State of AI in financial services; Ethereum’s coming ‘Merge’ could make or break crypto; The evolution of financial services building blocks;

Sam Boboev
19 min readAug 17, 2022


In this edition:

1️⃣ Coinbase racks up $1.1bn loss as crypto trading volumes slum

2️⃣ Top 100 Public Companies Investing in Blockchain & Crypto Companies

3️⃣ State of AI in Financial Services

4️⃣ Use of specialized software by business size

5️⃣ DEX performed better than CEX

6️⃣ ‘Dangerous’ precedent: crypto execs slam arrest of suspected Tornado Cash developer

7️⃣ The evolution of financial services building blocks

8️⃣ In defence of stablecoins

9️⃣ Ethereum’s Coming ‘Merge’ Could Make or Break Crypto

🔟 Binance receives in-principle approval to operate in Kazakhstan

Coinbase racks up $1.1bn loss as crypto trading volumes slump

Cryptocurrency exchange Coinbase said its revenue had plummeted 61 per cent in the latest quarter as crypto prices fell and trading volumes slumped.

The San Francisco-founded company on Tuesday reported an after-tax loss of $1.1bn, compared with the $1.6bn net profit it registered in the middle of the crypto boom a year ago. It said $446mn of the loss reflected an impairment charge on its crypto and venture investments.

Coinbase shares fell more than 5 per cent in after-market trading, having lost nearly 11 per cent earlier in the day.

The collapse in bitcoin and other crypto prices sent the company into the red in the first quarter, leading it to lay off 18 per cent of its staff in June. With trading volume down another 30 per cent compared with a weak first quarter, it suffered a $647mn operating loss before the impairment charge and its net cash position fell by about $400mn, to $2.8bn.

In an attempt to ease Wall Street concerns about its financial condition in the middle of what it called a crypto “winter”, Coinbase published extra information about its cash burn and said it had not experienced any credit losses, despite the financial stresses hitting some other crypto companies.

Chief financial officer Alesia Haas said that, along with long-term debt taken out when interest rates were lower, Coinbase believed its $6.2bn in available capital would enable it to keep investing through the downturn.

She added that Coinbase believed it could maintain its plan of keeping this year’s loss on earnings before interest, taxes, depreciation and amortisation to no more than $500mn.

However, the company also faced questions from analysts about its mounting non-cash losses, with its net loss of $1.52bn in the first six months equivalent to 77 per cent of its revenue.

Stock-based awards to employees equalled 49 per cent of revenue in the latest quarter, and Coinbase projected they would continue at the same level in the current quarter before dropping off next year.

Coinbase reported net revenue of $803mn in the quarter and a loss per share of $4.98. Wall Street had been expecting a loss of $2.65 a share on revenue of $832mn.

The shares have been volatile in recent days, jumping more than 30 per cent last week on an alliance with BlackRock. They had already given up two-thirds of the gain ahead of Tuesday’s earnings announcement.

Source Financial Times

Top 100 Public Companies Investing in Blockchain & Crypto Companies

BLOCKDATA dig into the blockchain investments top 100 public corporations made from September 2021 to mid-June 2022.

Forty corporations invested in companies in the blockchain/crypto space during this time. Samsung is the most active, having invested in 13 companies. UOB came in next with 7 investments, followed by Citigroup with 6 investments, and Goldman Sachs with 5.

In most cases, we cannot determine how much money these corporations have invested, as they participate in funding rounds with multiple or many other investors.

As a proxy of this, we can look at the total funding amounts of the rounds they participated in.

Based on this, the investors active in the biggest funding rounds are Alphabet ($1,506M in 4 rounds), BlackRock ($1,171M in 3 rounds), Morgan Stanley ($1,10M in 2 rounds), Samsung Electronics ($979M in 13 rounds), Goldman Sachs ( $698M in 5 rounds, BNY Mellon ($690M in 3 rounds), and PayPal ($650M in 4 rounds).

