Riding the tailwinds of Chinese E-Commerce

China will overtake the US to become the world’s largest economy by 2028, five years earlier than previously forecast, according to the UK based CEBR (Centre for Economics and Business Research)[1]. This bodes well for the e-commerce sector which accounts for a whopping 36% of total Chinese GDP today, compared to only 15% in 2008 [2]. In this post, I will look at the particularities of Chinese e-commerce and how and why retail investors should want a slice of this pie.

Market Size

Growth Trajectory

The global e-commerce market is expected to grow at 16% compounded over the next four years (having already grown at 17% compounded over the last seven years). For context, that’s 1.5x the expected growth rate of the global smartphone market, 2.5x expected growth rate of home ownership and about the same as the forecasted growth rate for the Electric Vehicle market over the same period…so it’s a pretty big number[3],[4],[5].

Global E-Commerce Sales Growth

Shopping Experience

This is being driven by changing consumer preferences over how and where shopping takes place and improvements in mobile technology to facilitate this transition. 67% of millennials already prefer to shop online (and 56% of Gen X-ers), when compared to brick and mortar. Of these, about half make purchases whilst sitting in bed, 23% at work (and 20% from the bathroom or in the car!), representing how mobile technology has been able to make shopping more convenient and fulfill our endless desire to do things on our own terms.[6] These structural trends are not expected to go away any time soon particularly as a result of COVID-19 which has spurred increased online sales in most industries.

China led growth

Whilst the current global e-commerce market is around $4tn (roughly 5% of Global GDP), the bulk of this market share is in China, which boasts an e-commerce market almost twice the size of the US equivalent. E-commerce accounts for almost 35% of Chinese total retail sales, compared to only 10% in the US[7]. To put this into perspective, approx. 50% of global online transactions each day, happen inside of China and the largest online retailer in China, Ali Baba, has almost double the number of active users on its platform that Amazon has globally (600m vs 330m).[8]

As the first major economy to post GDP growth in 2020, the Chinese e-commerce market is facing two very strong tailwinds; a global structural trend away from brick and mortar to online, and high sensitivity to a booming Chinese economy, which is set to continue to grow 5-8% per annum over the coming decade.

Chinese e-commerce market

The Chinese e-commerce market is not identical to other markets.

  1. Mobile shopping

Mobile platforms generate 80% of retail ecommerce sales in the country (and 95% of total e-commerce activity) vs 64% globally. Put simply, this is because the average consumer in China spends more time on their smartphone (5 hours, compared with 3 hours globally)[9]. Given smartphone penetration is still far from reaching saturation (50% of the Chinese population), existing operators need to have a large-scale mobile retail presence to continue to expand[10].

2. Mobile Payments

This mobile technology has gone hand-in-hand with mobile payment platforms (such as Alipay, Wechat and Union Pay). In fact, 80% of smartphone users use mobile payments compared to only 27% in the US. The opportunities for newcomers in this space however are limited with Alipay (owned by Ali Baba) and Wechat (owned by Tencent) controlling 92% of the Chinese mobile payments market combined[11].

3. Omni-Channel Retail & Brand Awareness

The third distinction is that Chinese consumers are more discovery oriented than US consumers (i.e. they don’t always know what they want before they’re online), indicated by the former having more touchpoints with the brand before item purchase (8 touchpoints versus only 4 in the US).[12] This has made omnichannel retail (selling across multiple platforms), international brands and interconnectivity with social networks, particularly important to drive retail sales in China.

Current State of Play

All three main players in the Chinese e-commerce market (Ali Baba, JD and Pinduoduo), cover some form of mobile retail, mobile payments and omni-channel retail but Ali Baba is miles ahead. This is because, whereas, JD and Pinduoduo both have secured strategic investments from Tencent to add additional services to their primary retail presence (e.g. mobile payments and social networking) Ali Baba has been able to integrate these functions across its existing businesses (such as Ali Pay and Lazada).

Resultingly, Ali Baba has the biggest slice of the pie with 60% market share, led by its Business-to-Consumer marketplace, Tmall. To put that into perspective, Amazon accounts for 40% of online retail sales in the US and is widely considered an e-commerce behemoth in the US[14].  

List of Chinese e-commerce players

Tmall China (Ali Baba) – The third most visited site in the world and a brand driven B2C marketplace.

