Comments on National Deep Tech Startup Policy (NDTSP)

Author note: A 3 month weekend course on public policy does not make me a policy expert. These comments are based on my current understanding and I might be misunderstanding some of the fundamental concepts involed and/or misunderstanding the policy itself. In addition, I have a B.S. & M.S. in Physics from IIT Madras and my wife has a PhD in Ecology from IISER Bhopal. I’ve seen and heard of the numerous issues that plague fundamental scientific research in Tier-I institutes in India. Please keep those in mind when you read my comments below.



4.B. Objectives: Deep Tech Ecosystem Priorities

4.B.9. Promote public and private adoption of indigenous deep technologies through
favourable procurement rules and innovative adoption mechanisms and tap into global
markets for Indian deep tech products.

In my current understanding, the USA (and the west) became the scientific and technological super power that it is because of supply-side initiatives before, during, and after the numerous wars that were waged over the past 100 years. There was a large demand for scientific and technological innovation by the Military-Industrial Complex which drove fundamental R&D and commercialization of research.

4.B.9 should be the first priotity.

4.C.2.4 The current National IPR Policy 2016 does not extensively address the eligibility of “computer programs per se, or algorithms” for patent protection. Most deep tech startups in India rely on advanced digital frontier technologies, which is heavily dependent on algorithms and APIs as the primary source of novelty. To ensure Indian deep tech startup competitiveness in the global market, it is imperative to recognise this uniqueness in the National IPR policy. Therefore, a well-defined mechanism should be established for evaluating the suitable amendments that encompass unique characteristics of digital frontier technologies within the scope of patent eligibility.



4.C.1.4 The difficulties in identifying high-potential patents as well as inefficient evaluation and prioritisation of publications and research result in missed commercialization and translation opportunities. To address these issues, this policy proposes the formation of an organisation - The Centre for Deep Tech Translation [CDT] to assess Indian research (publications, patents, etc.) for potential commercialization.

The top-down approach of formation of CDT will only get us so far. It needs to be paired with a bottom-up approach to educate the researchers regarding commercialization, IP generation and technology transfer. A larger cultural shift in Indian academia is also necesary but that is out-of-scope for this policy initiative.

Note: A lot has happened in the 10 years since I graduated college. In my understanding, there already are Graduate-level courses, in somecases mandatory, at Tier-I institutes regarding these topics. Such courses need to be emphasised and kept up-to-date.


4.C.1.7. The shift from being data-rich to becoming a data-intelligent destination is essential as deep tech startups play a crucial role in driving data demand. To catalyse innovation in Deep Tech, it is imperative to transform dynamic data repositories, underscoring the significance of data sharing and accessibility throughout the deep tech startup ecosystem. To this end, the policy aims to set up an Open Science & Data Sharing Platform to encourage collaboration and knowledge sharing within the deep tech community. Sharing research data, methodologies, and findings of startups will foster collaboration, transparency, and reproducibility. This can help in developing standardised solutions that are accessible and affordable to a wider audience. Implementing effective mechanisms to facilitate the transfer and commercialization of publicly funded academic research outcomes is essential to strengthen and propel the deep tech startup ecosystem.

Public-funded research should never be published behind a paywall.

4.C.3.2. Research and Development of deep technologies are time and money intensive. Long-term patient grants for selected science-based research institutions that form as sprawling hubs for deep technologies by routing CSR funding can be established. The Government may initiate a Committee to assess and analyse current CSR laws, recommend amendments, and advise on CSR grant disbursement such as contribution to incubators and R&D projects, as specified in item (ix)(a) and contribution to institutes/organisations, engaged in research and development activity (incl. Department of Science and Technology (DST), as specified under item (ix)(b) of Schedule VII of Companies Act, 2013. The existing CSR programs are usually short term (finishing the grant in the same FY or 2-3 years), or deployment of existing developed technology. The programs are also looking to create impact in quantifiable terms. The CSR fund, if donated to deep technology initiatives, must provide leeway for a) completely blue-sky research b) longer gestation period c) delayed impact. In case of the long-term patient grant, the CSR funding clause may suitably be amended to clearly include eligible premier privately funded institutions, or privately funded institutes of eminence as well.

