Vivek Wadhwa is building the AI-powered anti-Theranos. Now he’s moving it to the anti-Silicon Valley–India
I just remembered when I was a young entrepreneur going door to door on Sand Hill Road looking for financing. I remember how venture capital firms would make me wait for hours and try to appear smart in front of their partners by finding small flaws in my business plan. They were kings and kingmakers and treated entrepreneurs like beggars. Fortunately, the cost of starting software companies quickly fell, so much so that crowdfunding and angel capital were a better alternative. This shift in the balance of power has forced venture capitalists to turn to entrepreneurs.
However, this change has only occurred in the software domain. In capital-intensive fields like biotechnology, little has changed. Additionally, due to the spectacular failure of Theranos, venture capitalists are more risk-averse than ever – literally traumatized.
With exponential advances in technologies such as artificial intelligence, computing, sensors and synthetic biology, not only have costs fallen, but cutting-edge innovation is now globalized.
Today, there are much better places than Silicon Valley to develop world-changing technologies. One such country is India, where you can hire top-notch talent for less than 10% of what it costs in the Valley, and where the pathology data needed to train machine learning algorithms is available in abundance.
While San Francisco is the global center of AI development due to positive network effects, technology leaders like Sam Altman say India cannot build complex AI technologies. However, what I have seen in India leads me to believe that Silicon Valley’s advantage will not last long – and that its insularity, arrogance and overconfidence could lead to its downfall. As AOL founder Steve Case has long argued, the rest of the world is booming.
What I’m trying to build with my startup, Vionix Biosciences, is much more ambitious than Theranos claimed to achieve. I believe that in India I can build early versions of the technology for just $1 million, less than 0.1% of the total cost. $1.4 billion that Elizabeth Holmes raised (and squandered).
I’m approaching medical diagnosis in a completely new way, leveraging fundamental scientific advances made by a Chilean company I invested in over a decade ago, instead of just investing put money on the problem as the Valley often does. And rather than keeping my devices a secret and doing everything myself, I plan to make them available to researchers at several universities so they can validate the technology and do things with it that even I can’t. not imagine.
The technology changes the molecular structure of water, converting it into non-thermal plasma and water using the same amount of energy as a hair dryer. It allows real-time spectral analysis of organic matter. Just as DNA sequencing opened a new dimension in medical research by converting biology into letters, this technology can convert biological matter into a light spectrum that AI can decipher. It is similar to the techniques used in the gold standard for materials analysis, mass spectrometry, but without any consumables, sample preparation and mass-to-charge ratio measurement. The technology can analyze not only water, but also human fluids such as blood, urine, saliva and breath. It can do it in less than five minutes for the cost of electricity alone.
The problem is that it requires complex AI training to understand light patterns as complex as genomic data, which took decades to decipher. Training this AI will require tens of thousands of medical samples for each disease or cancer marker, which would be virtually impossible for a Silicon Valley startup in the United States to obtain. However, India has an advantage with its population of 1.4 billion. With informed patient consent and privacy protection, it is not difficult to gain inexpensive access to hundreds of thousands of biological samples in the many pathology laboratories that already analyze these samples with equipment advanced medical diagnostics.
Recruiting quality AI talent is another challenge. In Silicon Valley, entry-level salaries often top $150,000 a year, and employees expect sophisticated benefits and 35-hour work weeks, as well as the right to work at black and holding two or more jobs.
In India, hundreds of thousands of computer science graduates earn between $4,000 and $7,000 per year. I interviewed several prospective graduates and found that they were much more motivated and eager to learn than their peers in the Valley. I hired a student on the spot when he told me he wanted to work for me as an intern without pay, learn the machine learning tools I planned to use, then work 70 hours per week when he graduated in June. Sure, I’ll pay him way more than that and let him live, but you don’t see that attitude in Stanford or UC Berkeley grads.
The Indian government is also supporting the AI ecosystem. Rather than creating regulatory hurdles, restricting immigration and stifling the tech industry, India is creating a new fund to support its AI ecosystem. Prime Minister Narendra Modi’s chief science adviser, Ajay Sood, told me that his mission was to do whatever it took to facilitate entrepreneurship and support startups and that the government would do everything it could to welcome foreign companies like mine.
With all these factors in mind, it became clear to me that India was the ideal location for the research and development of my startup. I decided to forgo several scheduled meetings with Silicon Valley investors in favor of the dynamic and supportive ecosystem that is India.
Vivek Wadhwa is an academic, entrepreneur and author. His book, From incremental to exponentialexplains how big companies can see the future and rethink innovation.
