation and making asset management fees. In a sector where risk concentration is already problematic you better hope they have decent controls and monitoring… but I certainly have concerns.”
When you dig around for pipeline numbers to hang around a story like this, the most commonly thrown figure is $1 trillion of investment into data centres over the next 5 years – essentially to 2030. That’s from Dell'Oro Group and it’s just a little to “rounded-up” for comfort… but it’s likely in the right ballpark. It does also say that AWS, Google Cloud, Meta and Microsoft will take up around half of global spending by 2026 – which feels about right too.
The amount of energy these data centres snarf is rather alarming. And while it’s all as green as the grass you skip upon, do bear in mind that they have huge, diesel-fired back-up generators for each of these sites with enough fuel to buy time for the supply to be reinstated… or for them to ship in enough go-go juice to keep that highly-polluting generator going.
One source of a legal persuasion identifies a rather interesting data point – data centres in the US are projected to take 10% of ALL power generated in the nation by 2030.
“That’s going to give rise to demand issues,” says the lawyer. “Captive power plant is not a bad solution. BlackRock and Microsoft both seem to be looking at their own power solutions.”
Meanwhile an old friend of the technical advisory type says: “I think there’s a hope the market gets frothy again in 2025 so they can punt the problems on to others. However, the music may have stopped and there will be more write downs by equity as we’ve seen at Thames Water.”
Oh, Thames Water… as the investors like to ask: “One lump or two?”
Darkest before dawn
As mentioned earlier in the piece – not everyone spoken to this week is a doom and gloom merchant. Some say it’s a storm in a teacup and recommend that knickers remain untwisted.
As one says: “That logic could be used for virtually any segment of the economy. Lending to a car plant and the power plants providing it power for example. We are actively recognising that data centres are creating lending opportunities along the chain.
“So what if you’ve lent to a power plant feeding the data centre. At least there’s an alternate use for the power plant.
“The bigger risk is the concentration of risk in the data centre user – typically Google, Microsoft, X, AWS, etc. That’s what people ought to be worried about. Are the offtakers almost the same across my ‘portfolio’ of data centres such that I don’t really have a ‘portfolio’ any more?”
Meanwhile, another asks: “Is there any example of genuine project-on-project risk where there’s a direct wire from a non-recourse solar to non-recourse data centre? Not to my knowledge.
“Far more cases are synthetic and that means the solar is grid connected for alternative route of revenue generation and / or the hyperscaler balance sheet is an intermediary between the 2 projects.”
So that’ll be just fine then…
Request a Demo
Interested in IJGlobal? Request a demo to discuss a trial with a member of our team. Talk to the team to explore the value of our asset and transaction databases, our market-leading news, league tables and much more.