Plugging the short-fall in R&D investments by the private sector by taking money out of CSR funding feels like the wrong way to solve the problem. Deep-tech, by definition, is risky and the R&D may or may not generate meaningful results for the society-at-large.

Instead, an increase in corporate tax rate should enable the country to fund the necessary work.


There are two key supply-side issues that haven’t been addressed in the policy.

  1. Incubating broad talent for Deep-tech startups
  2. Hardware ecosystem for Deep-tech startups

1. Creating talent for Deep-tech startups

The policy highlights a few ways in which talent can be groomed for deep-tech

  • how college courses can be modified to focus on project work instead of course work
  • how graduate students can be incentivized to pursue internships at deep-tech startups

But I strongly believe that the best way to increase the pool of talent for deep-tech talent is by increasing the monthly stipend that graduate students get. One-off grants, prizes/awards and internships will only reach a certain portion of the overall graduate student talent that the market can tap into for deep-tech startups.

2. Hardware ecosystem for Deep-tech startups

To my limited knowledge, a large number of hardware deep-tech startups (startups that involve significant hardware, electronics components) still rely heavily on China e.g. to manufacture the hardware, to procure the fundamental electronics components, to print the circuit boards. Without a local ecosystem that supports these activities, a hardware deep-tech startup will continue being reliant on China, have a larger lead-time from design to production because of procurement delays and become a riskier venture than it needs to be.

A policy to support deep-tech startups must fundamentally also support the creation of an ecosystem that these startups rely on. Large-scale consultation sessions need to be help with deep-tech startups individually to identify the ecosystem that they depend upon to ensure that the policy can take the necessary steps in supporting it.


@sudo_navendu - I would love to hear your comments on this draft policy, if you have the time.

1 Like

Brilliant breakdown @rahulporuri!

Monthly stipend of researchers in India is a valid point and I’m surprised that there is no mention of it in the policy.

Apart from the above-mentioned points, I feel that the policy gave less emphasis on the private sector and industry contribution to science and deep tech. There is a lot of governmentality with respect to creating the initial market through the public procurement route, providing seed funds through schemes like iDEX etc. But given the multiple pulls on the Indian government budget, increasing the size of Science will become a last priority on a rainy day.

Moreover, there has to be foresight in the identification of DeepTech sectors. We cannot bandwagon with the world and now focus on the technologies that are on the headlines. These technologies are on the headlines because people had been working on them for about a decade. Prof CNR Rao mentioned more than a decade ago why India should focus on nanotechnology but we have missed that bus as well. So a strategic effort on technologies that would matter after a decade or so will have to be identified and worked upon. The current definition of Deeptech startups would allow most startups that just mimic technologies being developed across the world for the Indian market.


The Office of the Principal Scientific Advisor (PSA) of the Government of India (GoI) is seeking comments on the draft NDTSP. The guidelines for comments are in italics.

Please follow the below guidelines before filling out the form:

1. Please substantiate your detailed comments basis the document priorities

2. If you have specific comments regarding any particular section in the document, kindly provide the section number and then your input.

3. Kindly provide your detailed comments before 11.59 PM IST on 15 September 2023

FOSS United’s draft comments on the policy are listed below. We request the community to share their feedback by 18th September, 2023. Note: We have restricted our comments to the IP section of this policy as that is the most relevant section for the FOSS community.

We sincerely thank the Office of the Principal Scientific Advisor (PSA) of the Government of India (GoI) for initiating this timely discussion on enabling and supporting the Deep-Tech Startup Ecosystem in India. Additionally, we would like to thank the Office of the PSA for allowing ample time for providing comments to the draft policy.

C.2.1 The deep tech startup ecosystem in India lacks specialised support in terms of filing and obtaining patents for frontier technologies. The intellectual property (IP) regime faces several challenges in this regard including: a) Ensuring the novelty of inventions. b) Managing conflicts related to open-source technologies. c) Ensuring the security of trade secrets. d) Conducting comprehensive prior art searches. e) Navigating international IP laws. f) Risk of intellectual property being lost to foreign entities. These challenges obstruct the effective protection and management of IP rights in the Indian deep tech sector which hinders innovation protection thereby leading to trade secret breaches, and limited IP enforcement. A deep tech centred Single Window Platform should be developed to enable:

We found the language of some of these sections to be unusual. For example, subsection a) talks of ensuring the novelty of inventions. In any patenting regime, novelty is one of the three essential qualities for an invention to be granted a patent, along with usefulness, and non-obviousness to a practitioner of the art. Novelty is an intrinsic quality of an invention and it either exists or does not exist. We are unclear as to how one “ensures novelty.” It is possible that we have not understood this point at all. Therefore, we would appreciate clarity on this point, or request that this be dropped altogether.

