Morgan Stanley says buy the AI dip, sees $3.4 trillion in hyperscaler spending through 2028

Morgan Stanley is urging investors to buy the recent weakness in AI stocks, arguing that the market has overreacted to the US midterm election results and their implications for data center construction.In a report titled Buy the Drop: AI Investment Opportunities Post the US Midterms, the bank says that the AI investment cycle remains intact, even if political opposition to data centers pushes some construction projects into 2029.The bank's hyperscaler capex forecasts remain unchanged at $1.6 trillion in 2027 and $1.8 trillion in 2028, although it flags some downside risk to the latter.Key takeaways:Morgan Stanley expects 20% of US data center projects planned for 2028 to be delayed until 2029, representing 5 gigawatts of capacity.US data center capacity additions are now expected to decline 4% in 2028 before jumping 38% in 2029.Globally, data center additions are forecast to grow 9% in 2028 and 20% in 2029.A shortage of more than 30% in powered data center capacity is expected to persist through 2028.MS favors Amazon, Meta, Alphabet and Microsoft among the hyperscalers.It warns that parts of the AI supply chain could face slower orders, inventory accumulation and margin pressure in 2028.The concern is largely centered on Texas and Ohio, where newly elected governors could introduce additional requirements for data center construction. Morgan Stanley doesn't expect projects already under construction to be disrupted, leaving its 2027 forecasts unchanged.Instead, it expects permitting and regulatory delays to push some 2028 projects into 2029. I suspect there were also some bottlenecks developing around power that would have pushed them back in any case.Further out, MS sees a potential geographical shift in construction, with hyperscalers moving some projects to more accommodating states such as Tennessee and Louisiana. The bank also sees the potential for as much as 20 GW of additional data center development across Asia and EMEA.Morgan Stanley argues that the delays could ultimately be healthy for the AI investment cycle, smoothing the supply-demand imbalance, supporting lease rates for existing compute capacity and extending the period of investment.The distinction is between hyperscalers and their suppliers.The bank is particularly bullish on the largest technology companies, which have the balance sheets, existing infrastructure and technical capabilities to overcome power constraints.Its preferred names include:Amazon (AMZN), Meta (META) and Alphabet (GOOGL): Overweight, with advantages in scale, capital access and AI development.Microsoft (MSFT): Overweight, with similar advantages.CoreWeave (CRWV): Overweight.Nvidia (NVDA): Overweight, with an important role in improving the amount of AI output generated per watt of power.All of them are up 2-8% today, with the exception of Nvidia, which is down 6% after the Jensen Huang honeypot allegations. CoreWeave reports on Tuesday and the leader, up 8%.Morgan Stanley highlights the potential for dramatic efficiency improvements in AI models, including Meta's target of a tenfold increase in compute efficiency in 2027.The bank argues that advances in model architecture, distillation, training techniques and specialized chips could allow AI applications to scale much faster without corresponding increases in computing requirements.Meanwhile, the outlook is more complicated for companies supplying the data center construction boom.If the bottleneck shifts from equipment availability to powered buildings, suppliers could experience reduced pricing power, rising inventories and slowing orders. Morgan Stanley warns that the sector could see a traditional downcycle in 2028, including margin normalization and earnings downgrades, even as underlying AI demand remains strong.That makes stock selection increasingly important.Among infrastructure-related opportunities, the bank highlights Equinix (EQIX) and powered-shell providers, including WULF, RIOT, HUT and APLD, which it sees as less exposed to Texas-related risks. It flags CIFR, GLXY and MARA as potentially more vulnerable to construction delays or moratoriums. All the mentioned names are up strongly (4-9%) except Cipher, which is down 2.2%.The financing outlook also favors established companies. Morgan Stanley expects AI-related debt issuance to rise in 2027, but says new data center developments and neocloud operators will face greater scrutiny from lenders.The broader message is that the AI buildout isn't ending, but the winners may be changing. The largest hyperscalers could benefit from tighter capacity, better compute efficiency and weaker competition, while some equipment suppliers face a more difficult transition as the construction timeline shifts.Morgan Stanley sees the recent selloff as an opportunity, but it's clearly not recommending buying every AI-related stock indiscriminately. Looking ahead, eyes are on the Anthropic IPO with a report on Friday saying it will go ahead this month. This article was written by Adam Button at investinglive.com.