Blog

Adam Jonas: Welcome to Thoughts on the Market. I’m Adam Jonas. I lead Morgan Stanley’s Research Department’s efforts on embodied AI and humanoid robots.

 

Tim Hsiao: And I’m c, Greater China Auto Analyst.

 

Adam Jonas: Today – how the global auto industry is evolving from horsepower to brainpower with the help of AI.

 

It’s Thursday, August 21st at 9am in New York.

 

Tim Hsiao: And 9pm in Hong Kong.

 

Adam Jonas: From Detroit to Stuttgart to Shanghai, automakers are making big investments in AI. In fact, AI is the engine behind what we think will be a $200 billion self-driving vehicle market by 2030. Tim, you believe that nearly 30 percent of vehicles sold globally by 2030 will be equipped with Level 2+ smart driving features that can control steering, acceleration, braking, and even some hands-off driving. We expect China to account for 60 percent of these vehicles by 2030.

 

What’s driving this rapid adoption in China and how does it compare to the rest of the world?

 

Tim Hsiao: China has the largest EV market globally, and the country’s EV sales are not only making up over 50 percent of the new car sales locally in China but also accounting for over 50 percent of our global EV sales. As a result, the market is experiencing intense competition. And the car makers are keen to differentiate with the technological innovation, to which smart driving serve as the most effective means. This together with the AI breakthrough, enables China to aggressively roll out Level 2+ urban navigation on autopilot. In the meantime, Chinese government support and cost competitive supply chains also help.

 

So, we are looking for China’s the adoption of Level 2+ smart driving on passenger vehicle to reach 25 percent by end of this year, and the 60 percent by 2030 versus 6 percent and the 17 percent for the rest of the world during the same period.

 

Adam Jonas: How is China balancing an aggressive rollout with safety and compliance, especially as it moves towards even greater vehicle automation going forward?

 

Tim Hsiao: Right.  That’s a great and a relevant question because over the years, China has made significant strides in developing a comprehensive regulatory framework for autonomous vehicles. For example, China was already implementing its strategies for innovation and the development of autonomous vehicles in 2022 and had proved several auto OEM to roll out Level 3 pilot programs in 2023.

 

Although China has been implementing stricter requirements since early this year, for example, banning the terms like autonomous driving in advertisement and requiring stricter testing, we still believe more detailed industry standard and regulatory measures will facilitate development and adoption of Level 2+ Smart driving. And this is important to prevent, you know, the bad money from driving out goods.

 

Adam Jonas: , One way people might encounter this technology is through robotaxis. Robo taxis are gaining traction in China’s major cities, as you’ve been reporting. What’s the outlook for Level 4 adoption and how would this reshape urban mobility?

 

Tim Hsiao: The size of Level 4+ robotaxi fleet stays small at the moment in China, with less than 1 percent penetration rate. But we’ve started seeing accelerating roll out of robotaxi operation in major cities since early this year. So, by 2030, we are looking for  Level 4+ robotaxis to account for 8 percent of China’s total taxi and ride sharing fleet size by 2030. So, this adoption is facilitated by robust regulatory frameworks, including designated test zones and the clear safety guidance. We believe the proliferation of a Level 4 robotaxi will eventually reshape the urban mobility by meaningfully reducing transportation costs, alleviating traffic congestion through optimized routing and potentially reducing accidents.

 

So, Adam, that’s the outlook for China. But looking at the global trends beyond China, what are the biggest global revenue opportunities in your view? Is that going to be hardware, software, or something else?

 

Adam Jonas:  We are entering a new scientific era where the AI world, the software world is coming into far greater mental contact, and physical contact, with the hardware world and the physical world of manufacturing. And it’s being driven by corporate rivalry amongst not just the terra cap, you know, super large cap companies, but also between public and private companies and competition. And then it’s being also fueled by geopolitical rivalry and social issues as well, on a global scale. So, we’re actually creating an entirely new species. This robotic species that yes, is expressed in many ways on our roads in China and globally, but it’s just the beginning.

 

In terms of whether it’s hardware, software, or something else – it’s all the above. What we’ve done with a across 40 sectors at Morgan Stanley is to divide the robot, whether it flies, drives, walks, crawls, whatever – we divide it into the brain and the body. And the brain can be divided into sensors and memory and compute and foundational models and simulation. The body can be broken up into actuators, the kind of motor neuron capability, the connective tissue, the batteries. And then there’s integrators, that kind of do it all – the hardware, the software, the integration, the training, the data, the compute, the energy, the infrastructure. And so, what’s so exciting about this opportunity for our clients is there’s no one way to do it. There’s no one region to do it.

