We grounded our investment in drone-maker AeroVironment

One of the pleasures of managing a technology fund is that seemingly every day, I encounter something new and interesting emerging from corporate or academic research labs. Some of these “wow” moments come while I’m reading traditional technology research, or reviewing the many blogs that follow various corners of the technology world.

It’s hard to say that one thing is more interesting than the next, but this week I saw a brief video that was stunning not only for, 1) what was shown; but 2) for the real-world potential embedded in the research being done: Here’s the video in question; it’s only about two minutes long:

(For those of you who didn’t/couldn’t see the video, it is a simple demonstration of one drone helicopter tossing a six-foot-long weighted pole to another drone helicopter, which catches the pole on its other end. This all takes place inside a warehouse-sized building, inside a room that is well-lit and padded.)

Now let’s break down what we’re seeing in a little more detail. The video showed two robotic flying machines, operating independently, but with awareness of each other’s proximity. One robot balanced a pole almost casually, re-calculating the forces of the robot/pole combination fifty times per second, and adjusting its power levels on its four rotors to keep the pole not only balanced, but also essentially motionless in mid-air.

Then one robot signals the other robot that it plans to toss the pole, and the two robots coordinate their actions so that the pole is launched from one robot to the other, landing nearly upright. The audio voice-over indicates that the software embedded on each robot has algorithms that allow them to “learn” and improve their performances, presumably dropping the pole less often (or not at all) and catching it with fewer course corrections.

The first couple of times I watched this, I kind of jokingly thought to myself: well, so much for restaurant waiters. I mean, why wouldn’t a restaurant simply use a drone like this to deliver meals? A drone that would never drop a platter, never be burned by the fajita dish, and never stepped out back for a smoke. Place your order with your tablet, and get your meal delivered flawlessly a few minutes later. “Hi, I’m Droney. Here’s your food. Goodbye.”

But let’s break down what we’re seeing into the separate technological elements:

  1. Drone aircraft flying autonomously;
  2. Machines learning;
  3. Machines communicating with each other wirelessly;
  4. Dynamically unstable aircraft (balancing the pole), adjusting their controls at rates far faster than could humans in order to remain stable; and
  5. A multi-camera video system that performs real-time motion capture, creating an indoor “GPS.”

In short: we’re seeing five (and perhaps more) leading-edge technologies integrated to perform a task that humans couldn’t do. And in many cases (bomb disposal? hostage-situation-in-progress surveillance? in-caldera volcano monitoring?) humans wouldn’t want to do.

The academics in Switzerland who organized this project have already commercialized some of this technology. And it’s fun to contemplate the positive ways these technologies could be harnessed for good (e.g., squadrons of infra-red sensor-equipped drones searching for avalanche victims, etc.).

So what does this amazing demonstration of applied technologies have to do with how we go about our business here at Crabtree?


Impressive as it is, this flying demonstration doesn’t alter what we look for in a Crabtree Technology model company, or how we go about assembling our portfolio and re-balancing it every three months. Our recent investment in AeroVironment (AVAV) highlights this. AeroVironment manufactures Unmanned Aerial Vehicles (a.k.a. drones), mostly for sale to U.S. government agencies, including the military.

We bought AVAV in early 2012 because it was found by our quantitative model. It seemed likely to possess the three Crabtree Attributes for which we look: 1) holding on to and possibly gaining share; 2) generating cash; and 3) executing on its own and Wall Street’s expectations.

But by November 2012, we sold AVAV because the company seemed unable to execute on its expansion plans – a troubling development given what we know about the huge growth taking place in the UAV (Unmanned Aerial Vehicles) space. We believe AVAV was in a prime position to capture even more market share. But we can’t abide by companies unable to take advantage of such a huge market tailwind. So it was gone.

And that neatly summarizes how we view technology: we’re capable of being impressed, but we’re more impressed by a good business model being shepherded by capable managers’ intent on profitably taking market share.

Fourth quarter “earnings season” got off to a start in January. A number of companies making January announcements met or exceeded expectations. Some were optimistic in their outlook for 2013; others were more conservative. And still others, like Spansion and Liquidity Services, are clearly under-performing. We don’t expect under-performers to survive our February re-balancing.

January performance for the Crabtree Technology model was mixed: strong on an absolute basis, but slightly behind these relevant benchmarks. The model gained 4.7% in the month, net of fees, compared with a 6.2% gain for the Russell 2000 benchmark and a 5.0% gain for the S&P 500 (SPX). Our internal benchmark, the Merrill Lynch Technology 100 (MLO) rose 6.7% during December.

We’ll provide the full details on our February performance in our next manager letter, along with the balance of earnings reports and the results of our quarterly re-balancing of the Crabtree Technology model. But if you have any questions or comments, you can reach us any time at info@crabtree-am.com.

The investments discussed are held in client accounts as of February 28, 2013. These investments may or may not be currently held in client accounts. The reader should not assume that any investments identified were or will be profitable or that any investment recommendations or investment decisions we make in the future will be profitable.