Wall Street's smart money is decisively bullish on data center infrastructure, with TD Cowen's latest supply chain checks describing record hyperscaler leasing led by Meta Platforms, Microsoft and Alphabet's Google, while Morgan Stanley has warned clients that the memory shortage will intensify into 2027 and 2028. That view lands after weeks of selling in data center names, creating a contrarian setup where institutional conviction and current tape are diverging.
"Our checks indicate a record ~9.6GW of 2Q26 DC leasing led by META, MSFT, and GOOG, with a record ~12.5GW leasing pipeline as OpenAI upsizes its roadmap to 30GW by 2030," TD Cowen analysts wrote in a note this week. The Wall Street Journal separately reported that OpenAI now plans to spend $750 billion on data centers through 2030, up from a prior plan of $600 billion, while Anthropic is ramping more than 1GW in international requirements.
Morgan Stanley analyst Joseph Moore, in a note relayed by commentator Tae Kim, reinforced the supply-side tightness after meetings with data center procurement managers. "The intensity of the shortages in that part of the business show no signs of abating," Moore wrote. "We see prices up at least 25 percent on a like-for-like basis from 2Q to 3Q, above our estimates and above third-party estimates. As importantly, the longer-term concerns that the memory shortage will intensify in 2027 and again in 2028 are still as strong as ever."
The data center buildout is now the single dominant driver of memory demand, a dynamic Moore described as "unusual" because it creates mixed signals in consumer electronics, PC and smartphone markets that periodically drag down spot prices and stock prices. He called the recent pullback in memory stocks a buying opportunity, saying memory is "quickly closing the gap" with Nvidia and Broadcom on risk-versus-reward.
The Numbers Behind the Buildout
Vertiv, the power and cooling infrastructure supplier, reported Q1 2026 revenue of $2.649 billion, up 30 percent year over year, with a $15 billion backlog and organic orders up 252 percent. The stock carries a mean analyst target of $379.20 across 22 buy or strong-buy ratings, yet has fallen 9 percent over the past month.
Micron Technology delivered Q3 FY2026 revenue of $41.456 billion, up 346 percent year over year, with gross margins of 85 percent and guided Q4 revenue of $50 billion. The consensus price target stands at $1,491.95 across 40 buy or strong-buy ratings. The stock sold off 14 percent in the month before rebounding this week.
Equinix logged record annualized gross bookings of $474 million, with 60 percent of its largest deals AI-driven. Its mean price target is $1,199.66, with Morgan Stanley recently raising its own target to $1,075. Digital Realty Trust booked a 200-megawatt AI inference lease, the largest hyperscale deal in its history, and runs roughly 3.0 gigawatts in place with about 6.3 gigawatts buildable.
CoreWeave, the highest-risk name in the group, rents AI compute against a $99.4 billion revenue backlog but posted a widening net loss and $7.7 billion in Q1 capital expenditures. A securities fraud class action adds legal overhang, and the stock has slid 26 percent in a month.
What the Gap Between Targets and Tape Means
Every major data center name has recently pulled back. Vertiv is down 9 percent over the past month, Equinix off 6 percent, Digital Realty down 5 percent. The Global X Data Center & Digital Infrastructure ETF bundles Equinix and Digital Realty with chip names for diversified exposure, with Equinix at 14 percent and Digital Realty at 13 percent of the fund.
The bear case rests on stretched valuations and the risk that analyst projections may not fully play out. But the evidence behind TD Cowen's leasing figures and Morgan Stanley's memory checks is quantitative, current and consistent with what these companies are reporting themselves. For investors weighing the AI infrastructure thesis, the setup is a compression of price against improving fundamentals and hardening analyst conviction — a gap that typically resolves in one direction.
This article is for informational purposes only and does not constitute investment advice.