Weekly Research Briefing: Bluebonnets in November

The Cook Political Report’s 2026 House Vote Tracker follows the national House popular vote as ballots are counted after the November 3 midterms. Counting can take weeks, especially in states that count mail ballots late. To give the most accurate picture of the 2026 results, the tracker runs until all 50 states certify.

First, federal support for basic scientific research is critical for expanding the economy and generates high rates of return. The government should also support applied research centres around the country to encourage commercial applications of new ideas. AI-assisted research will be important in driving future productivity growth, so reforms that allow modernisation and the expansion of electricity transmission will be vital.

Second, a renewal of support for open markets is needed. While present policy can justify exceptions to openness for reasons of national security or to protect critical supply chains, broad interference in trade is counter-productive. Across-the-board tariffs raise the cost of important intermediate goods for US industry, and studies conclude that American consumers and firms have borne most of the costs…

Third, while acting aggressively against illegal immigration is warranted, broad limits on high-skilled immigration — as evidenced in recent attacks on H-1B visas for skilled workers or optional practical training authorisation for non-US students to gain work experience in technical jobs — are inimical to growth.

Hedge funds dumped technology stocks at one of the fastest paces on record last week, and the momentum crash that followed was the biggest in about 40 years. Even so, few on Wall Street are ready to call time on the AI boom.

Goldman Sachs Group Inc. noted that hedge fund disposals of global technology stocks roughly matched the highest on record in figures going back to 2016. Tech was by far the most heavily net-sold sector globally and in the US, driving the biggest wave of single-stock net selling in four months, according to the bank’s prime brokerage data as of Nov. 6.

The trigger was the Democratic sweep in the US midterm elections, which raised concerns about tighter AI regulation and about how much data center capacity will actually get built. The Nasdaq 100 posted the deepest plunge since the Covid pandemic on Thursday, wiping $2.3 trillion off its value in a day. The S&P 500 notched its heaviest weekly loss since April 2025.

After a brief reprieve in September, Momentum-factor performance deteriorated sharply in October, triggering its largest single-month underperformance in over 40 yrs (see Figure 12, Long vs market, cap-wtd), and precipitated by two +6 sigma unwind days over the latter half of the month. This rapid unwind has compounded the sharp losses from July’s Asia-led Tech deleveraging, bringing the cumulative drawdown to the largest since ChatGPT’s launch, or -38.4% from the June peak. As a result, Momentum crowding has retreated from 100%ile to 76%ile vs. 40 yr history (breaching the same level as the Feb–Mar 2025 unwind and now at the same level as the 2024 tech selloff). This marks the steepest decline in our crowding metric since the GFC recovery.

What are the potential positives coming out of the midterms? For one, oil prices have come down sharply, aided by comments from the White House that markets interpreted as a sign that a deal with Iran could come together, though this remains uncertain. Second, our policy strategists have flagged the potential for a reversal of the SNAP and Medicaid cuts given the midterms outcome, which could be a tailwind for consumers. Third, we may see a pull-forward of business activity and M&A before the 2028 presidential election. Once again, the markets traded poorly into the midterms, as they typically do. This is one reason why they generally trade well in the 6–12 months after the midterms—buy the rumor, sell the news.

Inside HY the median bond is at the 39th percentile of its own 2-year range while 14% of bonds sit at 2yr wides (a tail repricing, not a broad one). The widening is concentrated in Hyperscalers / Mega-cap Tech, Semiconductors.

AI Infrastructure represents c.15% of the STOXX 600 and only 2pp of its expected 9% earnings growth in 2027, making the region a better place to find diversified stock performance than the more AI-sensitive S&P 500.

Nonresidential construction spending has been rising for the last six-months, up 12.6% SAAR over this period. Data centers have been responsible for much of the growth but growth outside of data center construction appears to have stabilized somewhat.

