ACA insurers drop as “beautiful” GOP law hits

Bitcoin-sensitive stocks tumbled Monday, enduring a much steeper drop than the keystone crypto asset itself, which was down nearly 4%, falling below $87,000, as of 12:20 p.m. ET.

Goldman Sachs’ themed basket of bitcoin-sensitive equities was down more than 8%. (It consists of companies tied to bitcoin, either through mining, digital payments, crypto investment, or blockchain technology.) It was one of the worst performers among Goldman’s thematically curated baskets of shares on Monday.

Nvidia’s “House of GPUs” is looking a little wobbly.

Shares of Applied Digital, CoreWeave, and Nebius — three of the four biggest equity positions held by the chip designer as of September 30 — are getting crushed on Monday.

Nvidia owned about $3.6 billion worth of these data center and neocloud stocks (with the overwhelming majority in CoreWeave) per its most recent 13F filing.

The AI credit risk that’s been most talked about in reference to Oracle’s widening credit default swaps spreads is also present in some of these firms, as well.

An Applied Digital bond due in 2030 is trading below $96 for the first time this month. That issuance was made to support data centers where CoreWeave will be the main tenant.

CoreWeave, which earlier this year received warrants enabling it to purchase a large chunk of Applied Digital shares as part of a data center leasing deal, sank last week after announcing a $2 billion convertible note offering that was later upsized.

Of course, it’s not just Nvidia-owned stocks, but the entire data center ecosystem that’s under pressure on Monday. Cipher Mining and IREN are also getting walloped — with Monday’s crypto tumble also likely weighing on these two bitcoin miners turned data center companies.

Analysts at Evercore ISI began coverage of AI energy play GE Vernova with an “outperform” rating and a price target of $860 on Monday, citing a number of reasons to be bullish about the maker of turbines used for power generation. Evercore’s price target implies gains of roughly 27% for the shares.

Analysts at the shop wrote of GE Vernova:

“1) Growth is strong and well supported by backlog in both Power and Electrification… with visibility into the 2030s. Despite headwinds from a shrinking Wind business, we see 12% CAGR 2026-28E, the strongest growth ex-Siemens Energy in our coverage.

2) Margin is expanding with operating leverage, pricing & productivity in both Power & Electrification. Full ownership of Prolec should drive another step up in estimate revisions upon closing (mid-2026). The equipment dynamics (pricing, margin expansion) should repeat in the service business 2030+.

3) Shareholder returns are very well supported, with EBITDA margins rising from 7% in 2024 to 21% in 2028 and FCF of >$5bn pa on average — recent buyback upgrade to $10bn (vs. $6bn prior) and dividend increase amplify an already attractive growth algorithm.”

There are some risks to the rally for the shares, which have more than doubled this year. For instance, the company’s struggling wind power division could weigh on results. Also, the high valuations on the stock — its forward price-to-earnings ratio is roughly 55x — make it vulnerable to rapidly shifting investor vibes toward AI, analysts say.

“Investment sentiment is tied up in the AI/Data centre cycle, so any suggestion of delays or diminished energy demand would weigh on the stock as investors would fear over-capacity,” they noted.

Nvidia extended gains in early trading after announcing an updated edition of its open models, the Nemotron 3.

This family of models comes in three “sizes”: Nano (available today), Super, and Ultra (both expected to be launched in the first half of next year). These sizes reflect the different parameters of each model, which govern the complexity of a given request it can handle.

The company highlighted the flexibility benefits of these models, saying they can be integrated with proprietary counterparts to produce cost savings.

“As multi-agent AI systems expand, developers are increasingly relying on proprietary models for state-of-the-art reasoning while using more efficient and customizable open models to drive down costs,” per the press release. “Routing tasks between frontier-level models and Nemotron in a single workflow gives agents the most intelligence while optimizing tokenomics.”

This strong start to the week helps reverse a substantial run of underperformance from Nvidia versus its peers. It’s the only member of the VanEck Semiconductor ETF that’s declined since the S&P 500 closed at an intermediate bottom on November 20.

Last week, the chip designer closed at its lowest level compared to this fund of 2025, falling below the trough seen in the wake of the DeepSeek freak-out, where nearly $600 billion in market cap was obliterated in a single session.