The 40 companies invested approximately $6B into blockchain startups between September 2021 and June 2022. Because some rounds involve participation from multiple investors, it is unclear how much each company invested in a project.

What use cases are the top corporations investing in?

A total of 61 blockchain/crypto companies received investments across 71 investment rounds. These blockchain companies are active across more than 20 industries and 65 use cases:

Nineteen companies offer some form of non-fungible tokens (NFT) solutions and services. Many of these belong to industries such as gaming, arts & entertainment, and distributed ledger technology (DLT).

In all, 12 firms are marketplaces, with some supporting the buying and selling of NFTs.

Eleven organizations provide gaming services. There is considerable overlap among use cases for the companies that offer NFT solutions, marketplaces, and gaming.

The popularity of NFTs can be mainly seen as an opportunistic move by corporations looking to capitalize on trends to meet where their customers are transacting. The startups raising capital are enabling commerce in decentralized worlds by developing platforms where users can buy and sell NFTs, including virtual land, clothing, and other branded items.

Different investment approaches by public companies

Samsung is placing bets across the blockchain ecosystem investing in companies focused on 15 different use cases, such as blockchain services, development platforms, NFT, and social networks. On the other hand, Alphabet and BlackRock are showcasing a completely different strategy by making concentrated bets on a smaller set of companies.

That being said, corporations are exploring specific use cases and portfolios that complement their core offerings. For instance, HERE Technologies invested in UNL, a blockchain-based location and mapping technology company, to implement blockchain-enabled security and transparency measures in a location/mapping setting.


State of AI in Financial Services

- Deep Learning Dominates in Capital Markets and Retail Banking, While Fintech Relies on Machine Learning

Across all sectors of financial services — capital markets, investment banking, retail banking, and fintech — over 75 percent of companies utilize at least one of the core accelerated computing use cases of high-performance computing (HPC), machine learning, and deep learning.

Capital market firms — hedge funds, asset managers, and exchanges — that need every possible edge to improve financial returns are the most prevalent users of deep learning at 58 percent. In contrast, 80 percent of fintechs — which have the enterprise AI capabilities available from the cloud but may lack the scale of data needed to enable many deep learning use cases — are leveraging machine learning.

- Impact of AI in Financial Services

Ninety-one percent of financial services companies are driving critical business outcomes with investments in AI. First and foremost, 43 percent of respondents stated that AI is yielding more accurate models. Along with model accuracy comes a host of other benefits.

The data reinforces the state of play for AI in financial services today: Companies not utilizing AI are in the minority and consequently more likely to deliver inferior customer experiences and less efficient operations, leading to reduced revenues and market share

- AI Use Cases

Respondents to this year’s survey reinforced the findings from last year, stating that AI is enabling a wide array of meaningful use cases for financial services companies. The top two of the top three priorities across the industry remain fraud detection and algorithmic trading, while conversational AI is a new entrant into the top three. More importantly, the percentage of companies investing in each use case jumped significantly year over year (YoY), with underwriting and acquisition, conversational AI, and anti-money-laundering (AML) and know-your-customer (KYC) fraud detection showing the largest percentage gains. Interestingly, nine of 13 use cases tracked in our study are utilized by at least 15 percent of respondents’ companies, whereas none of the use cases had more than 14 percent industry penetration in last year’s survey. This demonstrates the rapid adoption of AI across financial services, which is requiring banks to invest in enterprise #ai strategies and infrastructure.

When asked which use cases companies would continue investing in for the next six to 12 months, the top five varied by sector. Notably, fraud detection for payments and identity verification (#aml and #kyc ) were the only two use cases to make the top five in three or more sectors, revealing the diversity of applications that are important to different types of financial services firms.


Game on: how tech companies are betting on the metaverse

The $180bn gaming industry, twice the size of the film business, already attracts hundreds of millions of players. But as tech leaders vie to create the next iteration of the internet, gaming has become a battleground.