60% Market Share

JD.com (JD) – marketplace with in-house delivery and logistics.

20% Market Share

Pinduoduo – Group buying marketplace (like Groupon).

5% Market Share.

Others – (Wechat Store and Red/Xiahongshu)

~15% Market Share

Ali Baba as an E-Commerce Stock

Whilst Ali Baba is by no means a pure play Chinese e-commerce stock, 68% of FY 2019 revenues came from Ali Baba’s China based marketplaces (compared to 7% from international wholesale, 7% cloud computing, 6% entertainment and  4% logistics services). 85% of this in was via mobile sales. This illustrates that whilst Ali Baba touches many different industries, it is still very much a Chinese consumer focused, mobile e-commerce giant.

Ali baba strikes me as the most defendable business model within the Chinese e-commerce space. This is principally because it meets the threshold of having the strongest mobile retail penetration for the Chinese consumer (with higher gross margins), mobile payments integration and cross border retail presence to drive brand awareness.

Why Ali Baba is superior to competition

Overview of Ali Baba E-Commerce Businesses

Strong Mobile Customer Penetration (and a more profitable business model)

Based on Chinese internet users of over 855 million, Ali Baba boasts an 83% penetration rate for its B2C and C2C platforms (Tmall and Taobao), which are the two largest online marketplaces in China. About half of the Chinese population are users of Ali Baba and over 85% of total yearly sales on Ali Baba’s platform, come from mobile (with total GMV larger than Amazon and Ebay’s yearly sales, combined). This deep penetration has allowed Ali Baba to leverage network effects across its platforms (the platforms drive traffic to each other thereby lowering acquisition costs) and as a result, Ali Baba is usually the first shopping experience for Chinese customers. This deep penetration has allowed the company to maximize revenue per user relative to its competitors (GMV per active user was CNY 730 versus CNY 143 for Pinduoduo and CNY 362 for JD).

Direct Customer fulfillment

Ali Baba is often described as the Amazon of China whereas in fact, it operates more as an ‘Ebay’ of China. Tmall, it’s B2C platform, gives businesses direct customer fulfillment so they manage the warehousing and shipping of goods to the end customer. This allows Ali Baba to charge servicing fees for marketing on its platform but it does not have to take inventory on its own balance sheet. This vastly improves the overall net income margin for Ali Baba which is 22% (vs 5% for Amazon, 1% for JD) and Gross Margin OF 44% (vs 40% for Amazon, 15% for JD).

Mobile Payments Technology

Ali Baba also owns a 33% stake in Ant Financial, a mobile payments and microlending platform. Ant has the largest share of the Chinese mobile payments market (54%) in an industry which is set to double in size by 2025. Wechat is second in size at only 38% market share. This dominant position has allowed Ali Baba a virtual monopoly on mobile payments as it can integrate this with it’s online e-commerce platforms.

Omni-Channel Retail via cross border e-commerce

Outside of China, Ali Baba is the furthest ahead in integrating its business with global brands and cross border e-commerce. This is primarily facilitated by Ali Baba’s controlling stake in Lazada, a South East Asia focused e-commerce business (with 200million active internet users across these regions). In these regions, e-commerce accounts for a meagre 3-5% of total sales reflecting lack of existing providers and strong runway of growth ahead. In 2009, Ali Baba also purchased Koala from rival Netease (integrated into Tmall) to build out its import strategy; allowing global merchants direct access to the Chinese consumer. This will allow Ali Baba to accelerate import/export services, have (exclusive) access to international brands selling into China, and drive brand awareness.


The biggest risks to backing this tech behemoth is regulatory uncertainty in China. The two specific risks to the price of Ali Baba stock (though these developments still have a corollary impact on the broader sector) are:

  • Ant Financial

Ali Baba has a 33% stake in Ant Financial, whose long awaited IPO came to a grounding halt as Chinese regulators suspended the offering and proposed last minute changes to Ant’s business model. Whilst founded as an online payments platform, Ant has grown beyond this and this has become the sticking point for the regulators. 63% of Ant’s revenues come from matching consumers to lenders via its platform (this was 44% of revenues in 2017 speaking to the spectacular growth this division has experienced). Ant uses its own proprietary technology to provide customers with credit scores (Zhima credit scoring system), compiles lending risks and then securitize debt via third parties or to pass this on to conventional banks.