Subsection b) talks of “Managing conflicts related to open-source technologies.” Again, this is a very unusual language because open-source technologies are the foundations of all startups. A look at their technology stacks would reveal that 70-90 percent of the software used by startups is Free and Open Source Software (FOSS).

FOSS technologies reduce the time-to-market for startups, reduce the burn rate for startups, and ensure that innovative products can be built that are financially viable. The deep tech startup ecosystem especially is heavily reliant upon FOSS technologies. Annexure I(b) highlights the following subsectors within the larger deep tech space

  • Enterprise Software (including Cloud, Enterprise Mobility etc.)
  • Artificial Intelligence (including NLP, ML etc.)
  • Security Solutions (including Cyber Security)
  • Analytics (including Big Data, Data Science etc.)
  • Robotics (including Robotics Technology & Applications)
  • Computer Vision

In each of these subsectors, innovations are made possible by FOSS technologies. At least two generations of startup founders used OpenCV to build Computer Vision startups. For all practical intents and purposes, the entire Analytics sub sector relies on the FOSS programming language Python and the Scientific Python ecosystem to provide services and solutions. The FOSS technology OpenPGP is the most widely used email encryption standard, which forms one of the pillars of any startup in the Security Solution subsector. FOSS technologies like Docker and Kubernetes heavily contributed to the growth of the Cloud Enterprise Software subsector and practically all deep tech startups in the Artificial Intelligence space are built on FOSS technologies like Keras and Tensorflow.

The tone of this section seems to imply that FOSS is an adversary to deep tech startups, when the truth the exact opposite! We speculate that the intent of this section was to ensure better compliance with open-source licenses. This is essential given the massive amounts of FOSS being used by startups. We therefore suggest that Subsection b) be rephrased to read, “Ensuring compliance with open-source licenses.”

C.2.4 The current National IPR Policy 2016 does not extensively address the eligibility of “computer programs per se, or algorithms” for patent protection. Most deep tech startups in India rely on advanced digital frontier technologies, which is heavily dependent on algorithms and APIs as the primary source of novelty. To ensure Indian deep tech startup competitiveness in the global market, it is imperative to recognise this uniqueness in the National IPR policy. Therefore, a well-defined mechanism should be established for evaluating the suitable amendments that encompass unique characteristics of digital frontier technologies within the scope of patent eligibility.

There are a number of assumptions in the above paragraph that we would like to unpack.

Assumption 1: Granting patents on computer programs and algorithms will be good for Indian deep tech startups

This section advocates for an amendment to Section 3(k) of the Indian Patent Act, which says that, “mathematics, business methods, computer programs per se, and algorithms,” are not patentable subject matter.

Software and algorithms are applied forms of mathematics and logic. In most jurisdictions, mathematics and logic are not granted patents because their unfettered use is essential for human progress. Software is protected by copyrights and trade secrets, which are better suited to protecting abstract ideas like software. Researchers like Bessen and Meurer, authors of the book, Patent Failure, have pointed out that patents do not serve as a vehicle for protecting software because it is hard to draw boundaries around abstract ideas. This lack of boundaries results in the fact that software is the most litigated category of patents.

Software patents are also diametrically opposite to the four freedoms of Free and Open Source Software (FOSS):

  • The freedom to run the program as you wish, for any purpose.
  • The freedom to study how the program works and change it so that it does your computing as you wish.
  • The freedom to redistribute copies so you can help others.
  • The freedom to distribute copies of your modified versions to others – giving the whole community a chance to benefit from your changes.

While FOSS creates public goods that benefit everyone, software patents are state sanctioned monopolies that benefit a narrow elite, and hinder developer freedoms in the process. The Indian government, industry and startups have benefited immensely from the public goods of FOSS. India’s largest e-government projects are all based on FOSS. Therefore, we believe that allowing software patents in India would hinder key government initiatives like Digital India, Startup India and Atmanirbhar Bharat.