The Financial Times explores how gaming got so big, and whether it really can be the gateway into this new world. Or will regulation, integration, and plain old reality get in the way?

DEX performed better than CEX

Over the past five years, decentralized exchanges (DEXs) have emerged as a self-custodial, programmatic way for cryptocurrency investors to trade. DEXs allow users to swap between hundreds of trading pairs without an intermediary. And fifteen months ago, these DEXs for the first time eclipsed centralized exchanges (CEXs) in on-chain transaction volume.

While most CEX transactions happen off-chain on centralized databases and captured on their order books to save on transaction fees, every DEX transaction occurs via smart contracts on-chain. For this reason, as well as the rapid growth of DeFi generally, DEXs now have a confident lead in on-chain transaction volume: from April 2021 to April 2022, $175 billion was sent on-chain to CEXs, well below the $224 billion sent to DEXs.

The transaction volumes at centralized and decentralized exchanges are closely correlated with market performance. For example, CEX transaction volume reached an all time high in late 2017 as Bitcoin climbed to its all-time high. Similarly, DEX and CEX transaction volumes alike skyrocketed in 2021 as cryptocurrency prices again multiplied. But with the recent market slump, the amount sent to both exchange types declined, with CEXs proving slightly more resilient than DEXs in current market conditions.

The balance first shifted away from centralized to decentralized exchanges in September 2020, when centralized exchanges supported below 50% of on-chain volume for the first time. DEX dominance then reached its peak in June of 2021; that month, DEXs facilitated more than 80% of on-chain transaction volume. Today, their share of on-chain volume is more evenly split, with 55% happening on DEXs and 45% on CEXs.

Another possible explanation is economies of scale, an important mechanism for DEXs. DEXs with higher liquidity may be able to provide more stabilized pricing for even the biggest market participants, but smaller pools may struggle to do the same without causing considerable price slippage — an unappealing proposition for both consumers and liquidity providers. A third explanation is simply that competition is intense.

By contrast, the top five centralized exchange services supported roughly 50% of all on-chain CEX transaction volume during the time period studied. However, it is worth noting again that on-chain CEX volume represents only the flows into and out of CEXs, not the trading volume of their off-chain order books.

Centralized exchanges’ lower concentration may be due to greater competition among CEXs, greater focus on regulatory hurdles within and across jurisdictions, and/or greater variability in how much these services’ users also use personal wallets.

Source Chainalysis Inc.

‘Dangerous’ precedent: crypto execs slam arrest of suspected Tornado Cash developer

Crypto industry executives reacted with alarm at the arrest of a developer suspected of involvement in Tornado Cash.

The Dutch Fiscal Information and Investigation Service (FIOD) arrested an unidentified 29-year-old man in Amsterdam, it said on Friday, alleging that he was involved in concealing criminal financial flows and facilitating money laundering through Tornado Cash, a crypto mixing service that allows users to obscure blockchain-based transactions.

Ryan Sean Adams, founder of Mythos Capital Management and Bankless, was among the first to condemn the arrest. He tweeted about two hours after it was made public that the suspect may have written “code that served as a public good for people to maintain their privacy online.”

“They put a man in jail because bad people used his open source code,” Adams wrote. “This cannot stand in any free society.”

The arrest was also seen as an assault on privacy by Cinneamhain Ventures partner Adam Cochran. “Code is free speech. Unless there is more to this story, then arresting someone for making a privacy tool that was misused is an insane government overreach,” Cochran tweeted.

Robin Andre Nordnes, an analyst at Outlier Ventures, tweeted via his handle Outlier Ventures that the arrest was a “dangerous” precedent and a pivotal moment in the crypto industry that might have “huge consequences.”

Aave founder Stani Kulechov echoed Nordnes, tweeting that an arrest for writing privacy-preserving code was out of line: “This arrest makes all privacy/encryption developers a target,” he said, adding that “people use privacy tools on a daily basis online.”

