Dr. Palantir: The Great Healthcare Reset
While some investors scoff at Palantir's valuation, others are in awe of its potential. Why was this company, founded in 2003, taken public in the middle of a pandemic? Cynicism may dictate that this was so that management could cash in on high valuations and a rush for speculative assets, but if so, why has Palantir since deployed capital into its own SPAC investments?
From this chart we can see Google's 'beautiful compression' as the company gained scale and accrued value. This is despite the fact that Google went public in the middle of a secular bear market that lasted from 2000 to 2013. $100 invested in August 2004 would be worth approximately $5,750 today. (Image Source: author's work; data provided by S&P)
In the past, we have argued that Palantir is the Facebook of the industrial internet. In this article, we will dive into the complexity of the healthcare industry and examine the transformational potential embedded in data that is scattered across companies and government agencies. We will examine how a recent investment by Palantir gives clues into management's thinking. We make the argument that the alignment of industry and public interests provide the opportunity for dramatic changes to occur.
Making Sense of the Healthcare Industry
We will primarily analyze the healthcare industry from a U.S. domestic point of view. The role that the healthcare industry plays in the U.S. economy cannot be understated, comprising 18% of GDP. The U.S. spends significantly more than any other nation on healthcare, ($3.3T or $10,500 per capita) but ranks 22nd in life expectancy. The U.S. is home to 7 large pharmaceutical companies, over 20 major insurers, over 500 biotech firms, over 6,000 hospitals, and nearly 90,000 pharmacies. All of this is deeply integrated into approximately $1.3T in annual government spending.
On the surface, the healthcare industry seems tortuously complex. However, breaking this industry down into several different categories, as defined by its largest companies, makes this complexity more manageable.
(Image source: Author's work)
Though this is merely an approximation, it gives us a useful framework when considering the industry structure. While the large pharmaceutical firms (there are others not listed here, such as Novartis and AstraZeneca) typically serve as a political piñata, the reality is that it has been the large insurers (also known as 'managed care' companies) that have benefited most under this system.
Performance in the pharmaceutical industry has been mixed on a historical basis, but insurers have consistently broken revenue records. This has been matched by share prices steadily setting all-time-highs. This is because of decreasing IRRs for pharmaceutical R&D projects, a situation that the pandemic has alleviated.
While the pharmaceutical industry is moderately concentrated, hospitals and hospital systems are highly fragmented. Healthcare facilities in the U.S. range from industry giant HCA to non-profit and academic institutions operating in a single location. Astute readers will note that this means the data infrastructure held by these institutions is also extremely fragmented.
The Role of Government
To understand the structure of the industry we must lastly consider the overarching role that the government plays. Most of this is organized under Health & Human Services (HHS), a cabinet-level department that manages several key agencies:
(Image Source: Author's work)
Outside of HHS, the Department of Defense has also played a major role in the healthcare industry. The Defense Advanced Research Projects Agency (DARPA), is essentially the R&D division for the entire DoD. DARPA's early and consistent investments in Moderna (MRNA) show just how successful a public-private partnership can be.
"We have been thinking about and preparing for this for a long time, and it’s almost a bit surreal,” said Amy Jenkins, manager of DARPA’s antibody program, which is known as the Pandemic Prevention Platform, or P3. “We are very hopeful that we will at least be able to have an impact on this outbreak. We want to make a difference."
-The Washington Post (7/30/2020)
This support enabled Moderna to launch its first product, the world's first mRNA vaccine. While the U.S. federal response was characterized as unprepared and chaotic, Moderna began work on a vaccine several days before the WHO announced there was "no clear evidence human-to-human transmission", and nearly two weeks before the city of Wuhan was locked down by Chinese authorities. Furthering the case for public-private partnership, Moderna's vaccine was developed in partnership with the NIAID.
