Tech Newsround – Robotaxi & AI Datapoints

Robotaxis: Yet another player.

• Zoox joins the ranks of robotaxi companies all vying to supply services to riders in yet another sign that Tesla is not the leader and that robotaxis will be won by scale and cost and not by technology.
• Zoox will launch in Las Vegas, but it has a long way to go to catch up with Waymo, which already operates in 5 US cities and offers 250,000 rides per week.
• It also has a long way to go to catch up with Pony.ai, WeRide and Baidu, all of whom have been offering some form of taxi service for a long time.
• However, Zoox is unique in that it will be the only one so far to launch with a vehicle built from the ground up for autonomous driving and, as such, has no driver controls.
• This is meaningful because it means that there can be no safety driver in the vehicle, meaning that Zoox has to be better at launch than its rivals.
• Its vehicle is also unique in that it is designed to travel in either direction, making it ideal for city navigation, but its unique movement abilities need specific training data, as it can move in ways other vehicles cannot.
• I suspect that this is what has taken it so long, but at the end of the data, I am not sure it matters very much.
• The reason why Amazon owns Zoox is for the in-house use case of logistics and distribution, where it spends billions and where an automated driving offering will get enough scale to save substantial cost without ever having to sell a single ride to the general public.
• Hence, this looks like an experiment to me and if it goes well, then a bigger service may emerge and if it does not, then it is good training for the in-house use case.
• However, it adds yet another competitor to the robotaxi scene, further undermining the idea that robotaxis are a high-margin business with technology providing barriers to entry.
• Instead, I have long argued that robotaxis will be a bloodbath of cutthroat competition where the winner will be determined by scale and cost and not by how good it is at driving.
• I can’t see Zoox winning this battle and suspect that once it has what it needs to power Amazon delivery, it will withdraw, although it will act as a spoiler while it is present.

AI Rollout: Still going strong.

• Monster earnings from Oracle and recent commentary from Texas Instruments confirm that growth in data centres driven by AI remains very much on track, putting to bed fears of an imminent correction.
• Oracle revealed that remaining performance obligation (signed deals that have not been completely delivered) increased by more than 4x YoY and has passed Google’s own backlog.
• This means that Oracle may soon be growing faster than Google Cloud, meaning that at some point soon, there could easily be another hyperscaler in the market.
• This, combined with commentary from Texas Instruments at an investment conference where it expects the portion of its revenues that come from data centres to continue expanding by 50% YoY, paints a very rosy picture for data centre spending over the next 12 months.
• Despite worries about poor returns from AI investments and falling deployments, demand for AI services clearly remains very robust.
• This is the main difference between this bubble and the Internet bubble 25 years ago, as back then dark fibre was laid in the hope that traffic would come.
• Today, data centres are being built like crazy, and we are very likely in a phase of overbuild, but crucially, there is already a lot of demand for capacity such that sales at some providers are being held up because they can’t build the capacity fast enough.
• This means that when the correction comes, it won’t be as deep or as long as the internet correction, but it will still be pretty painful, and I expect to see a number of players go out of business and be acquired.
• This means that it is the picks and shovels of the AI boom where the most gains will be experienced.
• Here, I would highlight Nvidia, Samsung, Qualcomm, TSMC, MediaTek, SK Hynix, Micron and so on as the companies that are likely to benefit the most, as they have revenue exposure and many of them are still quite reasonably valued.
• It’s the high-flying LLM-creators that I would want to avoid.