Vinod Khosla: Future of Healthcare

Notes on a talk of Vinod Khosla, the founder of Sun Microsystems and the founder of Khosla Ventures, on the future of healthcare and moonshot thinking.


Vinod Khosla is the founder of Sun Microsystems and the founder of Khosla Ventures. In an interview with StartUp Health he talks about the future of healthcare and moonshot thinking for entrepreneurs and investors.


  • At the time we were thinking we are weeks away from a pandemic, we were in fact well into the pandemic already – we just didn’t know it
  • This has direct parallels in healthcare, for example: if you are a cardiac or Alzheimer patient, you only know it 20 years after the condition started
  • The extremely fast development of vaccines showed us as a society that we are not limited by what we can do but what we think we can do
  • Covid-19 made us much more moldable and open to experimentation
    • In the last year, much more has been achieved through quick experiments which would usually take many years of experimentation and adaptation
    • I.e. treatments with hydroxychloroquine have been tested and then discarded quickly
    • Before Covid-19, most physicians would never have done a telecare visit, today most of them have done
    • This experimentation has made the system much more dynamic and flexible which will stay for the future
  • Today we have artificial categories of specialities in healthcare: for example primary care, endocrinologists, gastroenterologists, cardiologists
  • The cardiologists know exactly what is happening in their field, but the problem is they don’t know what the endocrinologists are doing – as a result an integrative care is not available to the patient
  • These artificial categories of specialities will be broken up in the future
  • We are well on our way to leveraging AI to do 80% of what doctors currently do
  • In the 21st century and near future, physicians will hence become the human element of care and AI systems will do the majority of the work
  • Only the best expertise (i.e. top 10 percent of physicians/diagnosticians) will stay a part of our healthcare system
  • We don’t believe it’s possible, but it is already happening in narrow specialty areas today
  • Telemedicine today is not an innovation, it is basically a video-call with some security and privacy guardrails
  • Healthcare is hard to change is because people follow their interests: i.e. not wanting to learn something new or adapting new ways of doing things, or they don’t want to leave their current revenue models.
  • Innovation in healthcare will happen through verticalization and build upon current telemedicine solutions, by doing so ± 80% of healthcare can be provided remotely (digitally) and simultaneously healthcare costs will become 10x cheaper
  • The capabilities of these remote medicine tools will increase rapidly – this is called Learning Health Systems, which are not possible the way healthcare is done today
  • The AI used in remote telemedicine tools can learn from every single interaction
  • The adoption of these AI-based tools will make the current healthcare system more efficient by removing costs from the system which leads to a better treatment of the patient at a much cheaper price
  • We will see a decade of an increasing percentage of verticals being handled by AI enabled telemedicine solutions: diabetic care, hypertension care, cardiac care, muscular scalp, mental health, women’s health, …
  • There are two types of adoption:
    1. Doing something incrementally better withing the existing system which is easier but not revolutionary and will not in a change in the way medicine is practiced (i.e. a better surgical robot within a hospital)
    2. Radically innovating, renovating and changing certain verticals with new business models as well as new technologies (i.e. Livongo or AliveCor)
  • In most areas of technology and business not all hard things are valuable, but all valuable things are hard → People who take the long perspective and invest in core technologies or learnings will have better payoffs over the long term
  • A fundamental revolution is both possible and likely to accelerate the level of care and Khosla’s vision for 2030 is completely different and bigger than where we are today
  • Collaboration between startups and medicine or bio-technology companies is generally BS, collaboration only works at much later scaling stages of companies – not early when they are innovating
    • Imagine if Tesla had gotten advice from General Motors on how to build a car → if we take advice from healthcare providers in radical innovations we will end up in an equivalent place
  • Collaboration has a role in scaling up
  • When you need real out-of-the-box thinking, you can’t rely on experts in an area but the questions those experts ask are extremely valuable to the entrepreneur

