Coinbase, Elevance and Nvidia After the 2026 Midterms: What a Democratic Sweep Changes

Democrats won both chambers on Tuesday, including a Senate seat in Texas, and investors spent the next two days trading it as if the new Congress could write laws. Health insurers rallied on ACA subsidies coming back. Chip stocks had their worst day in about six and a half years on talk of a crackdown on data centers. I don't think either move matches how this Congress works. Trump keeps his veto until January 2029, and Democrats are nowhere near the two-thirds it takes to override him.

So here's what I'm changing in how I read the names I cover, and what I'm leaving alone. Coinbase and Robinhood lose the crypto bill until 2029, and that loss is real. Elevance already got the one thing a Democratic House can give it. And the data-center threat I'd take seriously came from county ballots and the Texas governor's race, which don't need Trump's signature, more than from Congress.

Kalshi traders had Democrats at 75% to take the Senate on November 2. They've won 54 seats so far, including Texas, where James Talarico beat Ken Paxton. Two races are still counting, per the Yahoo News live blog. Mary Peltola leads Dan Sullivan in Alaska by about 600 votes, and the ranked-choice count can't start before November 18. Scott Colom leads Cindy Hyde-Smith in Mississippi by under 2,000 votes.

Even if Democrats win both, 56 seats is short of the 60 it takes to end a filibuster. It's further still from the 67 needed to override a veto. The House majority, on track to be the largest since 2008, is short of two-thirds too. Trump has said he'd work with Democrats but wouldn't hesitate to use the veto. That leaves a Democratic Congress three things it can do without him: block bills, withhold the votes every spending bill needs, which means the threat of a shutdown, and hold up nominees in the Senate. Everything else on the post-election trade lists needs his signature.

The CLARITY Act, which would split crypto oversight between the SEC and the CFTC, got only 49 votes on a procedural motion on September 15, per 24/7 Wall St.. It needed 60. Its backers were counting on the lame duck. Stifel's Brian Gardner told Yahoo Finance it's "off the table for the lame duck session" and expects Elizabeth Warren, in line to chair the Senate Banking Committee, to investigate crypto firms. Robinhood fell about 12% on Thursday, per Yahoo Finance, and Coinbase fell 4.1% in early trading.

I'd take the crypto-law upside out of both stocks until 2029. A Democratic majority leader decides what gets floor time, and I don't see CLARITY getting it. That doesn't touch the case for Coinbase as I've written it, which rests on state licenses in every state and custody for most US spot bitcoin ETFs, none of which needed the bill. Robinhood has a second problem. My Robinhood analysis already names payment for order flow and prediction markets as a regulatory risk after the midterms, and Warren running Banking makes that more likely. That analysis predates Robinhood's last two quarterly reports, so I'll refresh it before changing anything there.

The insurers and hospitals that rallied hardest were the ones most tied to Medicaid and the ACA exchanges, per Medical Daily. Reuters gave two reasons: Democrats could delay the 2025 law's Medicaid cuts, and could help extend the enhanced ACA subsidies that expired at the end of 2025. J.P. Morgan told Reuters the veto would more likely lead to gridlock than to sweeping reform.

I'd split those. My read is that the part that's real is the block: Republicans can't run another reconciliation bill, the simple-majority route they used for the 2025 Medicaid cuts, without the House, and that helps Elevance's large Medicaid book. Delaying the 2025 cuts or bringing the subsidies back is a new law, and Trump has to sign it. The Medicaid work requirements start on January 1, 2027 in most states unless Congress acts. Centene, Molina and the hospitals rallied as if that deal were coming. Elevance rose the least of the Medicaid insurers, which looks to me like a price on the block and not much more, so I'm not changing anything on it.

Rules for pharmacy benefit managers are different, because Trump has signed one already. The Consolidated Appropriations Act he signed on February 3 moved PBMs' Medicare drug business to flat fees. Extending that to commercial plans would hit CVS, which owns Caremark, and UnitedHealth through Optum Rx. It's a win both parties can claim, so it's the health bill I'd watch for in the year-end spending deal.