 

. So, stick with us folks. There’s a lot of – not just revenue opportunities – but alpha generating opportunities as well.

 

Tim Hsiao: We are seeing OEMs pivot from cars to humanoids and the electric vertical takeoff in the landing vehicles or EVOTL. Our listeners may have seen videos of these vehicles, which are like helicopters and are designed for urban air mobility. How realistic is this transition and what’s the timeline for commercialization in your view?

 

Adam Jonas:  Anything that can be electrified will be electrified. Anything that can be automated will be automated. And the advancement of the state of the art in robotaxis and Level 2, Level 3, Level 4+ autonomy is directly transferrable to aviation.   There’s obviously different regulatory and safety aspects of aviation, the air traffic control and the FAA and the equivalent regulatory bodies in Europe and in and in China that we will have to navigate, pun intended. But we will get there. We will get there ultimately because taking these technologies of automation and electronic and software defined technology into the low altitude economy will be a superior experience and a vastly cheaper experience. Point to point, on a per person, per passenger, per ton, per mile basis.

 

So  the Wright brothers can finally get excited that their invention

 

from 1903, quite a long time ago  could finally, really change how humans live and move around the surface of the earth, even beyond,  few tens of thousands of commercial and private aircraft that exist today.

 

Tim Hsiao:  The other key questions or key focus for investors is about the business model. So, until now, the auto industry has centered on the car ownership model. But with this new technology, we’ve been hearing a new model, as you just mentioned, the sheer mobility and the autonomous driving fleet. Experts say it could be major disruptor in this sector. So, what’s your take on how this will evolve in developed and emerging markets?

 

Adam Jonas: Well, we think when you take autonomous and shared and electric mobility all the way – that transportation starts to resemble a utility like electricity or water or telecom; where the incremental mile traveled is maybe not quite free, but very, very, very low cost. Maybe only the marginal cost of the mile traveled may only just be the energy required to deliver that mile, whether it’s a renewable or non-renewable energy source.

 

And the relationship with a car will change a lot. Individual vehicle ownership may go the way of horse ownership. There will be some, but it’ll be seen as a nostalgic privilege, if you will, to own our own car. Others would say, I don’t want to own my own car. This is crazy. Why would anyone want to do that?

 

So, it’s going to really transform the business model. It will, I think, change the structure of the industry in terms of the number of participants and what they do. Not everybody will win. Some of the existing players can win. But they might have to make some uncomfortable trade-offs for survival.  And for others, the car – let’s say terrestrial vehicle modality may just be a small part of a broader robotics and then physical embodiment of AI that they’re propagating; where auto will just be a really, really just one tendril of many, many dozens of different tendrils. So again, it’s beginning now. This process will take decades to play out. But investors with even, you know, two-to-three or three-to-five-year view can take steps today to adjust their portfolios and position themselves.

 

Tim Hsiao:  The other key focus of the investor over the market would definitely be the geopolitical dynamics. So, Morgan Stanley expects to see a lot of what you call coopetition between global OEMs and the Chinese suppliers. What do you mean by coopetition and how do you see this dynamic playing out, especially in terms of the tech deflation?

 

Adam Jonas: In order to reduce the United States dependency on China, we need to work with China. So, there’s the irony here. Look, in my former life of being an auto analyst, every auto CEO I speak to does not believe that tariffs will limit Chinese involvement in the global auto industry, including onshore in the United States.

 

Many are actively seeking to work with the Chinese through various structures  give them an on-ramp to move onshore to produce their, in many cases, superior products, but in U.S. factories on U.S. shores with American workers. That might lead to some, again, trade-offs.

 

But our view within Morgan Stanley and working with you is we do think that there are on-ramps for Chinese hardware, Chinese knowhow, and Chinese electrical vehicle architecture, but while still being sensitive to the dual-purpose AI sensitivities around  software and the AI networks that for national security reasons, nations want to have more control over.

 

So  that’s a long-winded way of saying we want the Chinese body, but we may need to find a way to have the U.S. brain in that body. And I actually am hopeful and seeing some signs already that that’s going to happen and play out over the next six to 12 months.

 

Tim Hsiao: I would say it’s clear that the road ahead isn’t just smarter, it’s faster, more connected, and increasingly autonomous.

 

Adam Jonas: That’s correct, Tim. I could not agree more.  Thanks for joining me on the show today.