Vinod Khosla’s 2030 Vision

  • In essence a patient gets diagnosed based on their symptoms, which is fundamentally a bad idea
  • Most of the diseases started a decade before you noticed them
  • Eliminating symptoms-based medicine is a key goal by 2030
  • For example: you shouldn’t have to lose your memory before you’re diagnosed for Alzheimer’s → The biochemical changes should’ve been monitored
  • Besides of symptoms-based medicine, we also have tradition-based medicine → there are hundreds of care pathways in each hospital system
  • Healthcare has improved constantly and incrementally – in some areas significantly – but the practice of medicine has not been changed
  • An important step is to change the practice of medicine into the science of medicine
  • The medicine of science will indicate that you are going to have a problem in 10 years (instead of treating the symptoms in 10 years)
  • Khosla will be shocked if in 2030 we don’t have a test for every patient as part of an annual physical and annual full-body image
  • If we have 10,000 biomarkers we will also get past trying to pretend humans can look at it, and we will have AI-based systems
What was the most significant development in digital health over the past year – specifically with implications for low and middle income countries?
  • When systems will become digital and care become remote it will be a lot cheaper
  • When the AI telemedicine system can to 80% of what a physician does – we will scale the number of physicians – which has direct implications for low-income countries
    • In Tanzania, there is one physician for every 50,000 population (Khosla’s recollection) which is a real problem, future systems will immediately reduce this number to one physician for every ± 10,000 people equivalent
  • Most likely in 10 years we will have free primary care – almost like Google searches today – to every person on the planet
  • Scalable medicine will be cheaper medicine
  • We will rely on an AI with a human-relationship-person
  • Khosla calls oncology the easiest possible discipline to turn into an AI oncologist and encourages any listener to exchange ideas with him
  • Someday you will have better cardiac care in a village in India – which relies on an AI system – than at Stanford – which still relies on the cardiologists they just hired. The cardiologist will be an expert, but the AI will have 25 years of learning and improving every single day
Current Status Quo
  • There has likely not been enough innovation yet. But many companies, entrepreneurs and scientists have already built lots of bits and pieces. At one point we will have enough of those pieces to see an exponential adoption
  • We are already collecting real-time contextually rich data and within the next decade the system will be rich enough, and we will see an auto-catalytic system phenomenon
Building and Investing in Exponential Healthcare Startups
  • European venture capital has only worked for linear innovation
    • European investors turn every radical idea and find was to reduce risk to the point that the probability of success goes up
    • But as a consequence the successes are marginal to inconsequential
  • Moon-shots are a theme for Vinod Khosla and the Silicon Valley – I would rather have a 90% chance of failure and a 10% of changing the world
  • Most huge innovations necessarily mean you have to take a large risk – whether it is a technology risk, a science risk, or a business model risk
  • You have to have the moon-shot mindset in everybody in the ecosystem
    • If you call a senior executive in Germany and ask him to leave his 25 tenure at Pfizer and join a five-person company you are not likely to see that answer – in Silicon Valley the answer will likely be “Yes!”
    • In Silicon Valley everyone – from the scientists to the entrepreneurs and management think outside the box and investors say “I don’t care about these small outcomes, go for the moonshot”

My willingness to fail gives me the ability to succeed.

Vinod Khosla
  • It is significantly important that an entrepreneur finds the right investor
    • If the entrepreneur does not want or cannot take the risks for moonshots he needs an investor who matches this incremental and slow growth
    • It is important that the entrepreneur picks the investor based on his own personality
    • For all kind of risks and ambitions there is a good investor – but the investor will be different depending on the risks and ambitions
      • Set your goals and find an investor to match you
    • Example: it is really easy putting a physician visit on a video call, call it telecare and be successful, it has been successful, but it is not an innovation (anyone who says they do video visits and AI → it’s not possible, you can’t claim both and this is mathematically wrong)
  • “Tell me what you are going to do for the patient first, what radical innovation are you going to make?” later, along the way we will start building code, revenue models, and so on
  • Instead of focussing only on the domestic (i.e. U.S.) market where maybe only five percent of the world’s population is subject to the FDA but build solutions for the other 95 percent, who have no alternatives
  • First, find the MVP niche segment which can allow for radical innovation
  • Book recommendations:
    • Loonshots by Safi Bahcall
    • Books by Eric Topol on healthcare and medicine innovation
  • I have not seen somebody who is an expert in an area (only possible exception is deep pharma) and not one large organization innovating – no expert has innovated broadly in an area to have large impact
  • “The more you know in an area, the more limited you are in thinking what you can do as opposed to what you can really do”
  • Experts come with many biases of how things are done or not done
  • There is a huge difference between a zero million dollar company and a zero billion dollar company – the difference is not the business plan for the next three years but in how you think about what you are doing
    • Do you pick a base camp where you can get revenue and profitability or do you pick one that leads to Mount Everest?
    • Have the vision and build the teams you assemble for the long-run
    • Build the human-resources for the long run while you build the tactics for the next 12 months
    • The team is the most important ingredient of whether you become a zero million dollar company or a zero billion dollar company
    • You can get some experts on the team, but the guiding light must be the entrepreneur, if the expert says it cannot be done, ask for first principles thinking not experience-based thinking but never let them change your vision

Leave a Reply

Your email address will not be published. Required fields are marked *

I accept the Privacy Policy