On Thursday the Nasdaq fell 5.52% and closed in a correction. Chip stocks had their worst day in about six and a half years, and Nvidia dropped 9.3%, per CNN. Yahoo Finance framed it as Democrats preparing a crackdown on data centers, and some Democrats are pushing in that direction: Rep. Yassamin Ansari posted that the first-day agenda should include a pause on advanced AI development, which would hit data-center demand. Any national pause would have to be passed as a law, Trump has promoted data-center building, and I don't see him signing one.

The limits that actually passed were local. Ballotpedia counted 41 local data-center measures on November 3, and voters sided with limits in 40. They approved five full bans and twelve partial bans, mostly in Ohio and Kansas, three six-month moratoria in Maine, and voter-approval rules for big projects in places like Grove City, Ohio. Texas elected Gina Hinojosa governor. She campaigned on a temporary data-center moratorium and on ending data-center tax breaks, and Governor Abbott's current pause came as a directive to the grid regulators, per the Texas Tribune, so she can likely keep it going without a new law.

So I'm keeping the risk on the list and changing where I look for it. For Nvidia and Amazon, a ban in one county mostly moves a campus to another one. It cuts chip orders only if enough places say no at the same time that the build itself slows. The figure I'll watch is the 2027 capital spending guidance Amazon, Microsoft, Alphabet and Meta give with their fourth-quarter results in late January and early February. If the four together guide below what they're spending this year, the backlash bit faster than I think.

Power is where local siting cuts closer. Vistra sells into Texas, and my Vistra analysis already takes a growth hit for the Texas pause and Hinojosa's pledge. Constellation has more Texas exposure than its nuclear reputation suggests, because Calpine's gas fleet there sells into the same grid as Vistra's. My analysis already names the pause in its key risk, and I'm leaving the severity where it is, since most of its contracted book is nuclear in the PJM grid and New York.

Lockheed is the one name where Democrats hold real leverage without new law. Spending bills still need Trump's signature, but they also need Democratic votes, and the Iran war has to be paid for. JD Vance put the loss down to it on X: "The reason we lost is the war," per the Yahoo News live blog. I'd expect more stopgap funding bills and conditions on war money, and maybe a shutdown fight next fall, which would delay contract money more than cancel it. F-35 sustainment and the missile lines have no second source, so I don't see programs going away.

The strongest argument against me is that stocks price years out, and this result makes a Democratic trifecta in January 2029 a lot likelier. A Democratic president with a Senate this size could pass health care, crypto and AI rules with no veto in the way. If you own Coinbase or Nvidia for ten years, that's fair to price now, and some of this week's moves are probably about 2029 more than 2027.

Trump might also sign things. He's said he'll work with Democrats, Jefferies told Reuters that his comments open the door to Democratic legislation, and the incoming House leadership is already talking about an affordability agenda. A deal on ACA subsidies is the likeliest cross-party health bill there is, and the same shutdown leverage I expect Democrats to use on war money could carry it. If they attach subsidies or a delay to the Medicaid cuts to a must-pass spending bill, Trump has to choose between signing it and a shutdown. If he signs, Centene's rally was right and my Elevance call was too cautious, and the falsifiable claim below says so.

Local limits can also add up. Forty of 41 isn't a small sample, and 72% of Americans oppose a data center in their own community, per a Washington Post, ABC and Ipsos poll cited by Yahoo. If enough counties and a state the size of Texas say no in the same year, the build slows whatever Congress does. A Democratic Senate can also stall Trump's picks for the agencies that approve power hookups and oversee crypto, which matters even without a new law.

And some of Thursday may have little to do with Congress. Ed Yardeni told CNN, "We want a slowdown in the AI buildout. It was too frenzied and speculative." If the election just gave holders a reason to sell something they already thought was expensive, the lesson is about valuation, and my veto argument doesn't speak to that.

Before January 3, I'll watch whether the outgoing Senate takes up CLARITY, which Stifel doesn't expect, and whether commercial PBM language makes it into the year-end spending deal. On January 1 the Medicaid work requirements start unless Congress acts, which is the first real test for Elevance and Centene. Texas's legislature convenes on January 12, and Hinojosa takes office after that. In late January and early February come the 2027 capex guides from the four big cloud spenders, and the Medicare Advantage rate proposal for 2028, which will move Humana and UnitedHealth more than this Congress will.

For the rest of the names here, I'm not changing anything because of the vote. The two things I am changing are small: no crypto-law upside for Coinbase and Robinhood until 2029, and local siting treated as a real but slow risk on the data-center build.