 

Tim Hsiao: Thanks, Adam. Always a pleasure.

 

Adam Jonas: And to our listeners, thanks for listening. Until next time, stay human and keep driving forward. If you enjoy Thoughts on the Market,  please leave us a review wherever you listen and share the podcast with a friend or colleague today.

 

 

Isolation of WJ-MSCs

Discarded human umbilical cords were obtained at the University Hospital Plzen (Plzen, Czech Republic) from healthy full-term neonates (N = 4) after spontaneous delivery, according to all ethical guidelines. About 10 cm of the umbilical cord was aseptically collected, placed in sterile PBS with antibiotic–antimycotic solution at 4 °C, and transported to the laboratory within 24 h. After washing several times in PBS and brief exposure to 10% Betadine (EGIS Pharmaceuticals PLC, Budapest, Hungary), blood vessels were removed, the remaining Wharton’s jelly tissue was chopped into small fragments and the cells were isolated by enzymatic digestion in PBS-AA solution containing 0.26 U/ml Liberase-TM™ (Roche Custom Biotech, Mannheim, Germany) and 1 mg/ml hyaluronidase at 37 °C with constant shaking for 2 h. After removing undigested fragments with 40-µm cell strainers, cells were centrifuged at 450 × g for 10 min^6,38.

Isolation of AT-MSCs

AT samples were obtained from healthy volunteers (N = 4) who had undergone liposuction procedures for aesthetic reasons and processed as previously described⁶. Briefly, the lipoaspirate was repeatedly washed in PBS and enzymatically digested by 0.3 PzU/ml collagenase type I (Thermo Fisher Scientific) at 37 °C for 2 h with constant shaking. After centrifugation at 200 × g for 10 min, the stromal vascular fraction was washed twice with PBS and expanded.

Cell culture and expansion

Primary cell suspensions were diluted in a complete culture medium containing α-MEM (Gibco^®, Thermo Fisher Scientific), 5% human platelet lysate (HPL, Bioinova a.s.), penicillin/streptomycin (Gibco^®, Thermo Fisher Scientific), and GlutaMAX (Gibco^®, Thermo Fisher Scientific). Cells were cultured at 37 °C in a humidified atmosphere with 5% CO₂. Regular media changes were done twice a week. After reaching near-confluence, cells were harvested using the 0.05% Trypsin/EDTA solution (Gibco^®, Thermo Fisher Scientific) and reseeded onto a fresh plastic surface (Nunc, Roskilde, Denmark) at a density of 5 × 10³ cells/cm². Cells in passage 2 were cryopreserved in culture medium, supplemented with 10% DMSO, using a 1 °C/min cooling rate down to 80 °C, and stored at -196 °C until further use. Before use, the cryopreserved cell cultures were thawed at 37–40 °C and additionally expanded in the α-MEM medium containing GlutaMAX, supplemented with 10% fetal bovine serum (FBS), 100 units/mL of penicillin, and 100 µg/mL of streptomycin (all from Gibco^®, Thermo Fisher Scientific). Cells from passages 3–5 were used for subsequent experiments.

Adipogenic differentiation

For adipogenic differentiation, AT- and WJ-MSCs were cultured for 21 days in a medium composed of α-MEM, containing 10% FBS, penicillin/streptomycin, GlutaMAX, 0.5 mM 3-isobutyl-1-methylxanthine, 0.1 µM dexamethasone, 0.1 mM indomethacin, and 10 µg/ml insulin (all from Merck KGaA, Darmstadt, Germany). The medium was replaced every 3 days.

In separate experiments, the induction medium was additionally supplemented by 100 µM Oleic Acid (OA, Merck KGaA), 100 µM Linoleic acid (LA, Merck KGaA), or their mixture (50 µM each). Before the application, OA and LA were dissolved in DMSO and then complexed with albumin in FBS, resulting in an estimated BSA: FA ratio of ~ 2.4:1 to minimize toxicity while ensuring efficient binding and a minimal amount of unbound fatty acids. The final DMSO concentration during cell culture comprised 0.1%. After 21 days of culture, WJ-MSCs were harvested by trypsinization and centrifuged. The pellets were washed once with PBS and stored at − 80 °C for subsequent qPCR analysis. The other part of cell cultures was used for Nile red staining and TG content determination.

Nile red staining and fluorescence microscopy

MSCs were fixed in a 4% buffered formalin for 30 min at 4 °C and stained with a Nile Red (1 µg/ml in PBS) solution (Merck KGaA) according to the manufacturer’s instructions. The cells were assessed using a fluorescent microscope (Leica, Germany).

