Within the next decade, tens of millions of homeowners could find it economically advantageous to invest in solar combined with energy storage, further distancing themselves from their utility.

But will they actually sever ties completely?

Barring extraordinary circumstances, the economic case for grid defection is still very weak for U.S. consumers. The electricity system offers valuable backup in case a customer over- or under-invests in an on-site system. It also offers a transactive marketplace that can help a customer or solar installer earn a return faster, assuming there are tariffs in place that value the electricity produced by solar.

Even SolarCity, a company at the forefront of promoting hybrid solar and battery storage systems, believes grid defection is a bad idea. In a recent blog post on the issue, SolarCity's CTO Peter Rive called the idea "polarizing" and described why the company doesn't believe it is an optimal outcome.

"While this is technically feasible, SolarCity has no interest in this scenario. While cutting the cord enables one household to be 100% renewable and self-sufficient, it limits what these technologies can do. In short, the grid is a network, and where there are networks, there are network effects," wrote Rive.

For now, defection is more of a theoretical scenario in the U.S. rather than a definite outcome. But could anything change this dynamic and make it more likely that customers abandon the network?

According to analysts at Morgan Stanley, there is one major factor that could increase the likelihood of the grid defection over the next decade: fixed charges on net metering.

In a new report on the growth of the solar industry, Morgan Stanley researchers warn that high fixed charges could encourage more customers to pair lithium-ion batteries with PV systems and abandon the central electricity system altogether.

"There may be a 'tipping point' that causes customers to seek an off-grid approach -- higher fixed charges to distributed generation customers are likely to drive more battery purchases and exits from the grid," wrote the team of researchers.

Utilities around the country, including those with virtually no solar deployment in their service territories, are working to change net metering and add extra monthly fees to the bills of solar customers in order to pay for the fixed costs of managing the grid. There is much debate over how to structure these fees -- and also concern from the solar industry about whether they're needed at all.

The common wisdom is that adding monthly fees will kill the economics of solar. And while that may be true in some markets with low electricity rates, the Morgan Stanley analysts predict it could have the opposite impact in some states with above-average rates like California and Hawaii.

"The higher the fixed charge required of distributed generation (primarily solar) customers, the greater the potential that customers purchase batteries on a large scale and go completely off the grid," concluded the authors.

The analysts modeled a scenario in California where utility bills increase by 5 percent a year, solar reaches a 20 percent penetration level, and monthly fees on solar customers reach 50 percent of the average customer's bill. In that case, a residential bill could hit 26 cents per kilowatt-hour by the end of the decade -- a higher rate than if a solar-storage system supplied all of a consumer's electricity needs.

"Over time, many U.S. customers could partially or completely eliminate their usage of the power grid. We see the greatest potential for such disruption in the West, Southwest, and mid-Atlantic," wrote the analysts.

That finding is consistent with a recent analysis from the Rocky Mountain Institute and CohnReznick, which found that more than 20 million residential customers in those regions could "find economic advantage" in solar-storage systems by 2024 under a moderate technology improvement scenario.

The Morgan Stanley analysts are bullish on improvements to lithium-ion batteries, which could fall to $200 to $150 per kilowatt-hour over the coming years. With Tesla preparing to build a Giga factory for mass producing batteries, this scenario has become more realistic.

"For every $25/kWh reduction in the cost of lithium-ion batteries, we estimate the all-in cost of power to customers falls by about $.01, or about 15% of the residential customer price for grid charges," wrote the analysts. By 2028, Morgan Stanley estimates that Tesla’s 3.9 million electric cars and stationary battery systems will enable 8 percent of U.S. homes to store one hour of electricity.

However, predicting defection rates is a difficult exercise. Even as the solar-storage combination becomes attractive for millions of electricity customers, many won't want to detach themselves completely from the utility. There are many deciding factors that could prevent full defection, including convenience, capital cost and trust of the technology. 

As researchers at the Rocky Mountain Institute recently pointed out, simple economics likely won't be the deciding factor: "Customers will make decisions that serve their best interest based on the many factors...and while for a very few, defection might ultimately be the best outcome, mass defection from the grid is almost certainly suboptimal."

But the potential is increasing. And as utilities seek to recover their costs by charging solar customers, it could accelerate the process -- that is, unless they get in the market themselves.

"Utilities in some regions could adapt to distributed generation to minimize the impact on shareholders," concluded Morgan Stanley.