The Potential of Data
The willingness of the American public to participate in experimental vaccines distributed under emergency protocols (of which they have zero legal recourse in the event of adverse effects), is a measure of how much we value human life as a society. Many more lives can be saved or extended if we can unlock the power of the data that is buried in the healthcare industry.
The uses of data in the healthcare industry are broad and come in many different disciplines. Biostatistics, bioinformatics, computational genomics, computational biology, actigraphy, and electronic health records (EHRs) all describe ways in which data is used in healthcare. Some of this data is open-source, such as the NIH's Human Genome Project or Google's AlphaFold database. However, most of this data is siloed in hospitals, fitness trackers, government databases, research labs, and so on.
If this data could be combined, the possibilities stretch beyond our current imagination. For example, advances in ASICs and inexpensive storage have enabled powerful discovery analytics in many other industries. Similar tools could be applied here - with life-saving and profound results. Clinical trials, analyzing treatment effectiveness, understanding risk factors, patient monitoring, tracking diseases, are all areas where increased collaboration could create progress. Such data collaboration would open the door to advances in personalized medicine and holistic health.
A map of Palantir customers and partners. Palantir is well positioned to build a collaborative data ecosystem. (Image Source: Author's work)
*The blue and yellow M reflects the entity known as "Merck KGaA, Damstadt, Germany"
One of the lowest hanging fruit is "real-world evidence" (RWE), which is essentially the use of 'real-world' databases as opposed to controlled clinical trial data. Some readers may be surprised to find that the use of data in the healthcare industry is so nascent, but that is emblematic of the opportunity.
While the venture capital industry is buzzing with excitement surrounding a flurry of companies building "RWE" datasets and tools, few have realized that Palantir is already at the center of a budding ecosystem. Palantir has made its intention to enter this field clear, but outside the excitement of Palantir investors this has gone mostly unnoticed.
Palantir's advantage is threefold. First, Palantir specializes in handling unstructured data, such as physician notes or patient records. Secondly, Palantir has unparalleled experience connecting disparate datasets, stitching together unorganized data sources. Lastly, Palantir is already working with an array of healthcare companies and agencies.
All of this software is ready to go, fully packaged with proper access controls, integrations with modeling tools (such as R Studio), and the ability to export workloads to high-performance computing clusters.
A changing regulatory environment in the U.S. may act as an accelerant. Major changes to HSS policy are enabling consumers to have access to their health records, which they may choose to integrate into apps on their phones. This creates the opportunity for people to quantify and monitor their health the same way they quantify and monitor their social lives, with what will likely prove to be a much more positive impact on their wellbeing.
Hospitals may also sell some patient data. This is HIPAA-compliant provided that the data is anonymized to provide sufficient privacy, including genetic data. This is another area of expertise for Palantir, employees of which quite literally wrote the book on privacy architecture. Note that when companies such as Google purchase healthcare data, the massive data repositories that they retain (such as GPS locations from Google phones) increases the likelihood that they can deanonymize such data.
Healthcare in the U.K.
Unlike the United States, the U.K. has a publicly funded healthcare system, the National Health Services (NHS). Because this system is so different from the United States, it may be easy for investors to overlook the Palantir-NHS partnership.
The U.K.'s NHS is the largest non-military public organization in the world, employing a staff of 1.7 million. This system is impressively successful; British citizens have a greater life expectancy than Americans while healthcare spending is just a fraction of that of the United States.
NHS began working with Palantir in March of 2020, creating the NHS Foundry System. This system gives the NHS deep insights, organizing data from hospitals, primary care networks, community pharmacies, and vaccination centers. Palantir also provided Covid-19 tracking software to NHS. It is worth noting that the NHS also played a role in the development of Cambridge-1, a new Nvidia (NVDA) supercomputer that will 'accelerate U.K. healthcare research.'
