Tesla Officially Announces FSD Hardware 5.0 and How It Compares to Hardware 4.0 [VIDEO]

By Karan Singh
Tesla's naming hardware 5.0, AI 5
Tesla's naming hardware 5.0, AI 5
Not a Tesla App

At this year’s Tesla shareholder meeting, Elon Musk announced a slate of changes for Tesla’s FSD hardware. For the first time, Musk talked openly about details of Tesla’s upcoming FSD Hardware 5, or “AI 5” and the differences between hardware 3 and hardware 4.

Hardware 4.0 Divergence

Currently, hardware 4.0 runs Hardware 3 in emulation mode, but with the completion of Tesla’s new AI-focused Supercomputer cluster, hardware 4 will eventually diverge from Hardware 3. Emulation mode simulates hardware 3, which means that today, hardware 4 is very similar to hardware 3. Elon Musk confirmed that the new cluster will be entirely focused on hardware 4 training, rather than hardware 3.

Hardware 4’s cameras are approximately 4-5x better in terms of resolution and fidelity than hardware 3, and hardware 4.0 is about 3-5x faster and powerful in inference capabilities. These are some pretty significant numbers, and Musk said he expects hardware 4.0 to eventually be about 5x better than hardware 3.0.

Hardware 3

While hardware 3 is no longer being produced, Tesla has said in the past that it will be capable of self-driving at a level that is safer than humans. Elon Musk confirmed on stage that hardware 3 would continue to receive training and would still have new FSD versions developed for it, however, hardware 4 would eventually be “about 5 times” better than hardware 3.

At some point hardware 3 will hit the end of its service life, but that’s not expected for at least several years.

Sadly, Tesla has previously confirmed that they do not intend to have a hardware 3 to hardware 4 retrofit, as the size of the MCU and electrical harnesses differ between vehicle hardware iterations.

Hardware 5.0 - “AI 5”

Elon Musk also officially announced Tesla’s FSD hardware 5.0, which he says Tesla is now calling “AI 5”. AI 5 is expected to hit the production lines for customer vehicles in approximately 18 months - around December 2025, with a massive slate of improvements. It is expected to be approximately 10x better than hardware 4.0, and up to 50x better in terms of inference power alone.

Of course, these massive improvements don’t come without a cost – AI 5 will consume up to 800 watts of power. In comparison, hardware 3 and hardware 4 use about 200 watts today, so don’t expect any upgrades from hardware 4. Like most modern processors, AI 5 can adjust its power efficiency. It’ll be able to ramp up or down its processing power depending on the complexity of the scene and the power needed.

We can expect that AI 5 will likely run in emulation mode when it arrives, and improvements between hardware 4 and AI 5 will take some time, similar to how hardware 4 is still in emulation mode today.

Musk Talking About AI 5

Musk talks about Tesla’s FSD hardware below during Tesla’s 2024 shareholder’s meeting, where he officially talked about Tesla’s upcoming hardware 5.0 for the first time.

Smart Assistant 'Grok' Coming to Tesla, Demo of Grok 3 Tonight

By Karan Singh
Not a Tesla App

xAI’s Grok AI Model has been receiving rapid updates, similar to Tesla’s FSD. After waiting for a major update, the latest version—Grok 3—is set to launch with a live demo on February 17th.

Grok 3 is supposed to bring big improvements, as xAI has changed its approach to training the Large Language Model (LLM). Unlike many other LLMs, Grok can access real-time information and adapt its answers based on the latest news and discussions happening on X’s platform, making it more suited for current events or breaking news.

At the World Governments Summit in Dubai, Elon also mentioned that Grok 3 is already outperforming its competitors, so there is a lot to look forward to. Let’s dig into the new capabilities and then see why they matter to Tesla drivers.

New Capabilities

This time, Grok is set to feature self-correction, reinforcement learning, and synthetic datasets. Self-correction allows Grok to identify and fix its own mistakes, with human oversight ensuring that the process is guided. This helps refine Grok’s future responses as part of its ongoing training.