Actual impacts:

Privileging private monopolies over public goods: Granting monopoly status to something that should not be granted a monopoly will be extremely detrimental to India’s Deep Tech startups in the long run.

The term “patent” itself originates from the Latin word “patere,” which means, “to lay open.” In the modern era, patents became a state granted monopoly over an invention for a limited period of time, in return for disclosure of the invention. The aim was that inventors would get a defined period of time in which they can exploit their invention. Upon expiry of the patent term, the invention flows into the commons and all of society can benefit from it. We can draw a few conclusions from the above.

Firstly, our lawmakers have clearly said that software is per se (or intrinsically) not patentable. Amongst all patentable subject matter, our lawmakers correctly kept software out of the ambit of patentability through Section 3(k) of the Indian Patents Act. Secondly, only software has a large and flourishing open source community. Most innovation in software is happening in the collaborative world of open source where people voluntarily open up the source code. In other words, innovation is happening in the open, and every important emerging technology - from AI, ML, IoT, cloud computing and many others - is being built in the commons. Thanks to FOSS, billions of dollars worth of technologies are available to Indians free of cost. For an emerging economy, FOSS is a resource to be cherished and protected at all costs.

The Linux Foundation, a leading FOSS organization that hosts 300+ FOSS projects, estimates that these projects have created 1.15 billion lines of code worth $54 billion. Similarly, the Apache Software Foundation estimates that the 350+ projects it hosts have created FOSS worth $22 billion. These projects cover the most fundamental technologies from cloud computing, distributed computing, big data and analytics, blockchain technologies and many others.

While open source software is a public good, freely available to all, software patents are a private good available only to an elite few who have the time and money required to file patents. India has benefited hugely from open source. Therefore, for pragmatic as well as principled reasons, we must keep software out of the realm of patentable subject matter.

Negative impact on early stage startups: The ability to program freely is important for innovation. When developers write software, they combine countless methods, techniques and algorithms into one software program, just as musicians combine hundreds of notes to make a song. If notes of music were made patentable, no musician would be able to write music without first hiring a lawyer, vetting the song and then releasing it into the world. This will have a chilling effect on musical creativity. The same is true for programming which is a creative endeavour.

Software today is rarely built in isolation, it is invariably built with tools, packages, code modules and standards that were developed earlier. If patents are granted on programming methods, no one will be able to write software without violating someone else’s patents. Depending on the facts of the case, the costs of fighting a patent lawsuit can range from Rs 1 crore to Rs 3 crore. This can destroy early stage startups who have raised a few crores in their seed/angel rounds. MNCs and large, mature enterprises are the ones who have the time, money and the legal firepower required to file for patents. A common strategy they follow is to create patent thickets, by filing numerous patents in a specific area. These thickets ensure that no other person can enter that area without violating their patents.

Often, it is argued that software patents will help smaller companies defend themselves against larger firms. However, this argument falls apart quickly when you consider the costs of pursuing these matters in the courts. Whether pursuing or defending a software patents case, smaller firms rarely have the time, money and legal firepower required to win these cases.

Negative impact on users: It is commonly assumed that software patent litigation is an issue that affects only the IT sector. Unfortunately, that assumption is wrong. In the US, Innovatio IP Ventures LLC, which owns patents on connecting to a Wi-Fi network, sued department stores, restaurants and coffee shop chains for offering Wi-Fi to their customers. It also sued individual branches of the largest hotel chains in the United States, including Marriott, Hyatt, Wyndham and Best Western. Innovatio sought payment ranging from $2,300 to $5,000 per infringement, hoping that most defendants will cough up small payments to avoid costly litigation. Technically, Innovatio could also sue residential Wi-Fi users as well, because a patent grant gives the patent owner the right to exclude others from using that invention. While larger companies will be able to fight such cases, it is smaller organizations, individuals, and mom-and-pop stores that will settle with patent trolls like Innovatio because they cannot spend a lot of money on patent litigation.

In 2011, many app developers in Europe stopped listing their apps on American app stores fearing patent lawsuits. A company called Lodsys went after independent app developers like PCalc and Mix & Mash. These apps used Apple’s in-house “Upgrade Now” button that allows them to purchase a premium version from the Apple App Store. Lodys claimed that the “Upgrade Now” button is actually in violation of their patent on “Methods and systems for gathering information from units of a commodity across a network.” Fortunately, this patent expired in 2012. Such litigation threatens app developers, and reduces the choice of apps for end users.