Yearn core developer Banteg likened the arrest to detaining the founders of a firearms manufacturing company for facilitating a public shooting, or a pressure-cooker company founder for terrorism.

The arrest took place on Wednesday, just two days after the US Treasury added Tornado Cash and 44 associated Ethereum and USDC wallets to its Specially Designated Nationals list. The Treasury accused the crypto mixer of laundering more than $455 million for North Korea’s Lazarus Group amid Tornado Cash’s association with high-profile hacks including Ronin and Harmony. The sanction was believed to be the first by a US regulator against a DeFi operator.

Ryan Selkis, co-founder and CEO of crypto research firm Messari, slammed the Treasury decision. He wrote on Twitter that the US government’s crypto policies “reward scammers and bad actors,” while punishing innovators.

Source The Block

The evolution of financial services building blocks

What if you sliced these “as a service” layers even further, right down to the most basic primitives? In fintech, a primitive — meaning not developed or derived from anything else — could mean a basic ledger, or a reference library for sending a payment of a particular type. These primitives would be open source: they would benefit from the continuous collaboration of the smartest minds around the world. They would be composable, meaning they could be selected and assembled in any combination to satisfy any specific user requirements.

The broad availability of open source primitives would motivate novel use cases we haven’t yet imagined. For instance, we will be able to combine the crypto and fiat worlds, friction-free, so users could spend, save, or lend fiat and crypto dynamically. Alternatively, there has long been talk about how software should be able to help us make better financial decisions — consider the prospect of a banking account that could make spending and saving decisions for us and fully automate our financial lives. Composable, open source primitives would unleash thousands of such experiments; it’s hard to predict which might be the next billion-dollar fintech company.

Source Andreessen Horowitz

In defence of stablecoins

Interesting opinion by Miles Jennings general counsel and head of decentralisation at Andreessen Horowitz crypto.

Crypto critics are using the collapse of dollar-pegged virtual currency TerraUSD as ammunition to attack stablecoins and the crypto industry as a whole.

Lost in the conversation is, however, the root cause of the turmoil. A better understanding of what went wrong — and why — could help protect consumers while safeguarding innovation.

It is a crucial point that policymakers across the world need to consider as they draft legislation to prevent future Terra-like collapses. If legislators believe algorithms are to blame, they risk enacting counter-productive, innovation-stifling regulations. Poorly designed laws could disrupt markets, encourage regulatory arbitrage, and diminish Western democracies’ influence in the rising, decentralised internet economy known as web3.

Amid the recent market volatility, the vast majority of “decentralised” stablecoins backed by blockchain assets such as bitcoin and ether performed superbly, handling extreme price fluctuations and unprecedented redemptions without fail. Generally speaking, algorithms are not the issue with modern stablecoins. Instead, essentially all risk now arises from their collateral design.

The riskiest stablecoins are readily apparent: they are significantly under-collateralised (less than $1 of collateral is required to mint $1 of stablecoin), and they rely on “endogenous” collateral (collateral created by the issuer such as governance tokens that give holders voting powers on a blockchain’s rules and procedures).

Endogenous collateral enables dangerous, explosive growth: when an issuer’s governance token appreciates, users can mint many more stablecoins. That sounds fine until one considers the flipside: When the price declines — as is practically guaranteed during a bank run — cascading collateral liquidations to meet redemptions trigger a death spiral. See TerraUSD as an example.

Stablecoins can, indeed, be stable if they manage their collateral properly. For “centralised” stablecoins backed by real-world assets, the liquidity and transparency of reserves may be low, so collateral should include less volatile assets like cash, treasuries and bonds. Regulators can establish parameters regarding these types of collateral and require regular audits.

For “decentralised” stablecoins, the almost exclusive use of blockchain assets such as bitcoin or ether as collateral has trade-offs. Digital assets, while often volatile, are also highly liquid and can be transparently and algorithmically managed. Redemptions can happen nearly instantaneously, enabling much more efficient systems. As a result, decentralised stablecoins could, ultimately, be more resilient than centralised ones.

