AI in logistics: Truckload shipping

By | June 16, 2018

Truckload shipping industry overview

I’ll be focusing primarily on the U.S, but it’s important to note that truckload shipping is a global market and different regions will have their own unique opportunities. According to the American Trucking Association, trucks move approximately 71% of freight in the U.S. Revenues for the industry were $740bn in 2017 and represented 81% of freight revenues.

Moving to cross-border truckload shipping between the U.S and neighbors Canada/Mexico, we can see there’s still clear value generated by the industry. Trucks transported 60.1% of the value of trade between the U.S. and Canada in 2017, representing $327bn in total value. Similarly, in Mexico trucks transported 70.1% of the value of trade between the U.S and Mexico in 2017, representing $373bn in total value.

Top AI use cases in truckload shipping

While there’s numerous use cases for AI in truckload shipping, I see two major areas of importance:

  1. Autonomous trucks (well, semi-autonomous for the near future)
  2. Improving the booking process between shippers and truckers

Autonomous trucks demonstrate the obvious potential to replace most human labor costs associated with driving trucks. Almost all companies will improve margins in the long-term assuming the trucks are successfully developed to the point of minimal or no accidents, delivery speed does not decrease (will likely increase at level 4/5 autonomy) and costs to develop these autonomous trucks go down in the future as more companies are able to successfully create them, thus creating intense competition.

It’s important to note that the period between level 5 autonomous trucks and now will feature volatility that makes it difficult to determine who the real winners are. During this semi-autonomous phase, we can expect some pitfalls such as crashes that set back the industry. We can also expect backlash from the trucking community as their jobs become endangered. How the industry navigates this period will be extremely important. In an ideal situation, autonomous trucks and drivers work together as partners with drivers only taking over in the most complex situations such as abnormal city traffic.

Improving the booking process between shippers and truckers is the other major use case in this industry. I’ll highlight a few startups already tackling this later, but at a high level, there’s major inefficiencies with this entire process. Most booking historically takes places via email, phone calls and the use of loadboards, a process that can take many hours. For example, Loadsmart (profiled later) helped Daimler Trucks (one of the largest trucking companies in the U.S) improve their booking process from 5 hours to just 18 minutes. In addition, the industry is extremely fragmented with 90% of trucking companies operating 6 or less trucks, and 97% operating 20 or less trucks.

In this type of fragmented environment, I think there’s room for multiple winners. Growth will come from partnering with more trucking companies to gain access to their data and customer bases. The value of this data is immense as it can be used to feed algorithms analyzing traffic patterns, the impact of different routes on different types of vehicles, pricing, frequency of maintenance issues (for predictive maintenance) and more. As such, it’s difficult to imagine one startup that can successfully obtain a monopoly over all the different regions in North America. Not only is this difficult due to geographic differences in truckload shipping, but also because of the different types of goods shipped and customer bases. Specialization is needed in the beginning stages of this industry to offer the best platform for a specific niche.

Other AI use cases include predictive maintenance, fleet management, and route mapping to name a few. The reason I don’t think these use cases have as much potential is because they can all be easily implemented on an autonomous trucking or booking platform as side features. You can already see this through some of the AI startups in the space.

Company profiles: Autonomous trucks

WaymoLarge cap technology company

  • Latest news: Back in March 2018, Waymo launched a pilot program in Atlanta for its self-driving truck fleet that will carry freight to Google’s Atlanta data center
  • Platform: In terms of data to leverage, Waymo will use the same suite of custom-built sensors that power their self-driving minivan, but specifically tailor them for Atlanta. Their engineers and AI experts are leveraging the same five million miles already self-driven on public roads, plus the five billion miles driven in simulation. Human drivers will be stationed to take over in emergencies
  • Personal perspective: Uber abandoning their autonomous truck program is a huge win for Waymo, who is now the large corporate leader in my opinion. The combination of the amount of data generated from the existing autonomous vehicle program, and new data generated from Atlanta is a huge advantage over competitors. Atlanta is one of the major logistics hubs in America and is an ideal place for testing. Waymo is also a major candidate for acquiring smaller autonomous truck startups

