Tesla cuts Solar Roof: After the hardware tide recedes, how do software and services settle into long-term value

August 23, 2026 · Industry Observation · Source: Xiangming Technology Content Center

In late August 2026, Tesla officially removed the Solar Roof solar roof tiles from its product line, ending this hardware product that Musk once called "the future of building-integrated photovoltaics." Since its release in 2016, Solar Roof's cumulative shipments have never reached scale, its installation cost has long been about 2.5 times that of traditional photovoltaics, delivery cycles often stretch for months, and repair complaints remain high. A hardware category once full of promise thus takes its final bow. The signal it sends is far more than the failure of a single product; it is that energy technology companies are systematically contracting heavy-asset hardware and turning toward the value accumulation of software and services.

What is receding is not just the roof, but the "heavy hardware" delivery model

Solar Roof's failure is not an isolated case. While cutting roof tiles, Tesla is still expanding Powerwall energy storage batteries and Autobidder virtual power plant software. Behind this one cut and one increase is a clear expression of the same business logic: any hardware that requires deeply customized installation, long delivery cycles, and frequent on-site construction is being marginalized; any asset that can be shipped in a standardized way and continuously create value through software is being doubled down on.

From a technical perspective, Solar Roof's shortcomings are concentrated in three areas. The first is the supply chain: specially made glass tiles require dedicated production lines, yield ramps up slowly, and unit costs are hard to reduce. The second is construction: roof tiles must be installed piece by piece by certified installation crews, with cycles and labor costs far higher than panel photovoltaics. The third is operations and maintenance: replacing a cracked tile requires matching products from the same batch, and the after-sales chain is lengthy. These three constraints combined make it difficult for Solar Roof to replicate the "standardized product + large-scale factory" playbook of electric vehicles; in essence, it is a product that is "not superior in either software or hardware."

The combination of Powerwall and Autobidder is exactly the opposite: energy storage batteries are standardized hardware, while virtual power plant software aggregates dispersed battery resources into dispatchable grid assets, arbitraging between peak and off-peak electricity prices through algorithms and charging service fees. Tesla's energy business revenue grew by about 60% year over year in 2025, with the contribution share of software and services continuing to climb. When a company can use the same line of code to serve ten thousand households, rather than sending ten thousand construction crews to climb ten thousand roofs, the structural difference in its return on capital becomes obvious at a glance.

The value of software and services is hidden in the "long tail"

If we shift our gaze from the energy industry back to enterprise digitalization, this logic of "hardware receding, software sinking deeper" is almost everywhere. The traditional view holds that the value of hardware is one-time: once sold, delivery is complete. But truly sustainable profit often settles in the long tail after the device is sold—control, scheduling, operations and maintenance, data analysis, iterative upgrades.

Take the Internet of Things as an example. In a smart community solution, the hardware layer is merely sensors, gateways, and access control terminals, and gross margins have already been squeezed extremely low by the supply chain. The real value lies in the software platform behind it: whether devices can be uniformly connected, whether data can be analyzed in real time, whether anomalies can trigger automatic alerts, and whether energy and security can be coordinated and scheduled. For the devices across dozens of buildings in a community, manual inspections are a huge operational burden. Once capabilities such as device management, energy consumption monitoring, and visitor access are settled into cloud-based software services, operations and maintenance costs can drop by more than 40%, while forming sustainable annual subscription cash flow.

The same situation appears in the field of software development outsourcing. In the past, when a Shenzhen software company took on an app development project, delivery marked the end, and the customer relationship ended with it. Today, experienced teams turn the delivery endpoint into the starting point of service: after the app goes live, they continue to take on version iterations, backend operations and maintenance, data dashboards, and growth optimization, converting a one-time development fee into a long-cycle technical service contract. Hardware can become obsolete, and code can be refactored, but once a continuous service relationship is established, its value will not decay with the product life cycle.

Judgment: Hardware defines the product, software determines compounding

Over the next three to five years, a foreseeable trend is that the division of labor between hardware and software will become clearer. Hardware is responsible for "defining product form and reaching scenarios," while software and services are responsible for "determining the compounding space of the business." If a company still remains at the stage of "selling one device and earning money from one device," its ceiling is linear; but for a team that can convert the data and operations behind every device into subscription revenue, its growth is compounding.

The direct corollary of this judgment is that for the vast majority of traditional enterprises, there is no need to agonize over "whether to develop hardware in-house"; instead, they should focus on "whether they can software-ize existing device assets." Device connectivity, cloud control, data analysis, and scenario-based applications—each layer is a software opportunity that can be seized. The technical threshold is not in a single-point algorithm, but in whether these capabilities can be strung together into a continuously deliverable service chain. For a Shenzhen enterprise focused on software development, this is precisely where its core competitiveness lies—helping enterprises transform a "sell it and stop" hardware business into a service asset that "becomes more valuable the more it is used."

Xiangming Technology has repeatedly verified this point when serving local manufacturing and park customers in Shenzhen: the value of a management system lies not in how many functions were delivered on the day it went live, but in whether it can respond quickly to every subsequent business change. What truly keeps customers paying is never the code itself, but the service capability behind the code to "evolve with the business." Technology itself is not the goal; using technology to create business value that can accumulate and compound is the watershed for judging whether a software service provider is mature.

Shenzhen Xiangming Technology Co., Ltd. | Creating value with technology | xiangmingit.com