#361: PayPal’s Evolving Partnership With Apple Pay Could Come At A Price, & More

Last week, PayPal announced another partnership with Apple Pay. For some time, PayPal has processed Apple Pay payments through its Braintree merchant acquiring arm on the App Store and has competed with Apple Pay as a checkout option for consumers’ e-commerce purchases. Now, PayPal’s Complete Commerce service will offer Apple Pay as an option, enabling its merchants to accept payments not only through PayPal but also through debit cards, credit cards, and Apple Pay.

Complete Commerce is part of PayPal’s unbranded processing, which has increased from 19% of PayPal’s payment volume in 2018 to 30% in 2022. During the same four years, its core branded offering has lost six percentage points of share to Apple Pay, and other competitors, dropping from 36% to 30%.

By facilitating Apple Pay payments, PayPal probably enjoys healthy margins today, but longer term its promotion of Apple Pay could cannibalize PayPal’s branded offering and impact its bottom line significantly. According to our estimates, branded checkout accounts for ~30% of payment volume but, thanks to higher merchant price points and lower funding costs than those associated with credit and debit cards, roughly 70% of transaction gross profits.

Elon Musk’s recent tweet suggests that Tesla’s Full Self-Driving (FSD) system is making significant progress. Now exceeding one million miles per day, or ~350 million miles per year, FSD has soared from the 90 million cumulative FSD miles reported a few months ago in Tesla’s 2022 shareholder update. According to ARK’s analysis, Tesla’s current data acquisition doubling rate, or it’s Wright’s Law-based experience rate, accelerated 40% in the last quarter, an important stride toward its ambitious, fully autonomous software updates more swiftly than once anticipated.¹ Moreover, recent videos of FSD’s performance suggest that Tesla is closer than ever to launching a robotaxi service. ARK estimates that robotaxis could be the biggest driver of Tesla’s value over the next five years.

[1] Wright’s law states that for every cumulative doubling in production there is a corresponding increase in efficiency. In this case, the more miles driven the more efficient Tesla may be at rolling out software updates. Efficiency gains could come in the form of reduced time to output or resources used in the process.

The recent surge in advanced AI models has led to the potential for unprecedented automation and efficiency in various industries. The revolutionary developments also spark concerns about copyright issues associated with AI-generated content. The core concern for businesses stems from the fact that AI models like GPT-4 are trained on vast amounts of content, some potentially copyrighted, which raises questions about the intellectual property status of the generated content. As AI continues to reshape the business landscape, enterprises are seeking clarity on the legal framework for AI-generated content, including potential impacts to their operations.

One example is the ongoing lawsuit involving GitHub’s AI-powered code generation tool, Co-Pilot. Built on OpenAI’s Large Language Model (LLM), Co-Pilot helps developers write code more efficiently by generating suggestions and auto-completing lines of code. Plaintiffs in the case argue that in drawing on the code they contributed to GitHub, the tool infringes upon their intellectual property rights. The outcome of this lawsuit could set precedent for the legal treatment of AI-generated content and have far-reaching implications for businesses that use AI tools in their businesses.

In 2006, the discovery of Yamanaka factors catalyzed the field of epigenetic reprogramming. Four transcription factor (TF) proteins—called OSKM—allowed scientists to retrograde specific cells into stem cells. Now, induced pluripotent stem cells (iPSCs) have many applications in regenerative medicine.

More recently, researchers showed that exposing cells transiently to Yamanaka factors, without turning them into stem cells, rejuvenated them. While immature, interrupted—or partial epigenetic reprogramming—could become a viable therapeutic strategy for a range of diseases.

TFs work by latching onto DNA physically and controlling whether or not genes are turned on or off. By exposing cells to TF combinations, researchers can reprogram them. Comparing the genetic “settings” of current cells to target cells, scientists are starting to transform theoretical epigenetic reprogramming into an applied science.

Commercialization could start with T-cell therapy. Companies like Lyell Immunopharma (LYEL) are working on epigenetic reprogramming of T-cell therapies in solid tumors. Specifically, Lyell hopes to transform exhausted T-cells into naïve, memory-like T-cells. Unlike exhausted T-cells, memory T-cells are more capable of fighting cancer.