Editor’s Note: In just one week, on August 28, there’s going to be a major event in the artificial intelligence industry… and it will be the catalyst for a historic shift in the market.
That’s why Jeff called an AI Emergency Meeting last week… to help as many people as possible prepare for it. On the far side of this shift is the technological advancement the industry is racing toward – the next generation of artificial intelligence. Jeff calls it “Exegesis AI.”
This advanced artificial intelligence will eventually give us unlimited clean energy… drugs that will cure many diseases… personal humanoid robots… and much more. And we’re right on the cusp of it.
At his event, Jeff shared a handful of companies helping set the stage for this next generation of AI… as well as how you can prepare for the oncoming industry event set to shake up the market. It’s not too late for you to prepare for it… a replay of Jeff’s event is still available until midnight tonight.
Just go right here to access it. Then read on for today’s Bleeding Edge…
Who is Tesla’s biggest competitor?
If we asked just about anyone on the street, most would have an answer like Ford, General Motors, Toyota, Nissan, or maybe Rivian. And perhaps those more familiar with the China market for EVs would say Li Auto, XPeng, or BYD.
I’ve long maintained that Tesla doesn’t compete against traditional car companies. I don’t see Tesla as an automotive company so much as one of the world’s most advanced and successful artificial intelligence companies.
This is precisely why Tesla trades at valuation multiples consistent with a high-tech growth company, and not a car manufacturing company.
So if you asked me what company Tesla’s most important competitor is in the near future, my answer would probably surprise you…
Tesla’s closest competitor is Alphabet (Google).
Yes, I recognize their business models are entirely different. Almost 89% of Alphabet’s revenues come from the collection, and monetization, of our data through advertising revenues. That’s its business.
Tesla generates about 85% of its revenue from the sales of its electric vehicles (EVs) and another 8.6% from selling services, which include revenue from selling its full self-driving software (FSD).
But there is a fierce battle being fought between the two – a race most don’t see. And trillions of dollars are at stake.
These two companies have competitive visions for autonomous technology. How they monetize that technology is largely irrelevant.
This race began more than a decade ago with autonomous driving technology. Alphabet’s Waymo is now on its sixth-generation autonomous technology platform.
Shown below, this sensor-heavy autonomous vehicle comes with 13 cameras, 4 LiDAR units, and 6 radar units to enable its self-driving system.
Source: Waymo
Waymo’s autonomous driving software requires precise mapping that is developed by Waymo employees driving every inch of every street in a geofenced area. As a result, Waymo’s commercial services are currently restricted to small geographic areas in San Francisco, Los Angeles, and Phoenix, with some additional testing happening in Austin.
Waymo’s ride-hailing services now provide about 100,000 autonomous rides every week in those three cities. As of December last year, Waymo self-driving cars had driven 7.1 million miles.
Standing in stark contrast to Waymo, Tesla’s vision for self-driving technology is based on, well… vision.
We explored the “key to autonomy” in yesterday’s Bleeding Edge, Optimus – Working Around the Clock. That key is vision.
Tesla Model Y | Source: Tesla
As we look at the picture above, what do we see? Or better yet, what don’t we see?
None of Tesla’s models are laden with sensors around and on top of the cars. There are just eight cameras around each Tesla, and they are integrated so seamlessly into the EVs that most people don’t even see them.
Tesla’s autonomous software is trained and operates on vision, just like a human brain. Each one of a Tesla’s eight cameras has a 250-meter field of vision. Everything these cameras “see” is fed into Tesla’s FSD neural network… which analyzes the data and “makes sense” of it, all in split seconds.
Because of Tesla’s unique approach to autonomy, Tesla EVs are capable of driving anywhere, there is no geofencing at all. Any Tesla with full self-driving is fully capable of navigating roads that have never been traveled by another Tesla.
That’s because Tesla’s FSD AI has learned how to drive based on billions of miles of real-world video enabling it to apply those learnings to any driving environment. This is very similar to how we learn and apply our own knowledge and experiences to new environments.
Alphabet/Waymo is developing an autonomous operating system for cars that it eventually hopes to license to car manufacturers and robotaxi operators. Part of those future licensing agreements will be the right to both monitor and collect data on the passengers who ride in those cars using Alphabet/Waymo’s software.
Naturally, Alphabet wants as many cars as possible to use its software so that it can capture as much data as possible about consumers. That’s how it will monetize its investment in self-driving software: by selling access to that data to advertisers.
A similar battle is happening with robotics technology.
