Subscribe

1. Is the Marketing Department Running GM?

This week GM and Nikola formed a strategic partnership. Nikola gave GM an 11% ownership stake valued at $2 billion, and in return GM agreed to engineer, validate, and build Nikola’s hydrogen and battery electric Badger pickup trucks on a cost-plus basis.

Previously, ARK shared its research demonstrating that hydrogen fueled vehicles are more costly to operate than battery electric vehicles. In ARK’s opinion, the partnership with Nikola signals GM’s desperation, with management seemingly ceding control of corporate strategy to marketing and messaging.

ARK has questioned management’s strategic directions ever since GM unveiled its autonomous electric strategy with the slogan “zero crashes, zero emissions, zero congestion,” none of which is true. Instead, we believe autonomous electric vehicles will result in fewer crashes, lower emissions, and increased congestion.

Instead of focusing on the all-important battery platform, GM is changing the subject to hydrogen fuel cells and small-scale manufacturing. To survive in the transition to autonomous electric vehicles, ARK believes that GM must focus on technology, not marketing.

 

2. Where is the Digital Economy Heading?

In our last write-up of Apple vs Epic Games, we ended with a question: will their battle, if not war, set a precedent for the digital economy? Before answering that question, we must understand where the digital economy is heading.

On the surface, the fight between Epic Games and Apple seems to revolve around Apple’s 30% platform fee. In reality, we believe the war will determine the future of the metaverse.

In its latest filing against Apple, Epic Games declares that “a metaverse is a multi-purpose, persistent, and interactive virtual space.” In our view, the metaverse is inevitable, if not already here in the form of Epic’s Fortnite.

In an always-on virtual world, physical goods and digital goods should be equal in importance, suggesting that the metaverse will rival the real economy in size. Apple’s 30% tax on digital goods could stunt the digital economy’s growth.

Weighed down by a 30% ‘sales tax’, will the metaverse reach its potential? Epic Games would argue no.

 

3. Molecular Prognostic Testing Could Reduce the Overtreatment of Cancer

ARK believes that multi-cancer, liquid biopsy screening tests will detect dozens of cancer types earlier than ever thought possible, raising the concern of overtreatment. Not all tumors cause symptoms or are lethal. Slow-growing – or indolent – tumors could result in unnecessary surgeries and exorbitant costs. Overtreatment is the principle argument against widespread cancer screening.

In our view, molecular prognostics – or tests involving next-generation DNA sequencing (DNA) and machine learning (ML) – could lessen overtreatment by predicting the lethality of tumors more accurately. While not eliminating overtreatment, such tests could reduce patient morbidity significantly. Veracyte’s (VCYT) Afirma test set a precedent in categorizing the risks associated with cancerous thyroid nodules. With more ML training data, such risk-stratification should be possible for other cancer types.

Researchers from the University of Cambridge and the European Molecular Biology Laboratory (EBML) recently published a landmark study evaluating the predictive power of their molecular prognostic test on a form of esophageal cancer. The researchers trained an ML model on 15 years’ worth of retrospective NGS data from roughly 800 biopsies. Their model provided two benefits: (a) it directed the researchers to surgical intervention for the roughly 50% of patients with lethal cancer, and (b) it would’ve lowered the frequency of testing for the other 50% of patients with indolent cancer. Importantly, the accuracy of the model’s predictions (AUC) improved from 70% to 89% as the ML training set expanded from 28 to 164 patients, suggesting that such models improve with data and time.

Molecular prognostic testing could reduce the human and economic burdens of overtreatment. While test designs may vary, we estimate that a molecular prognostic test could be priced near $500 at scale. Importantly, patients with low prognostic risk scores could be retested less frequently, further reducing costs. ARK plans to include the costs of overtreatment in our forthcoming multi-cancer screening model.

 

4. Could Man and Machines Integrate? #Neuralink

Elon Musk’s neural engineering company, Neuralink, is attempting to create brain-machine interfaces (BMIs). If effective, robots would insert a chip into a persons’ skull possibly without the need for general anesthesia. Flush to the skull, the chip is battery-powered to last all day and wireless with five to ten meters of range. In July, the FDA granted it a breakthrough device designation, though it is not in human trials yet.

Predecessor technologies have demonstrated efficacy, but Neuralink will need to prove the safety, privacy and value of its chip. The Utah Array and Deep Brain Stimulation were efficacious in controlling robotic arms, obsessive compulsive disorder (OCD), memory, consciousness and movement disorders.

While the chip holds great promise for the long term, the surgical robot could have an impact in the short term. Enabling human physicians with much higher definition, robotic surgery should result in more precision and control.

Neuralink already has enabled limb-tracking in pigs, suggesting that its foray into humans could be controlling artificial limbs. In the future, neurological diseases could become the focus, a huge unmet need given the $16 trillion global market in mental health treatments expected in 2030.