Two Prices From One Election

On election night, the Nasdaq and the S&P 500 climbed on AI chipmakers as the returns came in ^[1]. The next session, CoreWeave fell 7.2% to $49.85, its unsecured 2031 notes at 79.5 cents on the dollar, and the junk bonds of Galaxy Digital, Cipher and SE Cosmos weakened ^[2]. Two prices from one election, moving in opposite directions within a day. The market hadn't changed its mind. It had been holding two separate questions all along. Washington, the first question, was settled before anyone voted. Trump had said he would not sign bills limiting AI expansion, and on election day he made clear what a Democratic win would change.

I may have to become "Mr. Veto" if Democrats win a majority — Donald Trump

JPMorgan's published model put a 66% probability on a Democratic sweep while assuming any resulting AI regulation would be vetoed ^[3]. Chips rising on election night was arithmetic, not complacency. So the second price was the real verdict, and it was not about Congress. It was not general risk-off — chips went up on the same day ^[1]. It was not macro — Meta sold $15 billion of Aa3-rated bonds the same week at premiums as low as 2.5 points over Treasuries ^[4]. What fell was the paper that depends on one site getting one wire. David Solomon named that channel weeks before the vote.

Localized political volatility around data centers. You know, we've got an election coming up. It'll be interesting to see. — David M. Solomon

Data center builders had parked billions of high-yield issuance until after the election ^[5], and the losses landed on exactly that paper — debt whose collateral is a building waiting on a connection, in states where the politics of the connection just changed. Voters passed 33 ballot measures restricting data centers and a single one allowing development, and Goldman's economists now expect short-term moratoria to touch roughly 45% of planned U.S. capacity expansion ^[6]. But the veto being repriced was neither new nor partisan. Weeks before the vote, a Republican exercised it. Greg Abbott — the governor who had marketed Texas as the "epicenter of AI development" — froze roughly 1,800 pending grid applications in August ^[7][8]. By late October, Morgan Stanley's plan to sell $5–7 billion of high-yield bonds refinancing the $15 billion Anthropic-linked data center in Hubbard, Texas, could not reach the market at all ^[9]. A hold on a wire stopped a bond sale, and there was no legislature anywhere in that transaction. No bill, no floor vote, no Washington veto pen — just a governor's desk and a frozen queue. The sweep didn't create this power. It ratified it, then handed it to someone who campaigned on keeping Abbott's halt ^[10]. That is what the bond market repriced: not whether the power exists, but who holds it now, in Texas and Georgia and Ohio. Some of the move was already in the price before any ballot was counted. CoreWeave had fallen 24% in the two weeks before the vote as rates climbed ^[11]; the sweep added a political layer to an existing squeeze rather than starting one. And the checks on the governor's hold were never in Washington. North Dakota's legislature rejected a moratorium outright in September ^[12]. In Ohio, the fight moves to a lame-duck session where pending bills could hand legislative leaders veto power over state agencies, hemming in the incoming administration ^[13]. The calendar turns the two prices into one question. The new governors take office in January 2027, and hedge fund managers warned on November 4 that a wave of private-credit maturities crests in 2027 and 2028, with syndicated loan default rates already above 2008 levels ^[14]. The industry will be refinancing across those desks in exactly those years, under exactly these governors. The prices set Tuesday and Wednesday were the market's first quote for that crossing.