"The states where Democrats are likely to further restrict data centers account for roughly 45% of planned US data center capacity, with Texas alone accounting for roughly 25%. However, since data centers are low on the list of priorities for voters this election, we expect that the issue will fade somewhat in importance in the ensuing months and that these moratoria will not be long-lived if they are enacted.” - Jan Hatzius, Goldman Sachs

The biggest market issue may be AI over the medium term; nearly every major local ballot measure opposing data centers won or is winning, with only exception in North Dakota. What we’re now watching: governor imposed constraints on new builds. This is not an issue for the 37 GW expected to be delivered in 2027, 75% of which is under construction. It’s more an issue for the 96 GW expected in 2028. Based on election results, 37.5 GW may be subject to a “pause” with another 15 GW subject to “conditions”. While pauses are the more impactful outcome, conditions can be onerous: requiring data centers to pay grid costs to protect ratepayers, ending tax incentives, imposing water limits, setting requirements for onsite backup power, mandating labor requirements and banning non-disclosure agreements during permitting (most of which seem sensible to me). If the data center buildout starts to slow at some point, upgrading to better NVIDIA chips might be the only way to generate more tokens.

Alphabet Inc.’s Google will spend $17 billion to build a data center in Wyoming, formally putting a price tag on the proposal two days after voters across the US delivered a broad rebuke to the development of such massive projects for artificial intelligence work…

Google said Thursday in a blog post that the facility in Laramie County, Wyoming, will be its first in the state. The company said it will pay for all the power the data center uses. Last month, American Electric Power Co. said it planned to invest $4 billion in the project, which will draw electricity in part from a 702-megawatt natural gas fuel cell development built with Bloom Energy Corp. fuel cells, and will be air-cooled to limit water use…

Google also said it expects to generate hundreds of millions of dollars in local property taxes and will form a Community Council with its partners to address growth concerns in Cheyenne, Wyoming.

Bloomberg reported in June that data center developer Crusoe had been pushed aside at the site after Google raised concerns about the costs and timetable, according to people familiar with the situation. Energy utility Black Hills Corp., which had been working to line up agreements to power the project, said it was moving forward without Crusoe. Google on Thursday named Black Hills Energy as a partner in the Wyoming project.

“I can also tell you that every hyperscaler is now thinking about ‘Do we need to build outside the United States?’ Because they’re enormous investments, hedging those investments is a wise thing to do” - Meta President Dina McCormick

“I do believe strongly that if the US hesitates in any way, shape, or form in terms of data center build-out, the rest of the world is ready and willing. There is just huge demand for this capacity” - Arm CEO Rene Haas

Banks and credit investors are upping their scrutiny of AI-related debt deals while showing reluctance to increase exposure to borrowers, jolted by the recent market selloff and changing sentiment toward the technology's growth trajectory.

Some banks, which just a few weeks ago were lining up to fund AI deals, are now telling market participants they would be reluctant to finance additional deals even when they involve guarantees from the sector's marquee technology companies, according to several banking and credit sources.

The lending pullback has been exacerbated by a recent drop in the stock market, which erased $2.5 trillion in market value on Thursday from the Nasdaq 100, and sent the index into a correction. The US midterm election results are also adding anxiety about more oversight of data centers.

A potential slowdown in AI data center construction could curb loan growth and dent investment banking fees from the loans, bond sales and stock offerings that have helped finance the sector's rapid expansion.

The change, according to two sources, has come in just the past few days. Some lenders have shown reluctance to finance deals even if they are backstopped by names such as Anthropic and Google, the sources said, without specifying names of the lenders.

Impact to data center builds & capex: expect a ~20% delay of US projects from '28 to '29, but remain constructive on the overall build-out. We believe newly elected Governors are unlikely to disrupt projects under construction, leaving 2027 projects unchanged. However, we believe some 2028 TX/OH projects will be delayed to 2029, as governors work with legislators to put in place new requirements for data center developers. We shift 5GW of US projects from '28 to '29 (representing 20% of US projects, 10% of global projects), which implies US data center GW additions down 4% Y/Y in 2028 and up 38% Y/Y in '29. Globally, this implies data center additions up 9% Y/Y in '28 and 20% Y/Y in '29…

For now, we keep our hyperscaler capex forecasts unchanged (in aggregate $1.6T/$1.8T in '27/'28) as we still expect the hyperscalers to spend and build as much capacity as they can to create and scale new products to billions of users enabled by large language models that are increasingly compute-intensive but still carry high 30%+ ROIC (see here). We also expect hyperscalers to attempt to adjust their build locations to other states that are more data center friendly (like Tennessee, Louisiana, and others).

Roughly a third of SpaceX shares are now part of the company's public float, up from 5% at IPO. Based on SpaceX's lockup schedule, about half of its shares should be part of the public float by the middle of 2027, consistent with the historical precedent from large IPOs with low initial floats.