Determination of intracellular triglyceride content

The triglyceride (TG) content in AT- and WJ-MSC culture after adipogenic differentiation was determined using the Triglycerides Quantification Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Briefly, the adherent cells were washed twice with PBS and lysed using the lysis buffer provided in the kit, followed by incubation on ice for 10 min to ensure efficient cell disruption. The lysates were centrifuged at 12,000 × g for 10 min at 4 °C, and the supernatants were collected for analysis. Triglyceride levels were measured using a colorimetric assay in the Tecan Infinite 200 (Tecan, Männedorf, Switzerland) microplate reader at 540 nm absorbance. TG concentrations were calculated from a standard formula:

where

Cstand: Concentration of standard (2.26 mmol/L);

f: Dilution factor of the sample before the test (equals 1 in our experiments).

The data was presented as Mean ± SD of 4 independent donor samples assessed in duplicates.

Analysis of lipidomic profile

The lipidomic profile of cells was explored by the liquid chromatography-mass spectrometry (LC-MS) in the Service Department of Metabolomics of the Institute of Physiology of the CAS, Prague, Czech Republic. Briefly, cells were grown on 6-well plates, treated as required, quickly washed with PBS, snap-frozen, and stored at − 80 °C. Metabolites were extracted using a biphasic solvent system of cold methanol, methyl tert-butyl ether, and water³⁹.

An aliquot of the bottom (polar) phase was collected and cleaned using an acetonitrile/isopropanol mixture. After evaporation, the dry extract was resuspended in 5% methanol with 0.2% formic acid, followed by separation in an Acquity UPLC HSS T3 column (Waters, Milford, MA, USA). Another aliquot of the bottom phase was evaporated, resuspended in an acetonitrile/water mixture, and separated in an Acquity UPLC BEH Amide column. Metabolites were detected in negative and positive electrospray ion mode (Thermo Q Exactive Plus instrumentation)⁴⁰. Signal intensities were normalized to the respective total ion count (TIC) before subsequent statistical analysis.

Quantitative real-time PCR

To assess the efficacy of adipogenic differentiation, specific human marker genes were selected:

• CEBPA(CCAAT/enhancer-binding protein alpha, Hs.PT.58.4022335.g),

• PPARG(Peroxisome proliferator-activated receptor gamma, Hs.PT.58.25464465),

• FABP4(Fatty Acid Binding Protein 4, Hs.PT.58.20106818),

• LPL(Lipoprotein lipase, Hs.PT.58.45792913).

Total RNA was extracted from cell pellets using the Total RNA Purification Kit (Norgen Biotek Cor., Canada), following the manufacturer’s protocol. The concentration and purity of RNA were assessed using a NanoDrop™ 2000/2000c Spectrophotometer (Thermo Fisher Scientific). The RNase-Free DNase I Kit (Norgen Biotek Corp., Canada) was used to improve RNA purity. RNA samples with an absorbance ratio (A260/A280) between 1.9 and 2.1 were used for subsequent experiments.

For the quantitative conversion of RNA into single-stranded cDNA, we applied a High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). RNA samples were normalized to the same amount of input RNA for reverse transcription. Reverse transcription was performed using the C1000 Touch™ Thermal Cycler (Bio-Rad, USA) set to these conditions: 25 °C for 10 min, 37 °C for 120 min, 85 °C for 5 min, and 4 °C for the end.

Quantitative real-time PCR was carried out using the CFX384 Touch Real-Time PCR Detection System (Bio-Rad, USA) following these cycling conditions: initial denaturation at 95 °C for 30 s, followed by 44 cycles of denaturation at 95 °C for 5 s, and annealing at 60 °C for 30 s extension for 72 °C for 10 s.

The following primer pairs (PrimeTime^® qPCR Primers, IDT) were used in the study:

The expression of target genes was normalized to PPIA, which was selected as the housekeeping gene due to its stable expression across conditions.

Statistical analysis

Multivariate analysis of lipidomic data was performed in MetaboAnalyst 5.0²⁹. Shares of fatty acids and various TG species in non-induced WJ-MSCs and AT-MSCs were compared using an unpaired t-test. For multiple group comparisons in experiments involving induced cell cultures, ordinary one-way ANOVA followed by Sidak’s post hoc test was performed. Data analysis and visualization were done using Prism 10.2.3. Statistically significant p-values are presented in figures as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.