EmbarkEarly stage startup

  • Latest news: Currently running a daily service between L.A. and Phoenix. Back in February 2018, Embark’s self-driving trucks drove from L.A. to Jacksonville carrying Electrolux refrigerators. The 2,400 mile journey builds on their previous test route of L.A to El Paso
  • Platform: Currently has five trucks in operation and plans for 40 by early 2019. The software handles all highway driving, but a driver is always present, ready to take over in an emergency. The current system represents level 2 autonomy, but Embark has goals for level 4 autonomy in a few years
  • Investors: $30m series B in July 2018. Has raised $47m to date. Investors include Sequoia (series B lead), Y Combinator, SV Angel, AME Cloud Ventures, and Data Collective (DCVC)
  • Management: Alex Rodrigues (CEO and Co-Founder), Brand Moak (CTO and Co-Founder)
  • Personal perspective: Successful February 2018 pilot route from L.A. to Jacksonville was a strong momentum builder. Was key for securing their series B round in July 2018. If they’re able to scale their fleet to 40 trucks in the near future and complete more cross-country trips, Embark will stand out as a leading startup in autonomous trucking

Tesla – Large cap technology OEM

  • Latest news: Back in March, another Tesla autonomous electric semi-truck was spotted in California. Elon Musk confirmed the truck was making its first trip transporting by battery packs from the Gigafactory in Nevada to Tesla’s Fremont factory in California
  • Platform: Like Tesla’s Model X and S, the electric semi-trucks come with semi-autonomous features. The “enhanced autopilot” feature helps the vehicle stay in lane, warns about potential collisions, and can even automatically apply the brake in the case of an emergency
  • Personal perspective: Although the platform is lagging other competitors in terms of miles driven, the Tesla brand and existing partnerships could help Tesla catch up as the platform improves. Seems like the immediate testing grounds will be transporting Tesla battery packs and other supplies between different Tesla factories. These aren’t the longest and most complex routes, so it will be interesting to see when Tesla starts expanding its testing grounds

Daimler Trucks – Large cap traditional OEM

  • Latest news: Daimler revealed its autonomous truck development strategy during Daimler Trucks Capital Market and Technology Day, held in Portland for Wall Street analysts and investors. In addition to making its Freightliner Cascadia semi-tractor a global test bed for developing self-driving semi-tractors, it also will open an automated truck research center at its U.S. headquarters in Portland, Oregon
  • Platform: Freightliner Cascadia semi-tractor is the main test truck (level 2 autonomous status). Immediate focus will be on highway driving. Engineers will collaborate with counterparts at Daimler Trucks locations in Stuttgart, Germany and Bangalore, India, who are working on automated driving and tapping previous research performed across other Daimler AG divisions such as the Mercedes-Benz car brand. The company also will expand automated truck research and development in Germany. In addition to self-driving features, Daimler is focused on developing “platooning systems” which allows groups of digitally tethered trucks to closely follow each other to reduce drag and increase fuel efficiency
  • Personal perspective: With $600mm of near-term R&D dedicated to autonomous trucks, I think Daimler is in a good position to be a leader among the traditional OEMs. Their timeline for level 4 autonomous trucks is longer than Embark though, with Daimler executives stating they will be ready to sell its trucks in ~5 years

TuSimple – Early stage startup

  • Latest news: In September 2018, announced plans for further expansion in Tucson, Arizona. The company has been testing its level 4 – class 8 autonomous trucks in the state of Arizona for over a year and recently began generating revenue hauling freight for commercial carriers in the state
  • Platform: Camera-centric system has a vision range of 1,000 meters—farther than any other perception system today—and can “see” 360 degrees around for a pixel-level interpretation of the visible environment, enabling the vehicle to locate itself within four inches of the road at all times. To achieve the 1,000 meter perception range, TuSimple spent two years developing the deep learning algorithms that are the instructions used to help the perception system understand terabytes (TB) of data created per vehicle every day. TuSimple turned to Amazon Web Services (AWS) to support the company’s development process, relying on AWS’s machine learning expertise and nearly unlimited compute and storage capabilities
  • Investors: $55mm series C in November 2017. Raised $83m to date. Investors include Composite Capital Management (led series C), Sina (led series B), Nvidia, ZP capital
  • Management: Mo Chen (CEO and Co-Founder), Xiaodi Hou, (CTO and Co-Founder, phD from Cal Tech in computation and neural systems), Jianan Hao (COO)
  • Personal perspective: The 1,000 meters vision range is impressive, and the successful June 2018 SAE level 4 test run between San Diego and Yuma, Arizona solidifies the prowess of their technology. Potential roadblocks for the startup include the rush to level 4 autonomous truck testing. A fatal accident for the startup is more likely than it is for Embark (who is focused on level 2 testing first). Given the history in Arizona with Uber’s accident, it’s unlikely they’ll survive post investigation

