Sunrun has tried, and failed, to lift its slumping stock. The company made its biggest attempt yet on Wednesday as it unveiled an "AI compute pilot program" that will pay homeowners to power artificial-intelligence data centers directly in their homes.

To expand into distributed edge computing, Sunrun plans to install "compute nodes" -- essentially, miniature data centers -- inside homes using its solar and battery systems. The company says participants will be compensated for hosting the hardware.

The nodes are designed to process AI workloads, specifically inference, directly in residential neighborhoods rather than at massive server farms. Sunrun, in turn, will sell that AI-processing capacity to unnamed "enterprise compute buyers."

The company is positioning the program to meet data center operators' demand for compute power. By leveraging its 1.1 million-strong customer base, Sunrun can add significant inference capacity without the years required to permit and build new data centers.

The program marks an expansion of a "successful proof of concept that demonstrated revenue generation and high demand for distributed compute, " the company said in a press release.

This move follows Sunrun's increasing pivot toward the booming AI market, which presents an opportunity for the company to offset losses as its core residential solar business faces headwinds.

Sunrun -- the largest residential solar and battery storage company in the U. S. -- has grappled with a combination of macroeconomic pressures and policy shifts that have reduced consumer solar incentives.

Last month, the company unveiled a partnership with Tesla and Renew Home to turn residential homes into a connected, virtual power plant. Under the program, customers can store cheap solar or grid energy in batteries and automatically feed it back to the grid -- or lower their home's power use -- when electricity demand spikes.

Investors welcomed the development, driving shares up 13%. At the time, JPMorgan analysts noted that the model addressed the power market's most urgent issue: speed to interconnection. By utilizing flexible capacity, the program could help data centers and utilities avoid the lengthy timelines associated with traditional grid infrastructure.

The latest announcement didn't spark the same enthusiasm. After rising in premarket trading Wednesday, shares abruptly reversed course after the opening bell and slid 6.1%. The Nasdaq Composite traded flat.

The muted reaction likely stems from uncertainty over what this program actually means for Sunrun's bottom line, and whether the initiative will succeed at all. Sunrun is planning to complete the pilot over the coming months and evaluate the results before embarking on a broader rollout.

The bigger issue is that Wall Street isn't buying the AI turnaround narrative. Counting Wednesday's losses, shares have fallen 38% this year.

While the slump kicked off with a disappointing full-year outlook at the end of February, Sunrun's problems run deeper than missing investor expectations. The company continues to grapple with fundamental flaws in its capital-intensive "power-as-a-service" business model.

Sunrun relies on a highly leveraged funding strategy to support the upfront installation costs for its residential leased systems. While the company has successfully paid down chunks of debt in recent quarters, its total debt burden remains high.

In short, a flashy narrative won't be enough to distract investors. Even in a booming AI market, shiny new tech can't substitute for a sustainable bottom line.

Write to Mackenzie Tatananni at mackenzie.tatananni@barrons.com

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

(END) Dow Jones Newswires

July 08, 2026 14:41 ET (18:41 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.