The opportunity in data centres explained

The concept confounds many an investor, and market commentators too; from the moment the market understands there's a whole lotta growth up for grabs on the horizon, share prices move to above market-average multiples which, at face value, makes the stocks in question look 'expensive'.

But are they really? Is the opportunity already gone?

In many cases the answer is: no, the opportunity is still there. If the promised growth comes through, and management executes on the opportunities available, there's often plenty of room left for upside surprises, which, when looking back with hindsight, only makes that share price from the past, 'bloated' though it may have looked in-the-moment, actually a cheap 'bargain'.

One such prime example from the recent past has been delivered by healthcare imaging services provider, Pro Medicus (PME) whose share price has been trading on forward-looking multiples above 100x. But as the emerging global leader in its field added new contract after new contract, it forced analysts to regularly upgrade already bullish forecasts, with forward-valuations and price targets rising further on the back of it.

In simple terms: Pro Medicus shares looked 'expensive' back in 2020, when the price crossed the $30 mark. Last week they surged above $110. And while most analysts have valuations that are well-below that price level, Macquarie believes they are being too conservative. AI and new services and customers are still on the horizon. Macquarie has set a price target of $120. For now.

Goldman Sachs sits on $134.

In investment terms, Pro Medicus shares gained 52.8% in 2020, followed by 82.8% in 2021, then retreated -13.4% in 2022, only to advance a further 77.2% in 2023 and, thus far in 2024, another 15%. That's one helluva return for a stock that along the way attracted comments ranging from 'absolutely crazy', to 'bubble', to 'egregiously overvalued'.

For good measure: nothing from the past four years indicates this company will not encounter a growth hiccup at some stage, or worse, and its shares might well come back down to earth if that happens. But shouldn't investors equally be aware that such concerns, and calls of 'the next example of irrational exuberance', have been expressed way too liberally, and way, way too early?

For all we know, this company is nowhere near to about to run out of growth. Certainly, management at the helm thinks so. Thus far any weakness in the share price has been but an opportunity to get on board.

Never ask a barber whether you need a haircut. Never ask a value investor whether to invest into the next emerging growth opportunity.

Data Centres Are In Strong Demand

A similar dilemma has opened up with the emergence of generative artificial intelligence ('GenAI'), mostly in the US, and in the slipstream of the next tech (r)evolution, the surging demand for data centres.

Already, fund managers have been taking profits on their Goodman Group (GMG) shares that have appreciated by 45.7% in 2023, and by a further 33.6% in the first three months of 2024. Shares in NextDC (NXT) have gained 52.7% and 29.6% respectively and have been described as 'overvalued'.

But what if both companies are still only at the early stage of a strong demand growth period that has many more years to run? Might those elevated valuations in the here and now mirror Pro Medicus shares from years past?

A fresh research update by analysts at Morgan Stanley is certainly challenging all who are questioning the ongoing opportunity on offer.

With the local market for data centres to more than double by 2030 (150% projected growth), it's rather difficult not to expect a whole lot more upside for companies leveraged to that demand, assuming, of course, management teams execute and the global landscape does not come irreparably unstuck, like through war or much higher bond yields.

Three things make this research exercise unique:

-it incorporates the latest updates and insights from US companies, today's heartland of GenAI and data centres

-the research is a deep, multi-disciplinary collaboration between analysts across technology, media and telcos (TMT), REITs, Utilities & Sustainability, and Mining/Resources sectors

-the research is uniquely focused on the Australian market

Let's start with the basic outcomes. Morgan Stanley's 12-month price target for data centres operator NextDC has been raised by 13% to $20 (Monday's share price $17), while the target for Goodman Group has lifted to $35.30 ($33.94 on Monday). The target for Macquarie Technology (MAQ) is $100 ($84 on Monday).

Supporting these upgrades is US feedback that growth in AI and GenAI, and the associated rapid rise in demand for computing power, is accelerating. This, in return, boosts demand for data centre capacity. Conclusion: a golden period has opened up for companies such as the three mentioned. Those worried about capacity catching up will have to wait many more years, all else remaining equal.

Not many investors would be aware, but Australia already is a global Top Five data centres hub, with capacity similar to London, but lagging the US, Europe, the Data Centre Alley of North Virginia, and Beijing/Shanghai.

Morgan Stanley's current projections imply additional investments made in new data centres will total between $21bn-$28bn over the next eight years, providing companies with an incremental revenue opportunity of $5.6bn-$8.4bn per annum. These numbers, states the report, could well prove conservative.

One major challenge Australia faces is the additional stress on local power supply networks from this robust surge in demand. Data centres, in particular the larger and more expensive hyperscalers, are power-hungry beasts. The report estimates data centres currently consume circa 5% of Australia's total power generation. This is forecast to increase to between 8%-15% by 2030.

Can the country manage this? Yes, states Morgan Stanley, but by 2030 network limitations and problems might start to impact. As most data centre operators opt for 'green' energy when offered, there's an obvious catalyst for more green energy and renewables, as well as, maybe, for uranium (as is already the case internationally).

