Tesla Semi comparison
Tesla Semi comparison
This juggernaut of a startup looks set to drive off with a truckload of the hefty $4 trillion dollars global freight market.
And with its fleet of futuristic self-driving lorries already in service across the US and China, Tusimple can indeed claim its miles ahead of its rivals.
So who are these upstart truckers, what makes their technology worth its billion-dollar IPO, and can they stave off competition from the most beloved name in transportation?
Join us today as we hit the highway and ask how Tusimple plans to compete with the Tesla Semi.
Although the trucking business in the United States alone is valued at $800 billion a year, the sector is facing a crisis.
Why?
Right now, according to industry sources, the sector is facing a 60,000-strong driver shortfall.
And as veteran drivers retire without being replaced by the next generation, this deficit is expected to grow to 100,000 over the next few years.
Long hours and extended periods away from home, with comparatively low wages, are apparently putting young people off joining the profession.
But produce still needs moving around, more than ever in fact.
So can technology hold the key?
When most people talk about self-driving technology they mean private cars that whisk you to work while you nap, or driverless ride-hailing services like Google’s Waymo.
But it turns out long-range haulage is the perfect application for contemporary autonomous tech. Why? Big-rig trucks of the kind seen on major US arteries typically do some 95% of their mileage on well-maintained, predictable highways.
Unlike city driving, which is full of surprises and liable to random shocks like children running out into the road, being able to regulate speed and occasionally change lanes on an Interstate is well within the reach of currently available technology.
And in fact, it turns out, the mammoth scale of modern semi-trucks makes them a great candidate for automation, as sensors can be spaced apart more widely than on a regular car.
Add to that the fact self-driving truckers will never be tired, moody, drunk, or be forced to stop for a government-mandated nap, and you start to see how Tusimple, a Sino-American firm founded in 2015, claims it can lower the cost of freight transportation in the US by about 30%.
How?
By running trucks all day and night, only pausing for refueling and essential preventative maintenance.
When you consider autonomous drivers are, in theory anyway, safer, insurance premiums should also be cheaper.
Not to mention one University of California San Diego study suggesting simple mechanical consistency in applying the gas pedal could reduce the sector’s carbon footprint by 10% at a stroke.
TuSimple is backed by automotive giant Volkswagen, as well as logistics titans Penske, the US Postal Service, Werner Enterprises, and Union Pacific.
Amazon and UPS have pre-ordered TuSimple’s autonomous trucks, which are manufactured in association with old-hand truckmaker Navistar.
So who actually is behind Tusimple?
Co-founded by internationally renowned AI, computer learning, and computer vision expert Xiaodi Hou, who presently serves as CTO, the company started life as a software outfit specializing in code for recognizing the make and model of cars by analyzing images.
Hou’s cofounder Mo Chen realized this kind of visual-recognition know-how could be monetized by moving into the autonomous driving space.
Freight seemed like the way to go, as cargo is far less demanding than human passengers, and, as we’ve said, highway driving is a lot easier for algorithms to manage.
Tusimple’s self-driving tech hinges on its superior perception technology. Up to 10 high-definition cameras are able to identify objects up to one kilometer away.
At highway speeds, claims Tusimple, this provides the truck’s onboard AI a so-called ‘planning horizon’ – for identifying hazards and taking action – of about 30 seconds, as good or better than most human drivers.
Add to this a sophisticated suite of shorter-range microwave radars and supplementary liar technology, and the autonomous truck is also safe in fog and heavy rain.
Safety aside, Tusimple’s one-kilometer outlook is key to cutting fuel costs and carbon output.
Because if a hazard is spotted in good time, there’s no need for rough braking and the necessary unhelpful acceleration after that.
At a fleet level, those gas savings really add up – to the point potential clients can get their money back on a Tusimple truck in just two years, according to the company website at least.
CTO Xiaodi Hou spent six years working on neural processing at Caltech, including long stints exploring deep learning and visual processing.
As such, the patented tech he’s developed for TuSimple, running on partner Nvidia chips, is supposedly capable of reacting to nasty surprises 15 times faster than a human driver, while processing some 600 trillion operations per second.
This is no small thing – according to US data, 95% of all fatalities involving large trucks are attributable to human error.
To be clear, TuSimple trucks on the road right now – 50 in the US, and 20 in China – aren’t fully level-5 autonomous.
Rather, they’re Level 4, capable of performing driving functions along the so-called ‘middle mile’.
So long as a human can get them from the depot to the highway at the start of the trip, and vice versa at the end of the trip, the system works.
To date, all such journeys have been conducted with a human safety driver – or two – aboard to take over whenever the AI struggles with a new situation.
Which does happen.
Right now TuSimple US locations are in Arizona, New Mexico, and Texas, delivering freight along clearly marked modern highways, in large clear, and sunny conditions.
It’s not expected to work, for now at least, on rural backroads. So how will the company cope with the threat from Elon Musk and his shiny Tesla semi? Investment is part of the story.
Both Tusimple founders became overnight billionaires when the firm launched its Initial Public Offering back in April.
‘Having an IPO is a real boost,’ said Hou at the time.
‘We get a lot of cash so we can do more hiring and build out the team more quickly so we can solve bigger problems and deliver sooner.’
Still, the 800-or-so strong team generated only $1.8 million in revenue last year, from its haulage runs on behalf of sympathetic partners.
And despite an apparently bulging order book of nearly 6,000 trucks slated for the likes of Amazon, customers were only required to put down a comparatively modest $500 apiece, and most likely have investments in other similar truck projects from the likes of Waymo and Embark.
While Tesla’s business model is about providing all-electric trucks with human drivers very much part of the package, Tusimple’s vision is to launch human-free autonomy for its customers, across the US and beyond, as soon as 2024.
Given its already delivering produce for its partners, albeit with safety drivers in the cab, Tusimple is ahead of Tesla in terms of market penetration.
And as long as Elon Musk and co struggle to manufacture enough batteries to build their semis at scale, Tusimple’s approach, working in tandem with existing truck manufacturers, looks to remain viable.
Of course, autonomy is a risky business. It’s one thing for a self-driving taxi noodling along in Phoenix traffic at 25 miles an hour to go wrong.
A fully laden semi gunning along at 70 can do an awful lot of damage, and it remains to be seen how and to what extent US regulators are happy to let these potential death machines loose on the public highway.
Could they be hacked?
There’s also the not inconsiderable problem of Tusimple’s deep ties to China, not only in the C-suite but through its hefty financial backing from Sun Dream, an investment vehicle associated with the People’s Republic’s dominant social media platform Sina Weibo.
The Biden administration is looking carefully at any and all ties to a nation increasingly seen as a bitter geopolitical foe, with far-reaching implications for hopeful multinationals like Tusimple.
Tesla’s Semi should – if Elon Musk’s timescales are to be believed – start hitting the road later this year.
Yet Tusimple’s founding whizkid Hou is nonetheless bullish about his firm’s prospects for dominating the haulage business.
‘Autonomous driving is a problem that can be solved,’ he insists.
Only time will tell if he’s able to deliver the goods. What do you think?
Is the risk of a fully-loaded artic jackknifing because it’s gone haywire, or been hacked, worth paying less for melons at the grocery store?