Reinforcement learning, meanwhile, rewards or penalizes the AI based on its actions. When it performs correctly, it receives a reward, reinforcing that behavior; penalties reduce the likelihood of undesirable actions. Combined with new synthetic datasets, xAI can accelerate Grok’s training process, much like how Tesla’s AI team has been refining FSD. There’s definitely a lot of overlap between the two.

Voice Support

Grok is one of the few remaining publicly available AIs that doesn’t have a live-voice mode. Its competitors - including Google’s Gemini and OpenAI’s ChatGPT - both have live voice response and reasoning and will also integrate and hook into 3rd-party services over voice.

That’s going to be a key feature as Tesla looks to Grok to power its upcoming voice assistant in their vehicles.

Tesla Demo

Musk has previously stated that all Tesla vehicles will eventually receive Grok as a Smart Voice Assistant—one far more capable than the current voice command system.

With changes to the Navigation Voice globally, as well as the “Hi” command finally returning a response, we’re on the verge of Tesla’s integration of a voice assistant.

Tonight’s demo, at 8 pm PT is expected to be live, potentially showing off the integration into Tesla vehicles. Millions of existing Tesla vehicles worldwide could soon take advantage of Grok as a voice assistant. There’s even a rumor of a wake word for the feature — potentially Grok.

Grok’s real-time capabilities mean you could ask it to add a stop to your route for food at the highest-rated local restaurant with minimal wait. Grok would gather and process that information while your car drives itself, then update your route, allowing FSD to take you directly there. You’ll also be able to ask general knowledge questions, such as the weather for tomorrow or the price of Tesla stock.

Controlling the vehicle will also be greatly upgraded. While the current voice command system can handle simple tasks like “Open Charge Port” or “Turn on Defrost,” the new system should be able to handle more complex requests and allow drivers to ask things in a more natural way.

We expect to find out more about Grok in Teslas tonight. Even if the demo tonight doesn’t specifically mention Tesla, a look at Grok’s new abilities could be a first look at what we’ll soon be able to experience in our Teslas.

Tesla Executives Confirm Robotaxi on Track for June

By Karan Singh
Not a Tesla App

At the Tesla Q4 2024 Earnings Call, we found out that Tesla now has a firm date and set location for the first deployment of its Robotaxi fleet. Now, several of Tesla's executives have re-iterated that deployment date and location in several different places.

Like most Tesla fans and investors, we’re excited to see Cybercab make its official debut on the streets of Austin, Texas, sometime in June. That’s potentially less than 100 days away - but who’s counting, right?

Elon Confirms

Elon Musk confirmed on X that Tesla is still on track to launch autonomous ride-hailing (Robotaxi) in Austin this June. Plus, they’re also working on rolling it out to even more cities in the United States by the end of the year.

Tesla’s benchmark for expanding deployments is ensuring that Robotaxi safety surpasses the average human driver by a wide margin. Based on the latest Vehicle Safety Report, Supervised FSD is already making strong progress, outperforming the average driver by 10.5 times—a full order of magnitude.

Franz Confirms

Franz Von Holzhausen, Tesla’s Head of Vehicle Design, also confirmed that Tesla will be offering Cybercab rides in Austin starting in June. What’s key here is that he confirmed the presence of Cybercabs finally deploying - it won’t be driverless Model Ys or Model 3s - it’ll be the Cybercab.

That means an autonomy-first vehicle without a driver’s seat, steering wheel, or pedals will be on the road and driving people from point to point. Major autonomy competitors like Waymo use heavily modified EVs that still have seats and vehicle controls intact. An autonomy-first design increases passenger room and storage while also improving efficiency by reducing vehicle weight.

Are you planning on making your way down to Austin to experience the Robotaxi experience first-hand or are you planning on waiting until it’s deployed in a city near you? Let us know your thoughts in our forums or social media.

Latest Tesla Update

Confirmed by Elon

Take a look at features that Elon Musk has said will be coming soon.

More Tesla News

Tesla Videos

Latest Tesla Update

Confirmed by Elon

Take a look at features that Elon Musk has said will be coming soon.

Subscribe

Subscribe to our weekly newsletter