Both AGL Energy (AGL) and Origin Energy (ORG) are expected to benefit from the extra demand for electricity. In terms of increased demand for commodities, the collaboration has identified copper as the most likely greatest beneficiary from building additional data centres; not just locally, but globally.

More data centres should equally result in additional demand for lithium (batteries) but Morgan Stanley's projections suggest lithium will continue to suffer from market surpluses until 2027, so no enthusiasm here. Similarly, any increase in the popularity of uranium reactors won't be felt until into the next decade.

For good measure: today's developments and forecasts made remain a moveable feast, and Morgan Stanley's research report acknowledges as much. Investments into AI and IT generally could slow down, in case of an elongated economic recession, for instance, or due to other, unforeseen causes.

A scenario that pushes up global bond yields to much higher levels could derail everything, in particular if those bond yields would then remain (much) higher-for-longer, because access to finance and future returns from investments would be negatively impacted.

Other potential threats include energy and power capacity constraints, overinvestments that create a surplus in capacity, which then leads to price discounting and lower sector returns, supply chain disruptions causing delays, but also: hyperscalers that build enough capacity for themselves and stop leasing capacity from third party data centres.

As things are stacked up right now, it looks like the bias leans towards much higher demand growth for data centres globally, for years to come, and Australian companies should be among major beneficiaries. There are, however, key differences in how companies might grab the opportunity.

A brief summation:

NextDC is part of the local Top Three for data centre capacity, with the two key competitors not listed. AirTrunk is currently owned by Macquarie Asset Management (MQG) and PSP Investments. A consortium led by MAM took a controlling interest in 2020, at that time valuing the company at $3bn. Currently the shareholders are reportedly looking for a buyer who's willing to pay $12bn.

Canberra Data Centres (CDC) remains equally unlisted to date. International industry giant Equinix owns and operates a network of 260 data centres in 71 major metros around the globe. To date, Equinix's capacity locally makes it the fourth largest player in Australia.

NextDC is expanding its footprint locally, with more expansion plans under consideration, and has already moved to New Zealand, Japan and Malaysia. Its business model is to lease out capacity to both businesses (consumers) and other data centre owners; hyperscalers such as Microsoft, Google and Amazon.

NextDC's future profits depend on management's ability to access fresh capital, acquire land and build centres, then fill capacity with new customers and contracts, in an environment of much higher costs than in the past. Morgan Stanley forecasts an internal rate of return of 12% which remains well above the estimated weighted average cost of fresh capital at 7.7%.

As different types of customers pay different prices, future return estimates are at best an educated guess, with any difference in mix altering the numbers. Data centre operators are able to pass on higher power costs. Were the broker's bull case scenario to unfold, the value of NextDC shares increases to $28 instead of $20.

A research report issued by stockbroker Morgans in April suggested a positive outcome in the years ahead could well catapult shares in NextDC to $40 in seven years' time.

Goodman Group has many more options available as it can finance and develop new centres on its own, in cooperation with third parties/customers, or on behalf of data centre owners and users. Generally speaking, the options that require larger upfront investments offer higher returns, but not as quickly as less-costly developments.

How much of the pie will descend into shareholders' pockets will depend on which options are chosen, and when projects might be finished, etc. The key attraction for Goodman shareholders is data centres raise the company's profit margin, and this already has triggered a re-rating for the shares.

Morgan Stanley can see more than 25% upside to current market consensus EBITDA projections by FY28, and it'll be all about data centres becoming a more important source of returns for the company. Goodman Group is generating high margins on data centre developments thanks to its pre-purchased land bank.

Morgan Stanley estimates Goodman Group can derive $20bn in additional value from its pipeline of data centre developments in the years ahead. It is equally assumed the company can fund its share of developments without additional capital, and there's always room to sell some assets, if need be.

Were the broker's bull case scenario to unfold, Morgan Stanley's valuation for Goodman Group shares would climb to $45.70.

In the wake of the data centres deep dive (82 pages), Morgan Stanley analysts reiterated their view shares in Macquarie Technology are undervalued, trading at around $84 on Monday. The broker's price target is $100. Macquarie Technology is not a pure play, combining data centres with more traditional telecommunication operations which also include a cloud & government business which is equally growing strongly, points out the broker.

Companies not mentioned in the report that might also benefit from demand for additional data centres in the years ahead include Global Data Centre Group (GDC), which is part owner of AirTrunk through the Macquarie-led consortium, and Megaport (MP1).

As was widely anticipated, Goodman Group management has upgraded FY24 EPS guidance to growth of 13% this week, having already upgraded to 11% growth in February with the release of interim financials.

As analysts at Macquarie noted in the wake of this latest upgrade: Goodman Group management has upgraded guidance at the March quarter update each of the past three years, and beaten guidance at the full-year result in each of the last six years.

FNArena offers independent and impartial market commentary and analysis, on top of proprietary data, tools, and analysis. The service can be trialed at (VIEW LINK)