Data Centers Q2 Preview

AI Demand is Here, but Largely Accruing to Private Players - We've tracked >50 MWs of AI leasing in Q2 from smaller GPU-based platforms like CoreWeave, Denvr Dataworks and Lambda Labs. There are also multiple 100-200 MW hyperscale requirements in the market, signifying demand remains very strong. Over the next 5 years, we estimate GPU deployments could equate to 1.5-2.0 GWs of annual demand, or 25-50% of total industry leasing for the third-party market (in addition to hyperscale self-builds).

However, we think DLR leasing will still be relatively light vs. 2022 in Q2 due to limited inventory and capital constraints on new builds. There may be retrofit opportunities for existing data centers longer term, although those could equate to ~$3MM/MW by our rough estimates to install advanced liquid cooling technologies. EQIX's opportunity on AI inference demand and incremental connectivity is likely longer term in nature, with more near-term activity with AI training investment.

Pricing / Limited Supply Remains Core Bull Case - We have heard that rents on new leases are up 10-20% on average across the U.S. (and more in certain tight markets like Ashburn), with in-place vacancy at an all-time low of ~3% across top U.S. markets. We could see upside to re-leasing spreads for DLR, particularly for >1 MW bookings. Power constraints in N. VA and California are starting to spread to other markets where lead times are elongating, including Chicago, Phoenix, and Quincy.

EQIX: Downgrade to Equal Weight, Maintaining PT - We are downgrading EQIX to Equal Weight but maintaining our $810 price target. It is a crowded long with less room for valuation upside - stock trades at +25x NTM AFFO, or a yield spread of 10 bps over the 10-year UST (near an all-time low). We expect guidance to be revised slightly higher, although that's reflected in consensus (1% above mid-point of guide). We still favor EQIX longer-term, particularly given the strength of its balance sheet.

DLR: Maintain Some Caution into Q2 - We maintain a more muted view on DLR after the recent run, based on (a) potential future equity issuance, with stabilized asset cap rates notably higher than its cost of equity; (b) likely downside to '24 ests, with another ~$1.5B likely needed next year from asset sales/JVs/equity; and (c) yr/yr leasing slowdown with many key markets constrained on supply. These risks are offset by improving rental rates and renewal spreads, which remain a tailwind.