Last week, we saw a torrent of news stories covering the launch of the DJI Phantom 4 and several WSJ articles about how drone investments have not delivered to full potential. As an early stage investor looking at this sector, I’m trying to make sense of these headlines

The Phantom 4 is a quadcopter drone made by China’s DJI Technologies. Currently, DJI claims 70% of the $2 billion drone market, having shipped an impressive 700,000 drones last year alone.

Retailing for $1,399   — the price of a GoPro plus an iPad  —  the Phantom 4 boasts a slew of features that make the quadcopter easy to fly and hard to crash. The key is autonomous flight with increased computer vision. The drone is incredibly easy to use. Just double tap and the DJI Phantom 4 takes off; tap once more and it lands.

The end result? A drone that shoots fantastic 360-degree video and takes brilliant photographs. Less helicopter, more camera. All the hobbyists out there who are still buying cameras will no doubt be converting. And since consumers won’t have to worry so much about crashing this drone, the price point will continue to drop.

Expanding consumer demand and the China factor

The FAA has required U.S. citizens to register their drones since Dec. 21 of last year. In the ensuing two-plus months, more than 300,000 drones have been registered. Believe it or not, there are now more registered drones than piloted aircraft in the United States.

Even though the market is still in its infancy, 40,000 drones were shipped to Chinese customers in the third quarter of 2015. While some U.S. consumers might have disposable income  —  some might even consider $1,399 for the Phantom 4 an “impulse buy”  —  the Chinese have shown they will spend on drones. The WSJ reported that DJI is opening its first flagship 8,600 square foot store in the Shenzhen shopping center, which will sell products ranging from $500-$4,000. DJI is marketing and getting more consumers to touch and feel to convert to sales. Research suggests 3.1 million drones will be sold in China in 2019  —  quite the step up from the current sales figures. These forecasts are precisely why DJI recently raised an additional $75 million, boosting its valuation to $8 billion.

So where are the early-stage investment opportunities?

Goldman Sachs estimates commercial drone sales market to be $20.6 billion and consumer sales to be $14 billion in the next five years. Even with all of this demand, early stage investors are not likely to invest in hardware or drones themselves but in the services that drones can provide. Why? Most drone hardware is available from China at minimal cost, and the basic software is open source in many cases, meaning that drone makers will face plenty of competition. On the other hand, the companies that adapt them to spray crops, inspect buildings, pipes, rigs, do mapping, and other useful tasks will make money. Like the computer industry, making hardware is easily commoditized. Operating systems are becoming low margin. But using software to solve a problem is high margin.

Near-term commercial demand is being driven by a variety of use cases in media/entertainment, journalism, agriculture, real estate, and public safety ,  just to name a few. In many cases, it is not economically efficient for buyers to own and operate drones themselves, so companies providing “drone as a service” businesses and tangential services have cropped up and could be interesting investments.

Venture dollars have already poured into the industry. Airware, which received venture funding itself and focuses on operating systems for drones, has now formed its own venture fund.

In precision agriculture, many companies have been pursuing drone-as-a-service models. For instance, Greensight is using custom drones and sensors to monitor turf grass for early signs of chemical needs, thereby lowering the cost for chemical companies that service large turf areas such as golf courses.

For inspections, companies like Percepto are using computer vision for drones to reduce the cost of wind turbine inspections. Others are targeting oil and gas rigs and mining. The construction industry is being disrupted by drones offering superior and cheaper surveying such as Skycatch. In all of these instances, surveying is time-consuming and expensive and can even be dangerous because of difficult terrain. A map is only as good as the number and accuracy of measurements taken. Drones are solving a large problem here.

Investing in drone protection

Drone maintenance is another area of interest. Like any other piece of machinery, drones will need to be maintained, and aircraft maintenance is already a large industry. For example, companies like Robotic Skies have as many as 60 drone repair centers sprinkled across the globe. In a recent interview, the company’s CEO said Robotic Skies isn’t content with simply offering repair services; it intends to add a slew of value-added services into the mix in the near future  —  like training, data capture and analysis, and custom component development, among other things.

Drone insurance is another emerging industry. Drones can crash. From startup insurance brokers to major players like AIG, this niche industry will only grow larger as more and more drones take to the skies. Today’s policies mirror standard aviation policies, but as usage grows, issues of operator safety, privacy, cyber security, property, and access rights are likely to arise. On the commercial side, there is very little data on which insurers can accurately price risk, but this will change with more usage. On the consumer side, most homeowner policies cover the use of radio-controlled vehicles, and insurers can add a rider to extend to drones.

Then, there is drone security. We are seeing infrastructure plays on the defense side that provide drone detection and response. Two companies specifically,(Dedrone, a European company, and DroneShield) are focusing on drone detection. Another company, Skysafe, is licensing its technology to prevent malicious drones from entering stadiums and prisons and to stop drones from smuggling drugs across the border.

In case you get injured by a drone, there is always Flirtey, which came out of Y Combinator. It was one of the first companies to do a test of a delivery within the U.S. It delivered medical devices or healthcare-related products about 3 miles.

There is also a shortage of pilots, so perhaps we will see companies emerge to address this need. And many other use cases will create other markets.

Regulatory risk

Thus far, the Federal Aviation Administration (FAA) has moved more slowly than anticipated to create a regulatory framework for drone operations. The FAA missed its September 30, 2015 Congressional deadline to draw up rules for widespread use.

In order to fly commercially, prospective users must obtain a Section 333 exemption and Certificate of Authorization from the FAA as an exception to the blanket ban on commercial drone use. Still, operators are largely prohibited from flying in ways that deliver the greatest returns (i.e., more autonomously, beyond line-of-sight, over people, and at higher altitudes).

Around 4,000 exemptions have been issued so far. According to CB Insights, as of July 2015, 44% of all the Section 333 exemptions have gone to film, photo, or video use cases; that includes companies using drones for movies and TV shows. Two other groups operating under the exception are artists and real estate agents trying to map properties. The second highest use case was in inspection and monitoring. Third was mapping and surveying in order to understand construction zones or city planning or general land and commercial construction. Fourth was precision agriculture.

Without a clear roadmap on regulation, drone manufacturers have been slow to develop key technologies that might not be compliant in subsequent years, even if these technologies might facilitate uptake, expand capabilities, and enhance safety today.

Of course, the most hyped use of commercial drones is delivery, where Amazon receives the most press in the U.S. In China, S.F. Express is already using drones for logistics delivery ,  but there are still many obstacles before we see this in mass. Already the FAA and NASA are actively recruiting work groups from the industry to solve the air traffic control issue.

So how do investors make sense of this? Clearly, we are seeing more activity in this sector targeting real enterprise issues that are mission critical, but we need to keep close relationships with the regulators. WSJ may be right that we have not seen the masses benefit yet, but clearly there are areas of early stage opportunity here to invest. The sky’s the limit !

[A version of this story originally appeared on the author’s personal blog.]

Nisa Amoils is an angel investor, advisor and former entrepreneur. She is a member of New York Angels. She invests in a variety of technology and is the co-chair of the frontier technology committee which encompasses AI/VR, robotics, drones, driverless cars, autonomous innovations etc. She blogs on Medium @AmoilsNisa.