Starsky Robotics – Early stage startup

  • Latest news: Will be making driverless deliveries in Florida by early 2019, with at least one truck in operation
  • Platform: Utilizes automotive-grade radars for increased reliability in measurements. Radar is good at sensing the existence of a potential obstacle and measuring its velocity. It’s not so good at identifying its precise location, and it also tends to create a lot of false positives. To filter out the false positives, Starsky Robotics fuses the data stream from the radar with the stream from the cameras instead of utilizing LiDAR systems. Each sensor covers the weaknesses of the other
  • Investors: $16.5m series A in March 2018. Raised $20.3m to date. Investors include Shasta Ventures (led series A), Trucks Venture Capital, Abstract Ventures
  • Management: Stefan Seltz-Axmacher (CEO and Co-Founder), Kartik Tiwari (Co- Founder)
  • Personal perspective: Lack of successful medium or long-range test pilots puts Starsky Robotics behind Embark and TuSimple. The platform is still in the very early stages of development.

Kodiak Robotics – Early stage startup

  • Latest news: Raised $40m series A in August 2018
  • Platform: Not many details available publicly. Engineers will focus on developing the full self-driving system stack from the company’s own hardware and software architectures. However, Kodiak Robotics is not going to build any sensors. Instead it will use sensors from third-party suppliers. They also plan to use light detection and LiDAR as well as camera, radar and sonar technologies
  • Investors: Battery Ventures (series A lead), Tusk Ventures, Lightspeed Venture Partners, CRV
  • Management: Don Burnette (CEO, former Co-Founder of Otto, former Google Staff Engineer), Paz Eshel (COO, former VP at Battery Ventures)
  • Personal Perspective: Recently emerged out of stealth mode. Technology platform is unproven, but the management team is strong. Don Burnette’s experience as Co-Founder of Otto (acquired by Uber) should be a strong advantage as the startup builds out its platform

Volkswagen / Toyota partnership – Large cap traditional OEM

  • Latest news: Launched the partnership back in April. Both companies expressed a sense of urgency with developing autonomous trucks and the need for “allies”
  • Platform: Specific platform is yet to be unveiled but should cover “procurement and logistics in addition to hybrid engines, connectivity and other technologies,” the Financial Times notes
  • Personal perspective: Volkswagen and Toyota are behind other competitors who are already testing autonomous trucks. However, both companies have enormous R&D budgets and existing autonomous vehicle programs that they should be able to leverage. Strong candidate to acquire an autonomous trucking startup to play catch up should the partnership get off to a slow start

Company profiles: Booking / shipping process improvement

Loadsmart – Early stage startup

  • Latest news: Partnered officially with Daimler back in January. Notable results from the partnership include decreasing Daimler’s time to execute a spot shipment from 5 hours to as little as 18 minutes. Around the same time they also hired a new CTO, Joao Bosco, who will focus primarily on furthering the company’s AI platform
  • Platform: The company’s first feature was instant pricing and booking for full truckloads, which when launched in 2015 was the first of its kind in the U.S., allowing shippers to book a full truckload in under 5 seconds. Interested shippers could open the Loadsmart web portal and could book a truck even without signing up, which was an instant success. The company then moved to integrating its solution into enterprise accounts using its API
  • Investors: $21.6m Series A in October 2018. Raised $34.7m to date. Investors include Connor Capital SB, Maersk Growth, Chromo Invest
  • Management: Ricardo Salgado (CEO and Founder, former MD in principal investing at Goldman Sachs), Felipe Capella (Chief Product Officer and Co-Founder)
  • Personal Perspective: Strong partnerships (Daimler in particular). Early mover advantage having entered the market in 2015. Loadsmart differs from other competitors in this niche as it has invested heavily in its engineering division instead of the sales division – thereby allowing it to assemble a team of machine learning engineers and data scientists. The company gathers data from every single one of the 272,000 carriers running on the U.S. roads and uses the data in its sourcing algorithm, to find the best possible match for its shipping clients. I strongly believe they’re a market leader in the U.S., and I can see them expanding with much success into Mexico and Canada

