By Christine Ji

Investors are concerned that Alphabet faces a fresh threat from Nvidia's plans to help fund the AI buildout alongside major Wall Street players

Alphabet shares are under pressure as more evidence builds that it will be challenging to take market share from Nvidia.

Alphabet's artificial-intelligence capabilities have received intense scrutiny in recent weeks as its latest Gemini model experiences delays. Now, another cornerstone of the company's AI strategy - its in-house hardware - is coming under question thanks to Nvidia's latest AI financing plans.

After falling 4% on Monday, Alphabet shares (GOOGL) (GOOG) shares were down 0.3% in Tuesday's trading. To some investors, Nvidia's (NVDA) $500 billion partnership with Wall Street firms to fund AI-infrastructure development poses a significant headwind to Alphabet's custom chips, according to Jefferies buy-side analyst Jeffrey Favuzza. By making Nvidia infrastructure available through favorable financing options, the deal could make custom silicon options less attractive for customers.

Nvidia's new partnerships with Apollo (APO), BlackRock (BLK), Blackstone (BX), Brookfield (BIPC), Goldman Sachs (GS) and KKR (KKR) aim to turn Nvidia's chips and AI infrastructure into an "investable asset class," the chip maker shared in a blog post earlier this week. These strategic partnerships will create dedicated pools of capital "at attractive rates for Nvidia customers."

It's the latest iteration of Nvidia's vendor-financing playbook, whereby the company has invested into various parts of the AI-infrastructure ecosystem that run on its graphics processing units. These strategies can help boost near-term demand and give customers more purchasing power than they could achieve alone.

Nvidia's moves are meant to help tighten its dominance in the AI ecosystem, and some investors are interpreting the latest $500 billion agreement as a "net negative for custom ASICs" such as Google's tensor processing units, Favuzza said.

Custom chips have grown in popularity in recent years as companies look for lower-cost alternatives to Nvidia chips. However, Nvidia CEO Jensen Huang has repeatedly emphasized that GPUs are better for AI workloads than custom ASICs, arguing that Nvidia has built a comprehensive compute platform beyond just chips.

The appeal of GPUs has also strengthened as the chips prove that their useful lives are significantly longer than the five-year depreciation schedule that many had been assuming. On CoreWeave's (CRWV) Tuesday earnings call, the company shared that it had signed a contract for Nvidia's A100 GPUs that extends to 2029, showing that Nvidia chips can reach a nearly 10-year lifespan. CoreWeave is a major strategic partner of Nvidia's, and Nvidia holds a nearly 13% stake in the neocloud company.

Google's tensor processing units have been a major selling point for the company's AI strategy. In addition to renting out its TPU chips in major compute deals with the likes of Anthropic and SpaceX (SPCX), the company has begun to sell its chips to third parties this year. But it's clear that taking market share from Nvidia won't be easy.

Additionally, recent departures and leadership shake-ups in the Google DeepMind division have added fresh uncertainty to the company's AI strategy, leading many to believe that the company is rapidly losing its reputation as a frontier lab. Google's latest Gemini 3.5 Pro model was originally slated for a June launch but still has yet to be released.

"It continues to feel like it's very challenging tactically for this to work when folks have a waning confidence in an updated Gemini 3.5 Pro model, coupled with the continued turnover in their DeepMind division," Favuzza noted.

-Christine Ji

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.