Source Financial Times

Ethereum’s Coming ‘Merge’ Could Make or Break Crypto

ETHEREUM’S $415.3 BILLION market capitalization depends on the Merge going smoothly, but so do the thousands of businesses that operate on the blockchain, plus millions of users. Some $121.5 billion of capital is locked in Ethereum’s decentralized finance (#defi) apps, according to tracker DappRadar. Most NFTs — also with a total value in the billions — use Ethereum.

The Merge will be a nail-biter because a lot could go wrong. There could be software bugs or hacks, or miners could create an alternative Ethereum network. During a 2020 network upgrade, a bug split Ethereum in two, wreaking havoc on its nascent DeFi ecosystem, the apps that let people trade, borrow, and lend without intermediaries like banks.

Most centralized crypto exchanges are expected to pause Ether withdrawals and deposits around the Merge as a precaution. DeFi apps may pause, too, if something goes wrong.

THE MINERS ARE causing the most worries. Many may quit the network right before the Merge, figuring they can make more money by selling their gear than by waiting to get the last of the rewards. Too steep a drop in the network’s mining power, or the “hash rate,” could weaken Ethereum’s security, spelling disaster for its token and the various apps using the network. Ethereum’s core developers have planned for that scenario.

Miners may also choose to fork #ethereum , by taking the existing proof-of-work software and continuing to support it. That would create two different versions of Ethereum that run in parallel: proof of work and proof of stake.

SHUTTING ETHEREUM’S old chain will send shock waves through the cryptomining industry. Scrambling to find other uses for their equipment, miners will migrate their machines to other, similar chains such as Dogecoin, Litecoin, and Monero. The hash rate on those other chains will balloon by 5 to 10 times overnight, says Sam Doctor SCO at BitOoda, a digital asset fintech company. Overall revenue for this type of mining could drop as much as 90%, pushing many miners out of business, he says.

INVESTORS MAY BENEFIT from the Merge. The number of new coins issued on Ethereum as rewards for ordering transactions should decrease 50% to 90%, as the proof-of-stake chain will offer lower rewards, Beiko says.

In the next two years the amount of Ether that’s used for staking will probably increase from 8% to 80%, according to staking services provider Staked. That will reduce the Ether in circulation, potentially pushing up its value.

Stakers will be able to use the Ether they receive as rewards for ordering transactions, but not the Ether that they stake — at least not until another software upgrade, expected six months or so after the Merge. Stakers are more likely to hold their Ether for the long term than miners, who often need to sell some to cover electricity costs, says Kyle Samani, co-founder of Multicoin Capital.

Source Bloomberg LP

Binance receives in-principle approval to operate in Kazakhstan

Crypto exchange giant Binance has received an in-principle approval from regulatory bodies in Kazakhstan to operate #crypto services.

According to a release on Monday, the Astana Financial Services Authority (AFSA) has given it the preliminary nod to operate a Digital Asset Trading Facility and Provide Custody in the Astana International Financial Centre (AIFC).

The AFSA is the first regulator in Kazakhstan to grant an in-principle approval to a Binance entity. This still requires Binance to complete the full application process, which is expected in due course. Then, Binance will be able to provide its services as an operator of a Digital Asset Trading Facility and Provider of Custody in the Astana International Financial Centre.

“Large investors seeking new markets need clear-cut and well-managed rules, as well as high standards of regulatory practice,” said Nurkhat Kushimov, CEO of AFSA. “When a regulator meets these requirements, it creates collaboration based on trust and an ecosystem where players can work safely and efficiently. We believe that Binance’s work will further develop this vibrant ecosystem of digital assets industry locally and regionally.”

The approval follows a smattering of regulatory agreements for the exchange across Europe and the Middle East. In July, Binance’s Spanish subsidiary was given the go-ahead to operate as a crypto service provider in Spain. The approval had been pending since the end of January and follows European registrations with both France and Italy. The company also recently received licenses to operate in Bahrain and Dubai.

