- The New Rules with Alan Pentz
- Posts
- Waymo vs Tesla and Natural Gas is the Answer to AI
Waymo vs Tesla and Natural Gas is the Answer to AI
What I’m reading, listening to, and watching
Alan Pentz
October 30, 2024
Welcome back to The New Rules
A quick message before we jump in:
If you’re reading this online, subscribe HERE and join over 8,000 people who want to understand the world better by receiving these emails every week.
And don’t forget about The New Rules podcast where I go deeper on my posts or bring in a guest whose ideas I find intriguing.
Let’s dive in…
The Information reports on a new Waymo funding round:
Autonomous vehicle company Waymo, a subsidiary of Google parent Alphabet, announced on Friday that it had raised $5.6 billion in a funding round led by its parent company. Other investors including Andreessen Horowitz, Fidelity and Silver Lake also participated in the fundraising, according to the company.
Waymo says it will use the capital to expand its robotaxi service and improve its autonomous driving technology “to support a variety of business applications over time.” The company has made efforts to commercialize its self-driving technology in different ways, including by running an autonomous trucking division. But in recent months it has been focused on expanding its ride-hailing program, including partnership with Uber to bring it to new U.S. cities including Atlanta and Austin.
I wrote about taking my first Waymo ride a couple of weeks ago in San Francisco. It was a pretty seamless experience. There are major questions about the ability of Waymo to get the costs down enough to make this version of autonomous driving work.
Waymo relies on a multiple sensors which makes the car expensive to produce. It’s unclear how much the cars cost all in but estimates have ranged from $160-300k. This raises the question about the scalability of Waymo’s approach.
In contrast, Tesla has chooses to rely on mostly cameras, vision, and AI for its version of self driving. It’s a cheaper way to go but Waymo’s approach has achieved a higher level of autonomy. As you can see in this graphic from NHTSA there are 6 levels of automation from 0-5. Waymo is at Level 4: the system can self drive within limited areas and with occasional need for (remote) humans to take control. As of now, a human only needs to intervene once ever 17,000 miles on average.
In contrast, Tesla is at Level 2: the car can provide assistance with driving as long as the driver is fully in control. Tesla believes it will win in the end for three reasons:
It’s cheaper.
Tesla has a large installed base already operating and gathering data.
It doesn’t rely on mapping limited areas. In theory, Tesla doesn’t need to pre-survey an area to operate. Any Tesla should be able to navigate based on its camera vision and AI. Waymo must pre-map an area in detail before it begins autonomous operations.
It seems like Tesla is making a bet on leapfrogging Waymo. If it works, it owns autonomous driving across the world. If not Waymo will slowly chip away at it going from city to city. I’d bet on Tesla long term but who knows?
One of the potential caps on the development of AI in the US at least is electricity production. Querying ChatGPT consumes about 10x as much electricity as a Google query does. Estimates of how much electricity we will need are exploding. The chart from Goldman Sachs below is one estimate but I’d say it’ll be even more.
There are all sorts of issues with expanding electricity capacity. Nuclear is probably the best way to do it but the timelines for building nuclear plants can stretch into the decades. Solar and wind can serve some needs but they are unreliable power sources and can’t scale to the need yet. In the end, it will be natural gas that does it. It’s cleaner than coal and oil and in the US it’s cheap and abundant but there’s still the issue of our overburdened electricity grid. New projects have to wait years sometimes to get hook in.
In an interview with Maggie Lake, Doomberg , an energy and commodity focused Substacker, introduced a novel idea. Companies could hook natural gas turbines directly into data centers and skip any connection to the electricity grid altogether. You could build the data centers right at the source. Looks like Texas is going to be the data centers capital of the world.
Keep learning,
Alan
P.S.
Reply