Uber Freight – Late stage startup

  • Latest news: Uber CEO Dara Khosrowshahi recently described plans to “Uberize the freight brokerage business
  • Platform: Uber launched its freight app a little over a year ago in a small town in Texas. After initial success, the app expanded to California, Illinois, and Michigan six months later. Within the past six months the app has expanded nationally and has addressed numerous “defects” found in the earlier versions of the app. Fleet mode was launched recently, which now allows multiple drivers from the same company to be part of the same booking
  • Personal perspective: Main edge will be the ability to sustain lower margins versus competitors, thus gaining favor with shippers / truckers. While I think costs are one important consideration shippers and truckers make, the quality of the platform is more important. Slow delivery speed for shippers, and inefficient routes for truckers will drive users away from the platform regardless of cost. Uber Freight was originally designed for one driver per booking and has only recently launched its app for multiple bookings per driver. In all, I think Uber Freight is lagging behind Loadsmart and will need to play catch up

Project44 – Late stage startup

  • Latest news: Acquired GateHouse Logistics, a Denmark based company that offers visibility solutions, on December 4, 2018
  • Platform: Originally built and optimized for less-than-truckload (LTL) shipping. Platform expanded to truckload shipping solutions post project44’s $10.5mm series A in 2016. Project44’s software automates the flow of data among shippers, carriers, and 3PLs by allowing customers to get data through APIs that directly link computer databases instead of relying on third-party interfaces, such as email or electronic data interchange (EDI). Their truckload visibility product is one of the few, if not only, that is API driven (most run off legacy systems or are manually batched)
  • Investors: Raised $45m series C in October 2018 at ~$500m valuation. Investors include Sapphire Ventures, Chicago Ventures, Pritzker Group Venture Capital, OpenView, Emergence Capital, 8VC
  • Management: Jett McCandless (President & CEO), Tommy Barnes (President)
  • Personal perspective: I think project44 has a unique market position with CEO Jett McCandless saying they’re only logistics technology provider with direct APIs for large truckload carriers, noting that the major ELD technology providers, including PeopleNet, Omnitracs and others, have said project44 is the first company to ever have an ELD with them that is authorized and certified. The acquisition of GateHouse logistics certainly expands their presence to Europe, a move that Loadsmart has not done yet

Flexport – Late stage startup

  • Latest news: Opened a Chicago office which marks their 5th office in the U.S.
  • Platform: Flexport holistically analyzes all its data to optimize shipping routes and simplify relationships with ports, truck drivers and anyone else that touches a container. They differ from Loadsmart and Uber Freight due to their focus on trucking routes from ports to their ultimate destinations, as opposed to all aspects and routes within truckload shipping
  • Investors: $100m corporate round in April 2018 at >$1bn valuation. Raised $304m to date. Investors include S.F. Express (corporate round lead), DST Global (series C lead), Founders Fund, Y Combinator, First Round Capital, Bloomberg Beta, Alrai Capital, MacroVentures
  • Management: Ryan Peterson (CEO, previously Co-Founder of ImportGenius)
  • Personal perspective: They’re a much smaller player when it comes to truckload shipping. Main business focus is on maritime shipping. Flexport is slowly building a niche in logistics between ports and first delivery points, but lacks the broader truckload shipping scale and experience of competitors

Manbang Group – Chinese late stage startup

  • Latest news: Mangbang has faced quite a bit of negative publicity recently. Dozens of cities in China have seen protests by truckers, who claim that Manbang has an unfair monopoly on the truck hailing market and has driven rates lower, thus negatively impacting margins for truckers (like the impact Uber had on taxi drivers)
  • Platform: In April 2018, Mangbang said 5.2mm of the 7.0mm freight trucks in China were members of their Full Track Alliance Group.  The firm is often described as China’s “Uber for trucks”. It runs an app that allows companies to connect with truck drivers – often independent contractors – tapping into demand for haulage in one of the world’s busiest markets for goods transport. The company plans to use the new funding from Google and Softbank to beef up their AI capabilities, develop autonomous trucks, improve user experience, and expand into new geographic areas
  • Investors: $1.9bn private equity round. Investors include China Reform Fund Management (lead), Softbank (lead), CapitalG, Tencent, Sequoia
  • Management: Wang Gang (CEO, also an angel investor in Didi)
  • Personal perspective: I expect Manbang to be the major player in Asia, and eventually Africa due to the region’s increasing reliance on Chinese technology. They have developed strong relationships with the Chinese government and secured investments from leading tech companies in the region. Manbang’s scale is unrivaled in China