Productivity and the product of productivity, economic impact

Increasingly there is economic thought estimating the big lift from AI as to at the minimum double economic product. Of course we are still early and the commercial market is very unsettled. If you look to the longer term estimates of some of the bright lights of the tech sphere, you will see the upside exceeds anything the industry leadership has ever seen in terms of potential. Now things come down to how to realize the absolute best future AI makes possible, and to see that benefits are widely felt in terms of public prosperity. The more leadership sees advancement, the larger they foresee the opportunity. 

Capacity

China is operating at something around 55% factory utilization for their automotive industry. Now that there is softness in the Chinese market, they are trying to expand their exports, which puts pressure on jobs and capacity in other global geographies. As we have repeatedly warned, China will bring pain to the United States' and European industries. That has already begun, with VW of Germany for example preparing to close European factories. Also all the manufacturers from the West are having to emulate Tesla related to software projects. The logical case we have been emphasizing, of the 80 plants being built to shift the supply chains to the US from China, are going to be met with fairly low demand as Ford and GM have ratcheted back plans at least four times each. This means they will have less demand, which means less factory utilization, which in turn means excess capacity, which in turn leads to unprofitable capacity utilization. The scale of the projects underway are almost certain to be met with lukewarm uptake. As I have stated before, it would be a lot better to see 90% utilization of forty plants instead of 45% utilization of 80 plants. 

The industry has sharply reduced and pulled back their plans, while government projects are operating in a disconnected fashion somewhat oblivious to probable outcomes and any balance between supply and demand. This means mega billion dollar messes may well be on our hands in a year or so. 

Burn Rate

Electric vehicle startups are burning too much capital. So are semiconductor geographical moves, so are start up AI companies that are not under the wings of a Super Tech. For Open AI or Tesla AI this equates to about 5 billion dollars a quarter. Many of the start ups may have to be sold to backstop the research budgets with cash. Certainly the very large investments needed to for example build out training systems with hundred plus thousand processors are investments only so many companies can make. Tesla did brilliantly pivot and made the bold move to rush a Nvidia majority supercomputer to preempt some of the competition. 

Larry Ellison for example has said that there is likely to only be one huge winner in AI, He invested alongside Musk to make his move. Similar actions were taken by Microsoft, Open Ai, Meta, and more. 

Units and Profitability

The US industry is reducing in annual units, that trend continues, as only a portion of the American Demand base can afford new vehicles. China is also softening, which makes the issues of capacity very relevant for all parties. Aligned with that is a profitability decline as electrics are not in virtuous abundance yet for good ev economics. China was also marginally profitable, now that has spread to the US. The US is now operating around 10% of new vehicles being electric, Europe around 15%, and China at 50%. Sounds like a lot until you realize Norway crushed it and is now at about 97% of new vehicles being electric. 

State of the Advancement

The pace of accelerated advancement is now such that in the next year it is expected PhD level results look to replace the current College Freshman level. There are also more reasoning type processes supported by software. AI Robotic systems work off the AI advancements so they are a beneficiary of this progress too.  There are plenty of improvement paths to keep the product advancing. So it looks like a positive track forward is quite realistic. 

Super Intelligence as compared with General Intelligence

The historic test was could software with computers astutely interact so that it seemed to meet human standards. Now though the game is moving on as once systems can really design, create, and improve upon both hardware and software designs then quickly there is little pause in general intelligence and quickly we move on to enhanced machine intelligence, quickly moving past human limitations as the Computer supplants all human based efforts. Increasingly leadership thinks there is not too much time between generalized intelligence and moving to Super intelligence.  At that stage what is possible starts to be beyond human estimation. 

Musk makes move for 100k processor training supercomputer

He invested a lot, moved fast with Nvidia, and partnered with Oracle. Some of the constraints of future AI are processing compute, electrical power, talented teams, and capital. At great risk Musk made his move, to be sure there are others fast following now. 

Meta right behind him

They also have made announcements of 100k processor training systems, it will be interesting to see what is ready for prime time and what is more press release than bridge to Super Intelligence. 

Larry Ellison of Oracle, comments

Ellison has seen tech battles before, and he emphasizes that there will be probably one outsized winner. As much of the industry views the caliber of the training system to be critical, many of the Super Techs have invested in training and data centers too. Temporarily everyone wanted to be on Nvidia's dance card as the favored partner. That type of relationship is not particularly stable. There will be moves and deceptions, as all try and build out an advantage in a rapidly moving market. 

Summary

Electric cars are expanding sales. China remains the champion in terms of production and market uptake in volume. Electrics are not generally making much money. AI is increasingly looking to advance as to make AI Robotics more capable than schedule had generally expected. There is still a growing rift between Industry supply intentions and US Government Capacity Build outs of probably excessive and underutilized capacity.and supply. So far it is very unclear if the government will adjust to this illogical disconnect.  

To be sure though, the intelligence level is growing rapidly and lends strong credence that AI Robotics is a high odds play of massive impact. 

Despite global wars targeting energy infrastructure, the price of oil is moderated, and alternatives are building out capabilities. China, running at a run rate of about 50% new vehicles being electrics is sending a message to all oil markets, now and in the future. Price runaways will be moderated thankfully by electric vehicles. And too, coming are Ai and Autonomous offerings that will moot the old car with one driver and the costly one owner ownership model.