The creative and talented team over at Munro Live has shared some exciting new information about Tesla’s highly anticipated HW4.
Last month, Tesla hacker and owner Greentheonly leaked photos of the HW4 computer, citing a myriad of new features and improvements. The new computer will allegedly be 2-4x faster than HW3. A portion of this computing power will likely go toward processing additional data from the upgraded 5MP cameras.
Along with the upgraded 5MP cameras come some additional changes that make it hard for Tesla to retrofit HW4 into previous vehicles. The layout of the HW is quite different and, as noted in Munro Live’s video, the board is significantly larger, among other things.
The Power of Tesla
One thing in particular the Munro & Associates team uncovered was that Tesla removed the separate graphics card and integrated it directly onto the MCU circuit board. The Munro team said this would increase performance by removing the communication bottleneck between the two boards. Furthermore, this integration will save Tesla around $2 per vehicle due to removing some small parts and it will help streamline manufacturing.
Not Only for Cost Savings
This speaks volumes about the power and manufacturing abilities of Tesla. While $2 per vehicle isn't a big deal on its own, the cost savings alone, which are expected to be up to 4 million dollars this year, will more than pay for the engineering time spent on making these changes. However, the bigger win will come in having fewer parts to handle and assemble when building a vehicle. It will also help increase the vehicle's reliability by having fewer failure points.
The AMD-powered infotainment processor is unchanged in HW4, but the processor utilized for taking the input from the cameras and sensors is now larger and received a performance boost. Tesla is expected to transition all vehicles to hardware 4 this year, including the upcoming Cybertruck (new Cybertruck features uncovered at Investor Day)which is expected to begin manufacturing this summer.
Watch Munro Live’s HW4 Teardown
Subscribe
Subscribe to our newsletter to stay up to date on the latest Tesla news, upcoming features and software updates.
During the unveiling of the Cybercab at We, Robot last year, we saw the Cybercab use a strange laser-like lighting system to project an interesting light pattern. It was a rather peculiar projection, and many observers passed over it without wondering too much - perhaps it was just a trick of the show?
But we went back and thought - what is it really? What kind of lighting technology could Tesla be using to accomplish this visual effect, and does it have a purpose? That led us down the rabbit hole of Structured Light, and we think that Tesla could be using this technique for a few different things.
Cybercab’s Autonomy
Unlike other autonomous vehicles, which are often retrofitted with complex and expensive sensor arrays, the Cybercab is purpose-built to be a mass-market autonomous taxi. Because of that, Tesla is using several new techniques to drive down the cost and simplify the vehicle, from the paint-impregnated panels to the easy-to-clean interior.
However, the drive to keep costs down hasn’t stopped Tesla from looking for ways to present a premium and safe product, which has been one of the Engineering and Design team’s goals for a long time.
Structured Light is a technique involving the projection of light using a distinct pattern (see it in action below). In particular, it’s a machine vision technique for capturing 3D information. Unlike normal, ‘passive’ cameras that interpret ambient light, structured light actively projects a known pattern, such as lines, grids, or dots, onto the environment.
A camera will then observe how the pattern deforms as it hits objects. By analyzing these distortions, the system can precisely calculate the 3D shape and depth of the scene in front of it. Structured light can generate dense, high-resolution 3D data, which is particularly useful at short ranges.
But is this useful for an autonomous vehicle? Potentially. It offers a close-range means of mapping the environment directly in front of the vehicle, and could theoretically complement the existing camera suite for precise maneuvering in tight spaces. It could also be used for detecting low-lying objects, although this task could be handled by the bumper camera that is arriving on newer Tesla vehicles as well.
Structured Light vs LiDAR
Not a Tesla App
The mention of an active 3D sensor brings to mind LiDAR (Light Detecting and Ranging) - which Elon Musk has repeatedly called a crutch for self-driving. But LiDAR is different than Structured Light - it emits a pulse of laser light and measures the time of flight to generate a 3D point cloud of the environment.
While both involve projecting light, structured light systems require specialized hardware distinct from Tesla's current Matrix LED headlights. Matrix headlights are designed to intelligently block parts of the main beam to prevent dazzling other drivers, using an array of controllable LEDs.
Structured light, conversely, requires a dedicated projector to cast a specific, known pattern (dots, lines, or a grid) precisely so a camera can analyze its deformation for 3D mapping. Implementing structured light would likely involve adding a new, dedicated projector component - perhaps another Cybercab exclusive feature.
Tesla already uses computer vision to navigate the environment, and adding Structured Light could potentially aid this process for Robotaxi challenges and edge cases, while not deviating from Tesla’s policy to use the same senses that a human uses to drive.
Structured light could be Tesla’s potential intelligent solution to an active 3D sensing system without adopting a more traditional approach like LiDAR.
Other Uses
Beyond the possibility of 3D sensing, could this light projection serve other purposes? Tesla is known for integrating novel new features into its vehicles - sometimes practical, sometimes just for fun. There are a few possibilities here.
Visual Communication - The lights could be used to project information onto a road surface - perhaps showing the vehicle’s intended path, or displaying warnings to pedestrians. It could even project a virtual crosswalk to signal that it is safe for pedestrians to pass.