Chinese regulators want to treat these type of platforms as financial institutions and not tech companies, and resultingly want Ant to put up its own capital against these loans (30% of capital for joint loans with banks versus current 2%). This resultingly limits Ant’s leverage and growth rate, particularly if Ant also has to apply for new banking licenses.

Whilst Ant is not Ali Baba’s core business, the elephant in the room here is the Microlending/P2P scandal in 2008. During this period, China saw a rise in scandals and mismanaged debt which the government is keen to avoid the second time around[15].

My take on this is that unlike in 2008, the tech giants are providing a service which conventional banks have been unable to provide. Credit cards are not pervasive in China and financial records are largely inadequate in rural areas. This has limited access to credit in remote areas and to date, the biggest four conventional banks give out 75% of their total loans only in the form of mortgages (in 1H 2020). This has left a gaping hole for new microlending entrants who have the digital footprint and know-how to provide this service (and pick good borrowers from bad ones). Whilst Ant may be hamstrung by new regulations, this will likely come at the expense of short-term growth as this service cannot feasibly be performed by existing conventional banks or other private businesses without the same scale and expertise. In other words, the overall business opportunity remains intact but Ant will have to wait longer to achieve the same outcomes or at least, until regulators are more comfortable.

  • Anti-monopoly Laws

The second major risk to Ali Baba is an overhaul of the existing competitive framework in China for e-commerce. As we established above, Ali Baba relies on its massive network to be as effective within the e-commerce space. However, in 2020, the Chinese government first applied a monopoly law to fine Ali Baba (alongside others).

The newly contested issue is platform integration or exclusivity arrangements that Ali Baba may have for that prevent users from going on other platforms. This is a big blow to the company and so far, it is unclear how this will play out, particularly as there are so many avenues that they could pursue this with Ali Baba. For me, this is the single biggest risk to the stock as Ali Baba relies heavily on its expansive network to be effective.


I am bullish China and Chinese e-commerce. Ali Baba is the natural choice for investment, as it’s the biggest, most exposed and the most defendable business operating in the sector. Ironically, the biggest risk to this position is that regulators see these strengths and force Ali Baba to concede market share to allow new entrants in the market. Despite this, Ali Baba is sizably ahead of the competition and has cemented a strong position with the Chinese consumer. This should make policies short of actively breaking up Ali Baba, to be nothing but a slow burn; as what pieces of business Ali Baba may eventually lose to competitors will likely be more than offset by its levered position (and natural monopolies) in this fast growing industry.

Verdict: Cautiously Bullish (pending impact of regulatory investigation)
Timeframe: 3-5 Years

[1] https://www.bbc.com/news/world-asia-china-55454146

[2] https://tenbagroup.com/12-china-e-commerce-market-trends-2021/

[3] https://www.mordorintelligence.com/industry-reports/smartphones-market

[4] https://www.globenewswire.com/news-release/2020/04/14/2015310/0/en/The-Global-Electric-Vehicle-Market-is-expected-to-grow-from-USD-146-902-20-Million-in-2019-to-USD-359-854-56-Million-by-the-end-of-2025-at-a-Compound-Annual-Growth-Rate-CAGR-of-16-.html#:~:text=%2F%3Futm_source%3DGNW-,The%20Global%20Electric%20Vehicle%20Market%20is%20expected%20to%20grow%20from,(CAGR)%20of%2016.10%25.&text=The%20Battery%20Electric%20Vehicle%20commanded,Electric%20Vehicle%20Market%20in%202019

[5] https://www.mordorintelligence.com/industry-reports/us-homeowners-insurance-market

[6] https://kinsta.com/blog/ecommerce-statistics/

[7] https://kinsta.com/blog/ecommerce-statistics/

[8] https://business.cornell.edu/hub/2020/02/18/impact-e-commerce-china-united-states/

[9] https://tenbagroup.com/12-china-e-commerce-market-trends-2021/

[10] https://www.statista.com/statistics/321482/smartphone-user-penetration-in-china/

[11] https://www.paymentscardsandmobile.com/alipay-continues-mobile-payments-expansion-in-japan/

[12] https://digovation.com/2020/06/15/understanding-the-e-commerce-landscape-in-china/

[13] https://www.sellmasters.it/en/2020/01/28/e-commerce-usa-vs-china/

[14] https://www.emarketer.com/content/china-ecommerce-2020

[15] https://www.scmp.com/tech/innovation/article/3113410/how-chinas-online-microlenders-reached-tipping-point-worlds-biggest

3D Printing: Turning Science-Fiction into reality?