The construction of this new supercomputer is yet another hint that something powerful is brewing. If NHS Foundry is extended to R&D, it will create a new model for public healthcare, and vastly accelerate the field of health sciences. (Image source: Nvidia)
There have been rumors that Palantir's role with the NHS will gradually expand. Palantir's software platform could unlock the full potential of the U.K.'s public health data, while keeping it in the hands of NHS, who would also decide where and how it is accessed. This is quite literally how Palantir works, empowering the user.
So it might surprise you that Palantir's empowerment of the NHS has been challenged by 'community interest groups', an oxymoron if there ever was one. One such group is openDemocracy, which successfully bullied the U.K. government into an agreement "not to extend Palantir’s contract beyond COVID without consulting the public." OpenDemocracy is partially funded by the Rockefeller Brothers Fund. Note that openDemocracy has not attacked Rockefeller Foundation's support for efforts to create biometric ID programs.
Despite the controversy, Palantir has been pragmatic in its efforts, directly investing in London-based start-up Babylon, which will soon go public as part of the Alkuri Global SPAC. Babylon connects patients with doctors, provides AI-powered digital checkups, and helps users navigate the UK NHS system. In our view, it is one of the most under-followed companies in tech.
Services offered by the Babylon app, (Image source: Babylon Health)
It's not hard to see why. Imagine an app that provides you information about your health the same way that Bloomberg provides news about financial markets. Imagine deep health insights leveraged from millions of other people, the same way Amazon builds product recommendation engines. If Babylon, Palantir, and the NHS are able to broker a hypothetical partnership, the U.K. will have the most modern and sophisticated healthcare system in the world. Millions of British citizens will tangibly benefit.
With the recent changes in U.S. healthcare data regulations, it's worth noting that Babylon is also expanding domestically.
BGI Group (Beijing Genomics Institute)
While hypothetical cooperation between Palantir, NHS, and Babylon has the potential to modernize healthcare, China's BGI Group is revolutionizing it.
BGI Group was formed in 1999 to participate in NIH's Human Genome Project. It is closely tied to the Chinese government, if not explicitly state-backed, having harvested data from millions of women (including in the U.S.) in cooperation with the Chinese military. BGI Group's work with the Chinese military constitutes the largest "population quality" project since "rassenhygiene" policies in Nazi Germany. It's one of the most valuable private companies in the world, with only a subsidiary (BGI Genomics) trading publicly.
BGI's approach centers around genomics, which is sort of like the 'big data' version of genetics. BGI Group offers IQ tests, pregnancy tests (complete with genetic analysis), genetic tests for human traits, and more. BGI Group made a strategic acquisition of U.S.-based Complete Genomics in 2012, and now has enormous capacity to sequence human DNA samples, driving the cost down to just $100. BGI uses this to create genome databases that it can then analyze with military supercomputers. BGI recently unveiled a “million-scale prototype robot” designed to do this on an industrial scale.
Whole-genome sequencing, the technology that drives BGI’s business, is no longer particularly new. But Wang says genomics is about to become the core of modern medicine, for several reasons: Sequencing is becoming cheaper and more reliable; research is advancing to the point where genetic findings can underpin treatments; and governments—above all, China’s—are encouraging their deployment at large scale.…“Nowadays, medicine mostly comes from the industrial revolution. It’s physical, it’s chemical: Kill the tumor, poison the tumor, burn the tumor,” Wang says. “We go back to real biology."
—Bloomberg Businessweek, (11/13/2019)
BGI believes it will eventually be able to tailor medicine to individuals' unique DNA. As the US organizes a de facto bailout of Intel on the realization that the country is falling behind in the fabrication of advanced semiconductors, China is pulling ahead in yet another area. It is likely that a strategic response will be necessary, such as the creation of a national gene database. Such an effort, while still hypothetical, would most likely leverage the existing infrastructure that Palantir has already created.
U.S. government advisors warned in March that a vast bank of genomic data that the company, BGI Group, is amassing and analyzing with artificial intelligence could give China a path to economic and military advantage. As science pinpoints new links between genes and human traits, access to the biggest, most diverse set of human genomes is a strategic edge. The technology could propel China to dominate global pharmaceuticals, and also potentially lead to genetically enhanced soldiers, or engineered pathogens to target the U.S. population or food supply, the advisors said.
