Uber is currently an unprofitable business. Lyft is too. Both have aspirations to become profitable businesses in the next one–two years, but right now, they aren’t even close. How can they achieve profitability, and what impact would that have on your fare to get from the airport to your Airbnb?

There are a few estimates on how much Uber loses per ride. It’s enough that, up until its initial public offering, your Uber ride was substantially subsidized by Uber’s investors. That means if Uber charged you $10 for your ride, Uber might pay the driver $6, earmark $6 for the costs of operating a business, and cover the $2 loss with some of the money they raised from venture capital (VC) funding.

For a while, everyone was okay with this. Uber and Lyft were happy because both sides of their marketplace (rider and driver) were growing, and one side (riders) was pretty happy. Users stayed happy because they kept getting cheap rides. VC firms were happy, too, because the companies’ ridiculous spending begat ridiculous growth. Drivers were not happy, but nobody seemed to care, save for a few cities and states.

As anybody who paid attention to the markets in the past two decades is well aware, this was a pretty common model, and it’s what led companies like WeWork to try to approach an IPO — growth first, profitability later. Amazon popularized this, as it was famously unprofitable for years — and the strategy worked because the company crushed its competition with lower prices and eventually founded high-margin, profitable businesses (like Amazon Web Services) to prop up its growth. WeWork was the most recent signifier to the markets that this model just wasn’t going to work any longer, unless you had profitability and monopolistic aces in the hole like Amazon. So far, Uber and Lyft don’t.

Many have accused Uber and Lyft of running businesses that only really work when their drivers are underpaid.

After a leadership change and a disappointing IPO last year, Uber revealed its finances to investment banks (and then the world) and showed that the cost of spending on growth was a lack of profitability. (A substantial chunk of Uber’s losses also come from its Uber Eats food delivery business, but I’m focusing in this piece on how its core ride-share operation gets profitable.) Since this literally meant that acquiring more customers wouldn’t pay the bills, something had to give.

Both Lyft and Uber have committed to becoming profitable soon, and to be honest, they don’t have much of a choice. If they don’t get out of the red, the markets will crush them.

An obvious place to start would be cutting back on costs, which they’ve begun to do: Layoffs happened at Uber — they’ve already let over a thousand employees go since their IPO. Growth spending is heavily marketing driven, so it’s not surprising that 400 of those layoffs were in the marketing department.

However, layoffs alone don’t make a core business profitable, and a smaller marketing department won’t fix the unit economics of these ride-share businesses. Here are three more substantial ways that Uber and Lyft could stop the cash burn and become profitable.

Method #1: Increase fares for riders

This scenario is the most obvious place for Uber and Lyft to start when considering how to become profitable. Uber has already signaled it will start to do this as it attempts to stem operating losses. You might have noticed it happening for your own rides.

The problem is, Uber spent billions of dollars blitz-scaling and growth-hacking its way to a large user base. Pivoting means their stock becomes valued at a profit rate instead of a growth rate. Shareholders will most likely shed stock and push down the value as a result. What’s worse, if they charge too much then they risk losing the customers it cost them substantially more to acquire.

Uber and Lyft’s endgame — unless they want to charge more to riders, drivers, or both — has to be maintaining a fleet of autonomous cars.

How much more of a cost are you willing to tolerate for the convenience of using Uber and Lyft? Many people are already guilty of exactly the habit that would drive their customers away long-term: checking both Lyft and Uber to find the lower price between the two. If prices get too high for too long on both apps, customers may start to notice that a certain taxi company is cheaper than Uber or Lyft or that a different ride-share startup is offering lower rates or that maybe they can tolerate walking, biking, or taking public transit. And, just like that, we’re back to pre-ride-share consumer trends.

Given how much the competition between Uber and Lyft has forced the taxi unions to modernize, this might be good news for taxis, who still cling to a lot of the closest airport real estate.

Method #2: Pay drivers less

This, quite frankly, is not a good option.

One out of every five Uber drivers don’t drive for Uber longer than a year, so it’s a side of the marketplace that already has a notably high churn rate. Part of this is by design — Uber pitches itself as an “in-between” gig for its drivers — but it still means a lot of turnover in the drivers that keep Uber alive.

Uber’s also been fielding legal battles involving drivers who want benefits. California just enacted a law that would force Uber and Lyft to classify their drivers as employees (and give them benefits like health care, which would dramatically increase Uber’s and Lyft’s costs). Both companies have no choice but to fight this, but it reflects the continued tension between the companies and their independent contractors. Many have accused Uber and Lyft of running businesses that only really work when their drivers are underpaid.

For this reason, lowering payouts to drivers isn’t a viable option. Even if it’s an incremental payout decrease, this would likely mean fewer drivers willing to consistently drive for Uber and Lyft over time. Fewer drivers means longer wait times for rides and probably higher fares.

Method #3: Switch the higher cost of paying drivers for the lower long-term cost of maintaining a fleet of autonomous vehicles

It’s no secret that Uber and Lyft are researching and developing autonomous vehicles. Many other companies, like Tesla, Waymo (Google), and Cruise (GM), are in the game too. None of these companies has large-scale operations established yet.

Uber and Lyft’s endgame—unless they want to charge more to riders, drivers, or both—has to be maintaining a fleet of autonomous cars.

It’s not just because self-driving cars almost certainly are the future of transportation. More practically, it’s because Uber’s and Lyft’s current operating models demonstrate they can’t profitably pay their drivers. One way to lower those costs is by eliminating drivers (its own contentious issue, and one the companies will inevitably have to address head on) and paying engineers to maintain a fleet of autonomous cars instead.

Sounds great for Uber and Lyft, right? The problem is timing. Those companies I mentioned above are pouring billions into self-driving. But an at-scale, rigorously tested deployment of a self-driving fleet is probably not possible for half a decade at least.

So how will these companies resolve this dilemma? The most logical, or likely, scenario is a temporary price surge over the next one to eight years, followed by eventual price decreases or flatlining once self-driving becomes viable at scale. So, in the short term, don’t be surprised if you’re paying more for that ride to the airport. The era of VC-subsidized ride shares is over.

This article was originally published by Jameson zaballos on medium.