Source The Block

Use of specialized software by business size

The larger a business, the more likely it is to use specialized software to handle critical tasks.

The number of SMBs using specialized software for accounting, payments processing, and other critical business tasks rises dramatically once the organization grows beyond 10 employees. In other words, the bigger a business gets, the more pronounced the pain points become causing them to turn to software that can tackle complex workflows. This aligns with recent research by the OBIE4 in the UK which found that companies with at least 5 employees were three times more likely to use cloud accounting software that companies with fewer than 5 employees.

Findings by Codat also show that SMBs acquire financial software in a particular order. As a company grows, it is likely to purchase accounting software first, followed by software for incoming and outgoing payments, demonstrating how integral this software is to the day-to-day running of a business.

SMBs with 5+ employees are over 3 times more likely to use cloud accounting software than SMBs with fewer than 5 employees.

Are SMBs benefitting from more efficient operational software?

Companies sharing permissioned data to overcome inefficiency and generate insights are predicted to add $100tn to the global economy by 20301 . Likewise, the World Economic Forum (WEF) expects embedded finance to reach $7.2tn in the same timeframe.

These statistics are indicative of the ‘integrated ecosystems’ trend that has firmly taken hold in financial and software markets, giving rise to an #opendata and #openfinance model.

But are small and medium-sized businesses (SMBs) in the US benefiting from the new technologies and efficiencies through data sharing? Or are they just juggling more software and systems, experiencing increased complexity, and chasing lower costs in the face of the current economic uncertainty?

What do the ‘tech stacks’ (i.e., the combination of software a company uses to run its operations) of SMBs reveal about their day-to-day operations? What do they invest in, and what are the top features and qualities they look for when assessing a potential provider?

Codat worked with independent research agency, YouGov to answer these questions, and more. Together, we surveyed founders, owners, and directors at micro, small, and mid-sized US companies (up to 500 employees) to learn how they use, manage, and evaluate their digital tools. We surveyed 526 owners or senior decision-makers in SMBs representing different business types, ages, and revenues. The data was weighted to the profile of US SMBs.

Source Codat

In Defense of Pay-as-You-Go: Revenue volatility doesn’t mean risky

When a PAYG fintech company experiences high revenue volatility, we often hear investors jump to the conclusion that these companies are risky because no one can predict whether their revenue will go up or down tomorrow. Uncertainty around short-term revenue is expected in PAYG fintech, but volatility alone is a poor measure of the risk that actually matters: how revenue evolves over the long-term. Assessing this requires looking at many qualitative factors, like how core you are to a customer’s operations, or your rate of product innovation, or how you go to market. None of these things can be understood just from looking at a line chart in isolation.

In PAYG fintech, revenue may oscillate more along the way, but entrepreneurs should focus on maximizing revenue over a long time horizon. To illustrate this, let’s consider a hypothetical example of the revenue earned from a customer by four different companies.

While the software application company has the lowest volatility, all three PAYG companies capture more revenue over a ten-year period, despite the absence of a multi-year contract. For the PAYG companies, the smoothness of revenue depends on how recurring or episodic the customer need you’re solving is. A payroll provider processes payments multiple times per month, for instance, while a mortgage broker may only process a loan for a customer once a decade. Regardless of the volatility, the persistence of the need these businesses solve and how well they solve it determines long-term revenue. What matters is the area under the curve, not the smoothness of the line.

PAYG companies can be particularly hard hit in a downturn, and the short-term pain can be intense across layoffs, low morale, and more difficult fundraisings. A down cycle may also test the resolve of employees and investors, but this does not mean that important and high-quality businesses are not being built. We continue to believe in PAYG fintech companies and the opportunity for long-term growth ahead.

Source Andreessen Horowitz



Sam Boboev

Fintech fan with product and technology background. Subscribe