If citizens are largely willing to take experimental vaccines, we can only assume that they will be just as willing to create a national gene database for the greater good. Everyone would benefit. Given the diversity of the United States, such a project would be one of our greatest national assets. The genetic diversity of the United States is so valuable that U.S. intelligence officials warned that BGI Group was attempting to collect genetic data under the guise of Covid-19 tests.
Covid-19 as a Catalyst
The purpose of this article is to provide context to Palantir's opportunities in the healthcare industry; it is not an analysis of virology. However, we must provide appropriate considerations to the origin of the pandemic because it has been a critical turning point in the history of Palantir.
Some have postulated that the origin of Covid-19 is the Wuhan Institute of Virology (WIV), China's only BSL-4 lab. There are many anecdotes to this theory. Covid-19 potentially surfaced during the 2019 Military World Games, held in Wuhan. If so, foreign military personnel were among the first infected. One athlete noted that the streets were "empty" as rumors circulated that the government "warned the inhabitants not to go out." The alleged origin of Covid-19, Huanan Seafood Wholesale Market, is located just miles away from the WIV.
The NIAID repeatedly provided funding to an organization known as the EcoHealth Alliance for a project titled "Understanding the Risk of Bat Coronavirus Emergence". The EcoHealth Alliance in turn was funding gain-of-function (genetic modification) research of bat coronaviruses at WIV. The closest known relative of Covid-19 (96% genetic match) is a bat coronavirus that was collected by WIV researchers from a cave located approximately 560 miles away from Wuhan. According to a paper available from NIH, these researchers discovered a "rich gene pool."
And we have now found, you know, after 6 or 7 years of doing this, over 100 new SARS-related coronaviruses, very close to SARS…
…you can manipulate them in the lab pretty easily. Spike protein drives a lot of what happen with coronavirus, in zoonotic risk. So, you can get the sequence, you can build the protein, and we work a lot with Ralph Baric at UNC to do this. Insert into the backbone of another virus and do some work in the lab.
—Peter Daszak, President of the EcoHealth Alliance, December 2019
The author of the Lancet paper that condemned the lab origin theory as a baseless conspiracy was also the head of the EcoHealth Alliance. This individual also headed the WHO investigation that concluded Covid-19 did not come from the WIV. Dr. Anthony Fauci, the head of the NIAID, is unconvinced. Furthermore, some have speculated that the NIH was involved in funding as many as 60 projects at WIV.
Taken in aggregate, these facts suggest the possibility that Covid-19 has been genetically engineered. Some reports have speculated that the NIAID has been aware of the possibility for quite some time. If so, this may be unlike any pathogen we have ever seen. For example, Larry Brilliant, the famed epidemiologist that led the effort to end smallpox, has warned that the Delta variant may be the 'most contagious virus ever'. The Bulletin of the Atomic Scientists has likened it to opening Pandora’s box.
Containing and managing the virus may stretch on longer than expected, requiring the need for tools such as contract tracing indefinitely. A current roster of DARPA projects also suggests this possibility:
(Image source: Author's work; summarized from DARPA website)
DARPA also awarded an additional grant to Moderna for the development of portable vaccine production machines, capable of producing hundreds of doses in a matter of days. Note that ideas similar to the ones DARPA is working on were recently discussed by the CSO of IMEC, including the idea of 'lab-on-chip'.
Australia has lifted the veil completely with the government of New South Wales stating that it "will probably inoculate its residents against COVID-19 periodically for years to come," adding that "we need to get used to being vaccinated with COVID vaccines for the future." Australia has also constructed quarantine camps to 'remotely isolate' individuals that are infected with the virus.