Passenger Guidance - For a taxi without a driver, guiding passengers on how to approach and use the Cybercab could be useful. A projected welcome mat could show passengers exactly how to get into the Cybercab, reveal which passengers are being picked up, or provide other important information.
Improved Light Shows - We all know Tesla loves the little things in life. The projection could be used for more whimsical Light Shows in the future. We’re sure they’ll have more than just that in mind, though.
Wrapping Up
While we haven’t had any official confirmation that Tesla is using Structured Light, the possibility of the Cybercab using a novel new technology is relatively fascinating. It suggests that Tesla is looking into using its pre-existing tools and features in even more ways - something that they’ve always focused on. Whether this projection is for sensing or communicating, or even just for fun as a light show, we won’t know until Cybercab actually hits production.
One idea that consistently comes up when discussing the future potential of FSD is the ability to order a Tesla and have it arrive at your home—completely autonomously.
Picture this: a car rolls off the assembly line at Giga Texas, drives itself to the outbound delivery lot, and merges onto the highway—completely driverless. Its destination? A customer in upstate New York. Along the way, the vehicle independently plans its charging stops, navigates traffic, and drives nonstop, no breaks needed, until it pulls into the buyer’s driveway.
Of course, there’s a lot to consider and break down - but Elon Musk doubled down and said that he intends to have Tesla vehicles deliver themselves autonomously by the end of 2025. We’re sure there are some stipulations to that, but it’s an exciting idea that could save Tesla thousands per vehicle sold.
The core idea is that FSD Unsupervised is nearing completeness, with the upcoming launch of Tesla’s first Robotaxi network slated for June. Once Tesla can demonstrate that Unsupervised driving is safe, it will expand beyond simply transporting passengers from Point A to Point B.
That will include moving vehicles from factory to consumer, which will vastly streamline Tesla’s delivery process - and really make a statement about how far Tesla is ahead of the competition.
Reduced Delivery Costs
One of the biggest keys for Tesla throughout its history has been its ability to drive down costs. Delivery costs add to the price of a vehicle, and it’s not just the price of transporting the vehicle, but the cost of the whole delivery experience — delivery center and personnel included.
Instead, if a future Tesla drives itself directly to the consumer, Tesla can further reduce the price of its vehicles or improve its profits.
Improved Customer Experience
Money aside, the real game-changer is that no other vehicle delivery experience will come close to Tesla’s. What other brand will be able to deliver a car that drives itself to your home—straight from the factory?
Having a vehicle deliver itself can also provide other benefits, such as faster delivery times for customers. The customer also won’t need to drive to a delivery center to pick up their vehicle. The vehicle would go directly from the factory to the customer without any steps in between.
There’s something undeniably futuristic about getting a notification on your phone, glancing outside, and seeing your brand-new Tesla autonomously pulling into your driveway.
Limitations
Of course, with any new process or feature - there will be limitations and hurdles for Tesla to address, and these will take time and effort.
Let’s discuss some of the hurdles Tesla will need to solve.
Charging Infrastructure
Besides achieving true autonomy, another big challenge will be charging infrastructure. No Tesla can currently plug itself in at a Supercharger - they’ll need access to the upcoming V4 Superchargers with Wireless Charging to take on a road trip autonomously.
That will be a challenging rollout in the initial few years, especially as Tesla hasn’t been actively replacing older V2 Supercharger sites just yet - and is instead focusing on deploying new ones. We imagine that the rollout of wireless charging will likely also be limited, and it is likely that only Tesla’s more premium vehicles and Robotaxis will receive wireless charging compatibility.
Autonomy Policies / Geographic Limitations
Tesla has been consistently hamstrung by red tape, and unsupervised deliveries will also be a massive hurdle to clear with many states and provinces, as well as at the federal level. That will be an implicit challenge for Tesla in the infancy of the Robotaxi network as well, but once approvals spread wider, we can see these restrictions being lifted and relaxed.
Mileage and Damage
One item that could be a sticking point for customers is mileage and damage. If your brand-new vehicle was damaged on its trip over, what would you do? You’d have to take it from your driveway right to the Service Center, effectively invalidating the entire experience.
Of course, many people will also want to see a low odometer on arrival of a brand new, fresh-from-factory vehicle - and that’s understandable. Having a car arrive with several hundred or even a thousand miles of wear and tear on the tires and paint could be a big negative as well.
What Is Likely to Happen
We think that Tesla will continue to ship vehicles longer distances - but perhaps once they’re offloaded at your local Service Center, an employee taps a button, and begins the self-delivery process. That eliminates many of the above problems we’ve outlined, including charging and additional mileage, while keeping many of the advantages intact.
For any customers who order a Tesla in Austin later this year, it’s possible Tesla could have the vehicle delivered directly to you.
_300w.png)
![Tesla’s Missing Voice: Why a PR Team Matters More Than Ever [Opinion]](https://www.notateslaapp.com/img/containers/article_images/multiple-models/group_93.jpg/4e1056961f0480c7b9eff43dd2ec288e/group_93.jpg)