Earlier this month, Alphabet’s Google DeepMind division released some extraordinary research regarding artificial intelligence used in robotics. It didn’t get much attention, but it should have…
Shown below is a short video clip of the DeepMind AI playing table tennis.
Cleaning Up a Pan | Source: Stanford University
It’s pretty remarkable how well the DeepMind AI plays. To some, this might seem like a parlor trick and something insignificant. I assure you, it is not.
It takes years of practice for humans to reach a competitive level playing table tennis. To play well, one has to react in split seconds, play at high speed with precision, and use different shot techniques throughout a game against different opponents with different skill sets.
That’s what makes developing an AI for table tennis manifested in a robotic arm such a worthy challenge.
The Google DeepMind team provided its AI some limited video data on human-to-human play for training. Then it conducted further training in a simulation using reinforcement learning. After that, the DeepMind AI was trained further against human players to be able to play at a competitive level.
The results were impressive. The AI won all matches against beginner players, lost all matches against advanced players, and won 55% of the matches against intermediate players.
It’s not an exaggeration to say that DeepMind’s AI proved it was capable of human-level performance… and this was just its first version of this AI.
So, what?
Well, DeepMind has been working on what it calls Robotic Transformers (RT) and vision language models (VLMs) as they pertain to making robotics functional in the real world. Sound familiar?
This January, Alphabet, in collaboration with Stanford University, also released an open-source autonomous software platform for robotics – Mobile ALOHA. As stated, Mobile ALOHA is a “low-cost and whole-body teleoperation system for data collection.”
Let’s put aside the form factor of the hardware shown in the video below. The purpose of the team was to demonstrate the capabilities of the autonomous software based on vision inputs.
This is exactly the challenge Tesla is solving for with its Optimus humanoid robot.
Shown below is a short clip of the AI cleaning a pan after cooking. Those interested can find several more video clips right here.
Cleaning Up a Pan | Source: Stanford University
If we look back at the video of the robotic arm playing table tennis, it’s not Google DeepMind technology. We can see “ABB” in red letters on the top of the robotic arm. ABB is a massive Swiss conglomerate that has a strong robotics division.
The point is that Alphabet is hardware agnostic. It’s designing its software to run on any hardware platform.
Alphabet’s goal with robotics is the same as it is with cars. It is developing an AI – a robotic operating system – that it can license to robotics companies around the world, probably for free.
In exchange, those robotics companies will have to agree to provide the data collected by those robots to Alphabet – which will then categorize and monetize through advertising.
This is Google’s playbook.
They’ve done it before.
Google developed and gave away its Android OS smartphone and tablet operating system for free to the mobile phone industry… in exchange for the “right” to collect data on any user of an Android OS-powered smartphone.
This has been a fierce battle with Apple, but Alphabet/Google have the dominant market share (about 72%) of smartphones around the world.
That’s why Alphabet (GOOGL) will generate about $321 billion in revenue this year and is worth almost $2 trillion.
That’s also why Alphabet is racing in a battle to beat Tesla and its vision-based AI for its Optimus humanoid robot.
Tesla has developed a completely vertical technology solution – from the hardware to the semiconductors to the artificial intelligence.
Alphabet is focused on software and only hardware reference designs. Again, the intent is to make its software available, for free, in exchange for data collection.
Alphabet isn’t interested in the lower-margin hardware side of the business. It just wants the high-margin, future advertising revenues.
Even in smartphones, it has its Google Pixel smartphones contract manufactured, and they are really a reference design for the industry to follow.
The same is true for the Waymo autonomous vehicles. Waymo installs its software and sensor arrays on Jaguar I-PACEs, but the software can be adapted to any vehicle. The new sixth-generation car is from China-based Geely.
Alphabet can’t afford to have Tesla beat it to the consumer markets for both self-driving cars and home robotics. Losing those markets would result in trillions in revenues lost. The same is true for Tesla.
Tesla already has a massive lead over Alphabet/Waymo, having already driven more than 5 billion miles on Autopilot or full self-driving. And every Tesla shipped is capable of operating the FSD software.
And Tesla is making extraordinarily rapid progress with not just its Optimus hardware, but its autonomous robotic software, which benefits directly from the success of the autonomous driving software.
Both Tesla and Alphabet want to be dominant in autonomous systems, both for vehicles and consumer robotics.
They will monetize their technology differently, but it is very clear, they are competing for the same markets.
The Bleeding Edge is the only free newsletter that delivers daily insights and information from the high-tech world as well as topics and trends relevant to investments.
The Bleeding Edge is the only free newsletter that delivers daily insights and information from the high-tech world as well as topics and trends relevant to investments.