Increasing litigation, NOT innovation: The theory is that software patents will be great for innovation. However, anyone who takes the trouble of reading one or two software patents will quickly realize that software patents are a recipe for litigation. Upon reading a software patent, no software developer or their lawyer can clearly determine whether they are in violation of that software patent or not because abstract ideas do not work well as “property.” Contrast this with land records that clearly state the dimensions of the property, and state the properties that are adjacent to it in the North, South, East and West. This makes it easier to prove whether there has been an encroachment or not. Some developers call software patents a “welfare scheme for lawyers,” considering the amount of litigation that will inevitably follow this state granted monopoly.

In their book, Innovation and Its Discontents: How Our Broken Patent System is Endangering Innovation and Progress, and What to Do About It, authors, Adam B Jaffe and Josh Lerner explain how changes in the processes of the US Patents and Trademarks office resulted in strengthening the legal value of patents while reducing the rigor of patent examination has damaged the system.

“An unforeseen outcome has been an alarming growth in legal wrangling over patents. More worrisome still, the risk of being sued, and demands by patent holders for royalty payments to avoid being sued, are seen increasingly as major costs of bringing new products and processes to market. Thus the patent system intended to foster and protect innovation is generating waste and uncertainty that hinders and threatens the innovative process.”

Do we really want to import such a system into India?

Assumption 2: Copying ideas from the “Mature IP ecosystems of the west” is a great idea

Casting a critical eye on what the “mature IP ecosystems of the west” did before they became mature can be extremely instructive. In his book Trade Secrets: Intellectual Piracy and the Origins of American Industrial Power, Doron S. Ben-Atar, Professor of History at Fordham University in New York City, takes a look at the American understanding of intellectual property during the crucial time of national formation. He also explains how the United States emerged as the world’s industrial leader by appropriating

mechanical and scientific innovations from Europe.

Explaining the formative stages, Prof. Ben-Atar says, “The United States enacted a patent law in 1790 that restricted patent protection exclusively to original inventors and ruled that prior use anywhere automatically invalidated a patent. Alas, this principled commitment to absolute intellectual property had little to do with reality. Smuggling technology from Europe and claiming the privileges of invention was quite common and most of the political and intellectual elite of the revolutionary and early national generation were directly or indirectly involved in technology piracy. And they were following in the footsteps of their ancestors. Americans had welcomed such practices since the early days of European colonization. The American nickname ‘‘Yankee’’ originated in the Dutch word for ‘‘smuggler,’’ and suggests that violation of European economic restrictions had been second nature to the colonists from the early days of settlement.”

The economic history of countries around the world clearly indicates that countries prefer weak IP laws and weaker implementations of laws in the early stages of their economic development. In recent times, China has been an exemplar of this approach, which it follows till date. Once countries have attained a level of maturity by getting close to the apex of the value chain, they begin advocating for adoption of strong IP laws in other countries. With a billion plus population and a growing middle class, India is obviously an attractive market for developed countries. Many of these countries have lost their manufacturing prowess, and are using patents as a powerful tool to control global value chains. The balance of power in software patents is held by organizations in the developed economies who have acquired thousands of patents on basic functions like machine-to-machine communications, e-commerce procedures, cryptography, and many other fundamental computing techniques.

With their deep technical expertise and legal firepower, we might soon have MNCs file a flood of patent applications in the Indian Patent Office. Like Indian states that signed treaties with the East India Company, Indian software developers will have to sign treaties and pay royalties to patent holders to do even the most basic computing functions that they implement freely today. India has a vibrant software ecosystem, almost 100,000 startups that include 100+ unicorns and is a pioneer in Digital Public Goods that address the challenges of attaining the Sustainable Development Goals. All the hard work and success we have gained in IT over the last two decades could be jeopardized if we allow software patenting. Therefore, we must learn from our history of colonization, and protect our interests by rejecting software patents in India.

