Morgan Stanley Says the AI Sell-Off After Democrats' Blue Wave Is Overdone — Here Are Three Buy-the-Dip Plays

A midterm rout spooked AI bulls over data-center permitting risk. Morgan Stanley argues the sell-off creates entry points across hyperscalers, supply-chain picks-and-shovels, and software adopters.

The 2026 midterm elections delivered a Blue Wave — Democrats swept Congress and captured key governorships including Texas — and AI stocks cratered on the fear that tighter permitting and environmental review would choke off data-center buildout. The Nasdaq 100 posted its worst single-day drop since April 2025. Semiconductors, neocloud operators, and hyperscalers all got hammered.

Morgan Stanley's equity analyst Stephen Byrd pushed back in a November 9 client note, calling the market reaction "overdone." His core thesis: even if Democratic control slows the pace of AI infrastructure expansion, a longer capex cycle benefits incumbent suppliers and eases supply-chain bottlenecks that had been creating margin pressure. A slower rollout does not mean a smaller rollout. The bank's first bucket targets hyperscalers with the balance sheets and vertical integration to ride out permitting delays. Microsoft, Amazon, Meta, Alphabet, and SpaceX all receive overweight ratings. Byrd's logic is straightforward: when infrastructure constraints tighten, scale advantages compound. The biggest players capture share from smaller competitors who cannot absorb delays. The second bucket is the picks-and-shovels layer — companies building the physical substrate of AI. Morgan Stanley sorts these into powered-shell providers (Terawulf, Riot Platforms, Hut 8, Applied Digital) that have low Texas regulatory exposure, behind-the-meter power generators (Bloom Energy, Vistra, Talen Energy, Constellation Energy, and others) insulated from grid-level slowdowns, energy infrastructure plays (Williams Companies, Clearway Energy, Fervo Energy), and semiconductor/hardware names (Nvidia, Broadcom, TSMC, Arista Networks). The selection criterion: high current margins and historically low valuations. The third bucket is software enablers and AI adopters — the layer that historically outperforms last in a tech capex cycle. Morgan Stanley flags Snowflake (74.4% upside to price target), Datadog (54.6%), ServiceNow (53.6%), Cloudflare (51.3%), and Palantir (50.3%) as top software picks. On the adopter side: Devon Energy (54.7%), Grindr (48%), GXO Logistics (41%), Citigroup (38%), Blackstone (38%), and Citizens Financial Group (38%). The implicit framing is that this is a dislocation trade, not a fundamental reassessment. Nothing about AI demand, model capability, or enterprise adoption changed on election night. What changed was the expected velocity of physical buildout. Morgan Stanley is betting that the market conflated pace with magnitude — that investors priced in cancellation risk when they should have priced in extension risk. The note is useful as a map of sell-side consensus on the AI stack, but readers should note what it does not address: the actual regulatory mechanisms a Democratic Congress might deploy, whether permitting timelines have any historical correlation with final buildout volumes, or whether the current capex cycle already shows signs of overinvestment independent of politics. This is a buy-the-dip recommendation dressed as structural analysis.