When I initially stumbled across 3D Printing, I dismissed it as a futuristic and unproven technology, propogated by investors who were trying to offload their positions onto the next naïve Robinhood trader. However, when I purchased my new car a few weeks ago, I was surprised to find out that this technology was actually used in the creation of some of the parts that I now so ferociously admire. So I figured it was about time I gave 3D Printing another look… and I’m glad I did.

How does it work?

Additive Manufacturing (‘AM’), also known as ‘3D Printing’, strikes me as a sort of new industrial revolution taking place in modern, developed economies. There are a lot of tailwinds to this technology in Europe and the USA, as governments (on both sides) are battling to shore up more jobs and prop up structural declines in manufacturing, as these industries have generally moved to places where labor is cheap and regulation is lax, like China.

Put simply, the way AM works is by building objects layer by layer. Imagine a printer but one that makes real objects instead of paper documents. Now imagine connecting this to a computer program so you can design a 3D model, send this data to a machine and get your desired physical object out the other end. That’s how AM effectively works today.

This is unlike traditional or ‘subtractive’ manufacturing, where parts are made typically by casting or machining. In casting, a mold is used to create a specific part (and destroyed when that part is no longer required) whereas in machining, a billet is used to create the desired object. In both instances, the process takes more time, preparation and is inherently built for for mass production.

The 7 different types of AM

AM is a catch-all term for 7 unique technologies. Below is a list of these seven along with my simple explanation and my crude ranking for how they stack up.

Main advantages of Additive Manufacturing

Across these technologies, the key advantages of AM are the following:

  • Customization – Allows unique designs
  • Rapid prototyping – Saves time procuring machines or casts prior to production
  • Advanced geometries – Build intricate designs (particularly lightweight structure)
  • AI integration – AM can be programmed to reverse the process of manufacturing from producing a specific structure for a need, to finding what structures fits your need the best

To summarize, AM is increasing sophistication in an industry which has been largely stagnant for decades.

Market size and opportunity

The AM market is split into three key segments, Printing, Parts and Materials. The first two are around 40% each in terms of size of the overall market and that’s where the biggest opportunity lies.

The AM industry has gone through its early stages which is often referred to as Additive 1.0, starting in 2006. Additive 1.0 has seen the market expand from $1bn to over $12bn (20% CAGR). The key feature of Additive 1.0 has been rapid prototyping i.e. taking new complex designs and testing them quickly using AM technologies.

The new phase for this market is referred to as Additive 2.0. During this period, AM needs to prove it can go beyond rapid prototyping and start putting these prototypes into mass production. For me, this is the biggest test for the industry to date as this is something where AM doesn’t generally have a natural advantage. If it can pull it off, the industry is expected to grow from $12bn to $146bn over the next decade (CAGR of 25%).

How does the transition to Additive manufacturing look?

  • Bridge Manufacturing

Most factories have minimum order commitments (or producers inherently have to pay this to recoup investment in parts required in casting/machining processes). This poses a high cost of entry for new entrants and as a result, is where AM can make the biggest impact.

The below snapshot is taken from Desktop Metals latest investor presentation (2020). Assuming other Additive 2.0 print manufacturers are as efficient, what this shows is that for production <100,000 units, AM remains the most cost competitive technology. The market opportunity in bridge manufacturing should continue to expand as AM continues to see enhanced efficiencies (e.g. through material waste recycling, faster processing speeds and economies of scale in production).

  • Enhanced production

The second key transition that we can see taking shape is what I refer to as ‘Enhanced Production’. There is an inherent disincentive in the traditional manufacturing process to produce complex geometries or shapes in favor of block type manufacturing. This is because it is harder and more costly to impose complex geometries in casting or machining. Nonetheless, the opposite is true for AM; where block printing is actually more expensive (as it uses more material and takes longer to build). In industries where there is a need for lightweight, highly durable and specialized structures, AM has a natural advantage, even at mass production.

Investment Opportunity: Desktop Metal

Desktop metal (DM) sits firmly within the AM space and is currently the only pure play company set to benefit from Additive 2.0. DM’s portfolio is making this transition from prototyping to mid-level or mass production with a particular focus on metal Binder Jetting technology (which I ranked as my favorite technology above).