“Mature IP ecosystems,” can also make mistakes, as the US did by allowing software patents. Some of the fundamental innovations of the computer age – from compilers to spreadsheets to the Internet – were designed before software patenting took off in the 90s. In their book, Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innovators at Risk, authors, James Bessen and Michael Meurer of Boston University dedicate an entire chapter (Chapter 9) to abstract patents and software. The authors found that 25 percent of all patent litigation is around software patents, and one of the main reasons for this is that software is an abstract technology and, since the 18th century, patent systems have had difficulty dealing with abstract technologies.

Using the Karmarkar patent on linear programming (US patent number 4,744,028) as a case study, the authors state that, “The abstractness of the patented algorithm means that these determinations cannot be made with certainty. Patent law assumes that two technologies can be unambiguously determined to be equivalent or distinct; this sets the patent boundaries. Yet for software, this assumption simply does not hold. Although this assumption works for most other technologies, it distinctly does not – or does so insufficiently well – for software algorithms. And if computer scientists cannot make these determinations with any certainty, how can we expect judges and juries to do so?”

These mature IP ecosystems and other IT ecosystems around the world (including India), have also failed to factor in the growing importance of FOSS in patent discussions. The vast majority of software today is built using the Collaborative Innovation model of FOSS, which has collaboration, community and the shared ownership of knowledge as its three pillars. Nowadays, proprietary software is a thin layer built on top of FOSS. When the trend of software patenting picked up in the 80s, proprietary software was the norm and FOSS was the exception. Today, the reverse is true.

By definition, all FOSS are built in the commons, and this has spurred tremendous innovation. Our modern digital world would be unimaginable without FOSS. As public goods that are non-rivalrous and non-excludable, FOSS represents a wealth of technologies that can be used, modified, improved and shared by anyone for free. From stock exchanges to social media to search engines to supercomputers to the Mars Rover to airline scheduling to pandemic predictions, FOSS touches every aspect of our lives.

National Relevance of FOSS

For any country, maintaining control of its critical infrastructure is essential for protecting the interests of its citizens. India has consciously built its identity, taxation, payments and other systems on the foundations of FOSS. This ensures that we are in full control of these citizen services. Compared to proprietary software, the availability of source code in FOSS helps ringfence us from potential denial-of-technology regimes. It also helps avoid vendor lock-in, which leads to higher prices and monopolies. Since these critical projects are built on FOSS, the project implementation teams have the freedom to audit the code to ensure that the software does not have back doors for surveillance, or other malicious intentions.

The Collaborative Innovation model of FOSS also ensures that India has a say in innovation and technology leadership. For example, RISC-V microprocessors, which are based on FOSS software, are expected to see significant growth in the next few years. With India’s electronics import bill at $54.33 billion, second only to its oil import bill at $82.35 billion in FY 2021, we need to produce our own microprocessors to reduce the outflow of foreign exchange. Today, India imports even the basic chipsets that go into simple devices like power meters. An initiative to produce a range of FOSS based RISC-V chips in India can help with import substitution for devices ranging from basic electronics to sophisticated defense equipment. Thus, India can leverage its domestic capabilities in semiconductor design, FOSS based RISC V technologies and its large domestic market to further the goal of building a self-reliant economy.

In emerging technologies like 6G, quantum computing and others, FOSS enables India to participate in the development of the technologies, as well as benefit from the Collaborative Innovation model. This would be in contrast to earlier generations of technologies like GSM, 4G and CDMA, where India was merely a consumer and licensor of these technologies; an approach that involves heavy foreign exchange outflows, and dependence on international vendors.

In sum, FOSS is politically relevant to India as it enables us to contribute to, and benefit from open technology frameworks. What is unique about FOSS is that it enables us to leverage the best of global technologies while reducing dependence on global supply chains that have been disrupted recently due to Covid-19 and other factors. This enables us to build technology leadership, and lower the barriers to serving our large domestic markets.


The terms “Free Software” and “Open Source” are used by people in an interchangeable manner. However, there are important philosophical differences between the two. While the Free Software community considers software that is not free to run, study, modify and share, as non-free and unethical. The open source software community takes a more practical and business friendly approach to software licensing. The term “Free and Open Source Software” or FOSS is often used as an inclusive term that respects both communities. It must be emphasized that the term “free” here refers to freedoms to use the software and not the price of software.



This comment was submitted to the office of the Principal Scientific Advisor today (19th September, 2023).