US yields just hit a 19-year high: these 5 stocks could take the biggest hit

US yields just hit a 19-year high: these 5 stocks could take the biggest hit
Devesh Kumar
15-Sept-2026, 14:08 PM

powered by

Invezz
Nvidia (NVDA)

Buy Nvidia. Even with 10-year yields above 5%, NVDA’s current AI compute dominance (about 80% share cited) supports earnings power, so the market can reprice the multiple without breaking the business. It’s the cleanest way to stay long the AI capex cycle while others get hit harder by duration risk.

Key Risk: AI demand weakens enough that investors cut near-term revenue/earnings expectations, not just the valuation multiple.

CoreWeave (CLOU)

Sell CoreWeave. It’s the most rate-sensitive AI infrastructure name: it needs massive capital for data centers and just took an $8.5B loan, so higher borrowing costs hit both valuation and funding flexibility. The article flags contracted but not yet sold power capacity—exactly where a financing squeeze can force renegotiations or slower buildout.

Key Risk: A fast refinancing/cheaper debt window that removes the funding squeeze and lets contracted capacity convert to revenue on schedule.

  • A 5% Treasury yield is raising the bar for Wall Street’s growth trade now.
  • Capital-heavy AI plays face a tougher test as borrowing costs stay elevated.
  • Long-dated growth stories risk multiple compression even without EPS cuts.

A 5% Treasury yield is becoming another challenge for Wall Street’s growth stocks.

The US 10-year yield climbed to 5.03% on Tuesday, its highest since 2007, as expensive oil, inflation concerns and expectations for further Federal Reserve tightening drove another bond selloff.

The move follows calls to slow frontier-model development that hit AI-linked stocks.

CoreWeave stock faces a financing squeeze

CoreWeave stock may be the clearest example because higher rates affect both its valuation and its operating model.

The AI-cloud provider requires enormous capital to build data-centre capacity. It secured an $8.5 billion loan in March, taking its equity and debt commitments over the previous year to about $28 billion.

That makes elevated borrowing costs relevant just as investors debate whether slower frontier-model training could soften infrastructure demand.

“The most exposed name in our coverage is CoreWeave,” Bernstein analyst Madison Rezaei told Yahoo Finance, citing the company’s exposure to contracted but not yet sold power capacity.

Marvell stock's $30 billion opportunity is years away

Marvell’s case shows how higher yields can hurt a stock even when the underlying AI opportunity remains intact.

Bank of America analyst Vivek Arya highlighted a $30 billion opportunity tied to Marvell’s custom-chip connectivity and networking portfolio by the end of the decade.

That long runway is also what creates duration risk. Investors are paying today for profits expected closer to 2030.

When Treasuries yield more than 5%, those distant cash flows become less valuable in present-value terms, raising the return investors can demand before paying premium multiples.

AMD stock must clear a higher valuation hurdle

AMD’s investment case increasingly depends on share gains across server CPUs, AI accelerators and agentic-computing infrastructure.

Stifel analyst Ruben Roy said AMD remained “confident and, notably, more expansive on server CPU than on GPU,” after meetings with company executives, according to Investing.com.

Stifel kept its Buy rating and $635 target.

That confidence supports the long-term thesis, but it does not eliminate rate sensitivity.

If investors can earn more than 5% in Treasuries, AMD may need to provide stronger evidence that forecast AI earnings and market-share gains will arrive on schedule to justify a premium valuation.

Nvidia stock can be repriced without weaker earnings

Nvidia is different because its current AI business remains exceptionally strong.

Piper Sandler analyst David O’Connor called Nvidia the “outright leader in AI compute” in commentary carried by TipRanks, citing an estimated 80% market share and continued supply constraints.

Yet strong fundamentals do not make the shares immune to higher yields. Nvidia is still the market’s clearest vehicle for expressing confidence in a long AI-capex cycle.

If the discount rate rises, investors can assign a lower multiple to those future earnings without analysts cutting near-term revenue forecasts.

Tesla stock shows this is broader than chips

Tesla demonstrates why the issue extends beyond semiconductors.

Morgan Stanley analyst Andrew Percoco recently described Tesla as a “credible emerging competitor in autonomous trucking.”

The bank sees long-term potential from autonomous Semi software and other mobility businesses.

Those opportunities, alongside robotaxis and robotics, sit years into the future.

Higher yields therefore reduce the present value of Tesla’s future profit streams while potentially keeping consumer auto-financing costs elevated.