Currently, they sell three printers (Fiber Q4 2020, Studio System Q4 2018, Shop system Q4 2020) and expect to sell a high-speed mass production system in the 2H 2021 (P1). The focus for DM is on speed of manufacturing and they currently have the fastest metal 3D printing technology (roughly 100x faster than other technologies).

Company Financials

Whilst DM sits in a space which is growing and is likely to experience significant tailwinds from broader industry growth, the thing I like most about this company is the recurring revenue stream.

In their latest P1 Production System for example, cumulative gross margins grow from circa 28% on sale of the initial printer to approx. 58% by Year 10, through the sale of consumables, services and materials. This is truly remarkable and should cement strong Cash Flows and profits as DM continues to expand its order book for 3D Printers. Given the highly specialized nature of AM technology, I would also expect customer price elasticity after the initial sale to be very low (i.e. allowing for greater mark up on subsequent sales). As of now, DM has 90+ orders for it’s Production System through 2024.

Operating Leverage

DM completes its manufacturing through contractors. This has allowed the business to remain asset light and have a high degree of Operating Leverage as it continues to scale its customer base. With strategic investments from Ford and BMW, it has unparalleled access to the automotive industry (Ford and BMW are already using DM’s printing technology). This should lead to continued FCF and EBITDA growth over the next years.


The AM industry has proven itself to be successful at a localized level having grown at over 20% the last decade and it is set up to continue to grow at a faster speed, as it transitions into mass production. This is a paradigm shift for this industry, going from initially complementing traditional manufacturing to now trying to actively replace it. The backdrop of protectionist agendas in developed economies and more sophisticated, personalized products should act as tailwinds to this sector. The biggest risk here is that it takes longer to play out than planned but for the patient investor, this is great risk/reward opportunity.

Verdict: Bullish
Timeframe: 5-10 Years

I’m long Gold. Now what…

A week ago, I got long Gold and I’ve been frustrated with my position since. Not because the investment has gone against me (I bought calls on a Gold Miner, Kirkland Lake Gold, which are up 30%), but because I’m still not clear on what I’m actually investing in. To make things more confusing, everyone seems to have their own reason for buying Gold, so I don’t even know what to root for. In this post, I wanted to run some (simple) historical analysis to try to understand what is really driving Gold and whether I still believe in this position.


EV Charging Stations: The smart(er) entry point into the EV market

As of right now, Electric Vehicle (EV) sales account for 3.7% of total vehicles sold globally[1]. That number is forecasted to grow massively over the coming years; growing to 25% in 2025, 27% in 2030 and 35% in 2040[2]. The average range for EV’s on our roads however, is expected to remain within 200miles on a full charge, leaving a huge requirement for EV charging infrastructure to support this fleet. In fact, it’s estimated that the US alone will need 1-2million public charging stations alone. Whilst I do not know who will ultimately win the race to sell EV’s, there is undoubtedly huge demand for charging services to come, and that’s what I want to look into today.


Rare Earth Metals – Invest in the future

This week, I want to look at Rare Earth Metals (REM). It’s a nice segway from my previous post on Electric Vehicles, as these metals are a crucial building block of not just electrification technology, but most technologies we rely on today (Consumer Electronics, Medical Research, Defense…).


Fuel Cell Electric Vehicles – believe the hype?

As someone who missed out on the meteoric rise of the Battery Electric Vehicle (BEV) market, the question I have been asking myself more recently is what will emerge as a viable alternative in the years to come. One such alternative I came across (on an episode of Jim Cramer’s, Mad Money) are Fuel Cell Electric Vehicles (FCEV’s).

Current state of play: Electric Vehicles

Despite having less than 2% global market share (as a proportion of total cars sold), BEV stocks have soared well above conventional car manufacturers in both market cap and YTD stock performance, even when some of those same conventional manufacturers compete in both markets.

Market Capitalisation of Electric Vehicle Stocks vs Conventional Vehicle Stocks
Electric Vehicles Market Share (2019)

How do FCEV’s work?

Put simply, FCEV’s have three key components – (1) an anode (2) a cathode (3) electrolyte membrane. All three rely on one crucial ingredient – hydrogen.

Skipping the technicalities – in goes hydrogen, and with the help of a catalyst, it generates electrical current, water (and some heat).