At first glance, it may be easy to dismiss Australia's actions as totalitarian overreach, particularly when the country held a nationwide manhunt for a guy who sneezed in an elevator. But consider that just months into the vaccine rollout, new vaccine-resistant variants are already spreading, including in the United States. Pfizer is working on a twice-daily Covid-19 pill. The CEO of Moderna outlined the possibility of seasonal waves, the possibility of declining vaccine efficacy, and suggested that countries may want to 'stockpile boosters'.
These facts indicate that there is a lot that we do not know about Covid-19. Not only is it likely that the pandemic will last much longer than expected, a possibility that has been suggested by experts, but effectively managing it may also require cooperative data analysis on an unprecedented scale. The CDC is already using Palantir's Covid-19 tracking tools. Foundry's privacy architecture manages the risks of overbearing surveillance. By design, Foundry distributes access permissions and allows data to be retained across many different organizations, as opposed to centralizing both.
If this cooperative data infrastructure materializes, it may create permanent changes in the healthcare industry. In particular, it may lay the groundwork for cooperative genomics research, if the U.S. comes to the conclusion that it is lagging China's grand ambitions in this strategically important field.
We can talk about statistics... We have five billion rows of data. We have a million covet patient EHR records, starting prior to Covid.
But what really is amazing about N3C is... In the public health world often data is splintered and it and it exists in silos all over the country. And what we've been able to do at N3C, really for the first time, is create a nationwide data set that everybody has been willing to give so we can address this pandemic.
—Ken Gersing M.D., NCATS Director of Informatics, speaking at Palantir Double Click (4/14/2021)
To some extent, this thesis is already playing out. To combat the pandemic, NCATS (a division of NIH) orchestrated a collaborative program, N3C, that is "systematically collecting data derived from electronic health records from different institutions" to create a "national data resource." This data is being used for the explicit purpose of combatting the pandemic, but this approach could be extended to other areas of research. The N3C platform runs on Palantir via AWS GovCloud.
(Image source: BloodCounts!)
Palantir was also a sponsor of the Trinity Challenge, which selected BloodCounts! as a finalist. The U.K.-based effort to re-analyze the 3.6B blood tests performed globally each year, repurposing them as an early detection system for outbreaks. Cambridge's Prof. Carola-Bibiane Schönlieb detailed this project in a video that featured '#Palantir' in the description.
Quantum Computing as a Catalyst
Astute readers will note that in order for futuristic personalized medicine to become feasible, we will most likely need far greater computing power than what is available today. We return to quantum computing (see earlier discussion here), with an update on its progress.
PsiQuantum and GlobalFoundries, working in partnership, recently announced that they already have some of the critical components to a quantum computer in production. PsiQuantum says that 'by the middle of the decade' they will have all of the necessary manufacturing processes to build a quantum computer at scale. Google made a similar vow, with Google's quantum computing team setting the goal of 'within the decade'. Google has already mapped the entire human proteome using conventional computers running AI/ML algorithms.
Quantum computing is a critical catalyst because it will create the computing power necessary to run molecular and atomic-level biology and chemistry simulations. At scale, a functional quantum computer would quickly reshape the field of medicine.
The world seems ready to unlock the opportunities at the intersection of biology, chemistry, and data. Doing so will require a rethinking of how data is handled in the healthcare industry, and extraordinary levels of cooperation. We believe that these are challenges that Palantir is uniquely suited to handle, with a software platform that empowers users.
Success will rest upon Palantir's ability to overcome controversies and broker industry partnerships. The outcome has the potential to not only reward investors, but reward society as a whole.
(Image source: Author's work)
Even in an optimistic scenario where the pandemic subsides quickly, the healthcare industry has been significantly augmented. Pharmaceutical firms are loaded with cash. That money should be spent on R&D, not squandered on buybacks. We should learn from the mistakes of Intel (INTC).
To enable the next generation of medicine, national data assets should be built and expanded. Palantir has the software needed to orchestrate this. We should start right away.