Honda-GM Deal Highlights the Difficulty that Traditional Automakers Will Have in Autonomous Mobility

Follow Tasha on Twitter @TashaARK

This week, Honda announced a $2.75 billion investment in Cruise Automation, GM’s autonomous vehicle subsidiary. In the planned deal, Honda will provide $750 million in equity upfront and $2 billion over a 12-year period. Previously, Honda and GM had collaborated on fuel cells, and Honda had relied on GM’s batteries for its own electric vehicle efforts. The new partnership will create electric autonomous vehicles without steering wheels, taking the steam out of Honda’s negotiations with Waymo, which may have broken down over technology rights.

ARK believes that the autonomous mobility-as-a-service market should be valued at $2 trillion in the equity markets today. Given the magnitude of the opportunity, we are surprised that GM does not want to maintain full ownership of Cruise Automation. GM seems to be providing the battery and autonomous technology, the crucial building blocks of an autonomous car, making it difficult to understand what value Honda could contribute to justify the dilution in GM’s ownership.

Softbank also diluted GM’s ownership in Cruise Automation in May, giving GM access to a number of ridesharing companies and a path to international expansion. Yet, this week we learned that GM is paying Softbank a 7% annual dividendfor Cruise, which is a pre-cash flow project. Perhaps GM took outside investment because the investment in Cruise is proving much higher than it expected initially, a source of concern to traditional auto investors now wondering if autonomous will be worth the ultimate investment.

In The Information’s piece on Honda and GM, one source claims that Cruise experienced a setback in the performance of its autonomous software when the team switched to the new Bolt prototype. Autonomous software and systems seem to be highly sensitive to the differences among vehicles, potentially another problem for traditional automakers working with a number of models and partners. Companies like Nissan have at least 40 models.

In summary, GM still seems to have an advantage over most traditional automakers as the Cruise team has demonstrated impressive technological capability in the past and has laid out some very aggressive commercialization timelines, supported by Mary Barra, GM’s CEO. If it launches a commercial autonomous taxi service in 2019 as planned, GM could beat many of its peers to market with significant first mover advantage, as autonomous taxi services should submit to natural geographic monopolies. That said, the equity deal with Honda highlights some of GM’s hurdles as a traditional auto player trying to break into new age transportation.

Is China Using Tiny Chips to Hack the US?

Follow James on Twitter @jwangARK

In a blockbuster story this week, Bloomberg reported that China has hacked several US companies and government agencies by slipping a tiny chip into computers destined for American server maker, Supermicro. Smaller than a grain of rice, the chip allows hackers to open a backdoor to Supermicro servers, enabling them to steal information from otherwise secure datacenters. Supermicro supplies servers to Amazon, Apple, Facebook, and the Department of Defense, among other companies and government agencies.

In a surprising twist, both Amazon and Apple issued detailed statements refuting Bloomberg’s report. Both assert that they have not found malicious chips inside Supermicro servers and highlight factual errors in Bloomberg’s story. Apple further volunteered that it is not under any gag order and has not signed non-disclosure agreements with any suppliers involved.

The wildly differing narratives suggest the truth lies somewhere in the middle. Perhaps the details of the hack were withheld from upper management, or Bloomberg’s sources exaggerated the extent of any hack. Given the severity of the claims, other sources likely will emerge in the coming weeks to shed light on this important controversy.

Digital Wallets Are Becoming the New ATMs for Financial Services

Follow Bhavana on Twitter @bhavanaARK

Several decades ago, ATMs emerged as an alternative to bank branches, offering customers easy access to financial services and helping banks to lower infrastructure costs. Their role in retail banking evolved considerably over the years to include check deposits and many information services. Yet, over the last few years, ATM growth around the world has flatlined, averaging 47 per 100,000 adults globally,with the US at roughly 225 and China now stabilizing at fewer than 100.

At the margin, digital wallets are taking share today and, eventually, could dominate customer-facing banking ecosystems globally. Wallets like Alipay and Mpesa already are providing access to a wide range of financial services to customers who are difficult to reach in China and Africa.  Even with a well-established financial infrastructure in the US, digital wallets are taking off and, at ~20 million today, have swamped the number of ATMs, as shown below.   

Miner Equipment Competition Is Escalating

Follow Yassine on Twitter @yassineARK

China-based Bitmain launched its last generation of Bitcoin mining machines, the Antminer S9, in 2016, with marginal improvements since then. Recently, it announced plans to roll out its long awaited 7 nm Application Specific Integrated Circuit (ASIC) in a new generation of mining machines at some unspecified time in the future. In the meantime, its competition has escalated considerably, as shown below.

These graphs suggest that its competition has surpassed Bitcoin in both performance and efficiency.  In 2016, the S9 had 3 times the hash power of its nearest competitor and was 80% more efficient in terms of both cost and electricity consumption. Today, the Antminer is in the middle of the pack on most metrics.

As Moore’s Law hits its limits and the ASIC design space loses its competitive advantage, the potential for exponential improvements in proprietary production will disappear. Consequently, Bitmain’s machines could continue to lag behind competitors not only in performance but also in cost.

Manufacturing Matters: A Single Cell Could Undermine the Impact of CAR-T Therapy

Follow Manisha on Twitter @msamyARK

Recently, one rogue cell among the 2 billion T-cells infused during CAR-T (chimeric antigen T-cell) therapy caused an aggressive relapse in the cancer of a patient just 20 years old, ending in his death.  Immunotherapy has resulted in the complete remission of cancer in hundreds of other patients, so what possibly could have happened in this unfortunate case?

The first two CAR-T therapies approved in the US in late 2017--Kymriah and Yescarta--treat acute lymphoblastic leukemia and aggressive non-Hodgkin’s Lymphoma, respectively.  Extracting a patient’s own T-cells, the therapy turbo-charges them with a CAR gene that can detect and kill cancerous cells, and then transfuses them back into the patient.

CAR-T therapy has been quite successful, so this rare case triggered a search for an explanation of its failure. In a Nature Studyreleased this week,  researchers reported a manufacturing error that probably prevented the  success of CAR-T therapy in the 20-year old patient.

Among the 2 billion CAR-T cells transfused into the patient was one leukemic B-cell also re-engineered with a CAR gene. When extracting a patient’s own T-cells, leukemic cells often are extracted in tandem. Post-hoc analysis discovered the leukemia B-cell was re-engineered with an “anti-CD19 CAR” gene, instead of a CD19 CAR gene, that camouflaged the cancerous cell and made the patient resistant to CAR-T therapy. The transformed leukemia B-cell caused the proliferation of a number of new leukemic cells without the CD19 cancer marker which went undetected until the patient experienced an aggressive relapse.

During manufacturing, the flow cytometry process was not sensitive enough to catch the rogue B-cell. While only 1 in 369 patients fell victim to this manufacturing error, it was fatal.

ARK believes that allogeneic, or off-the-shelf, CAR-T therapies will emerge to prevent this complication and dominate the CAR-T market in the future. Allogeneic administration involves healthy donor T-cells, eliminating the chance of transfusing cancerous cells into patients. Importantly, a single donor can create up to 1,000 patient doses at a fraction of the cost and time required to develop personalized, autologous CAR-T therapies. Allogeneic approaches to CAR-T therapy should ensure that the field of immunotherapy will continue the push toward a cure for cancer.