The beauty of this process is that you don’t need combustion like a conventional vehicle, it operates silently, has no tailpipe emissions and allows you to store energy in the form of (normally liquified) hydrogen in your gas tank.

This is broadly similar to a BEV but replaces the need for lithium batteries as power is generated directly in the gas tanks of the vehicle.

The benefits

FCEV’s are pitched against BEV as a competitor to create zero emissions (in the driving process). In many ways, FCEV technology helps to overcome some of the major problems associated with BEV’s. Namely, they take only a few minutes to fill up compared to closer to an hour for a BEV, have a longer range (over 480km in many cases) and they reduce reliance on heavy (limited supply) lithium batteries, making vehicles lighter and more fuel efficient.

Chicken & Egg problem

Yet despite those efficiencies, the biggest challenge to FCEV’s taking off is the classic chicken and egg problem. It goes something like this – without a viable network of hydrogen fueling stations, FCEV’s struggle to gain in popularity as they have limited appeal to someone who can’t reliably use them. However, without sufficient customers, infrastructure spend on fueling stations remains largely limited.

Whilst BEV vehicles had a similar issue to start, FCEV’s are uniquely disadvantaged. A typical hydrogen fueling station can’t be plugged into the power grid like a charging station can. Any provider will need to insure they have a hydrogen distribution network to safely store and handle liquified hydrogen, a very flammable substance. This makes hydrogen fueling stations costly to set up and increases the cost of entry for new participants.

The average station costs $1.1m vs $600k for a charging station (with 4x 150KWH charging points). No surprises then how sparce hydrogen fueling stations are in the US and Europe. As of 2019, there were only 177 hydrogen stations in all of Europe combined, and 39 stations in the US (35 of which are located in California). The infrastructure is vastly underbuilt.[1]

Hydrogen Fueling Stations – USA

The big picture

You may argue however, that if FCEV’s help us fight the challenge of climate change and CO2 emissions – maybe it’s risk worth taking. Here, yet again, FCEV’s come up short. Hydrogen is typically created by electrolysis – a process of separating hydrogen from oxygen, in water. Whilst we have an abundant supply of (sea) water, electrolysis is an exceptionally wasteful method of creating energy. In fact, the overall efficiency rate of electrolysis in producing energy to powering a car is 30-50% of that compared to a BEV[2]. In other words, for the same $ amount of electricity spent to generate electricity, you could obtain more than twice the amount of energy to power a BEV than you would to power an FCEV.

This is particularly problematic when considering the cost of entry into the market. The average FCEV costs almost 1.5x the price of a BEV before energy costs are considered. That delta only widens when you include the higher cost of fueling your FCEV.

Localised demand

So FCEV’s cost more, struggle from lack of infrastructure and are less efficient (from a cost of energy perspective). So they have and will likely continue to struggle to expand passenger vehicle sales. However, one area where FCEV’s can and have done well, are in localized demand centres such as airports, city busses and utility vehicles.

Companies such as Hyundai and Plug Power have done exceptionally well here as they can cut the inefficiencies associated with heavy batteries and support the investment required to build hydrogen stations, where they can tap into reliable, repeat customers such as city busses or factory utility vehicles. These locations however, often go hand-in-hand with cheap access to renewable power sources to obtain the hydrogen such as Hyundai’s partnership in Switzerland[3].

Final say

One thing that made BEV’s great is having a transformational company such as Tesla pioneer the change. You need a sexy, elusive brand to help drive the change in customer behavior, which then justifies the return in building out charging infrastructure. FCEV’s have an even larger struggle – they require more expensive infrastructure, cannot claim to be more efficient and as for pioneering companies… Honda, Hyundai, Toyota don’t personally strike me as being the right brand to drive the transition to FCEV’s. Whilst industrial vehicles and long range trucks may be the ideal candidate for fuel cell technology, I’d want to see better vertical integration in those companies particularly around renewable power generation, before I part with my money.

Verdict: Cautiously Bearish
Timeframe: 1-3 Years

[1] https://www.automotiveworld.com/articles/why-battery-evs-have-raced-ahead-of-hydrogen-fuel-cell-vehicles/

[2] https://www.bmw.com/en/innovation/how-hydrogen-fuel-cell-cars-work.html

[3] https://tech.hyundaimotorgroup.com/article/the-future-lies-in-ev-or-fcev/