The cost of space launch has plummeted. At an average of $3,868 per kilogram to orbit, it is at an all-time low. In all likelihood, it will continue to fall.1 Correspondingly, the demand for space launch has never been higher. Orbital launches nearly quadrupled in a decade, from the mid-eighties in 2015 to well over three hundred in 2025. Annual launches now exceed anything seen even during the height of the Cold War.2 And most of this growth comes from privately developed rockets.

Counterintuitively, satellite firms are struggling to find launches.3 The record-low price and the closed launch manifest come from the same source—SpaceX. Three-quarters of the world's payload mass now flies on SpaceX Falcon 9 rockets. The company is now beginning to wind down Falcon as it prepares to transition to Starship.4 Customers report difficulty buying a Falcon launch after 2028.5

The crunch indicates a scarcity of private launch vehicles. The study that produced the $3,868 figure found a structural break in 2010 and attributed it to Falcon 9 entering service on a NASA contract.6 The fall in launch costs can be traced back to state-backed anchor demand. Launch capacity, while it is being developed, is contingent on a set of state decisions. Market forces can only operate atop these decisions:

1. The state issues anchor contracts

2. The state grants access to launch sites

3. The state tolerates failure

4. The state assigns orbital rights

These four decisions operate at different points in the market but converge on the same outcome: launch capacity. Contracts create the demand; launch sites create usable supply; tolerance for failure lowers risk; and orbital rights precede space assets that justify frequent launches. They form the major pillars of launch capacity. Private firms cannot make these decisions for themselves and so private investment and competition respond to those incentives.

When these conditions are in place, private firms can compete over execution. Meanwhile, these four allocations stay in state hands. This edition demonstrates how that plays out.

Every launch company whose privately developed rockets first reached orbit since 20087 is in the table below, the corpses included.8

The state issues anchor contracts

Emerging launch firms mostly sell to one customer—the state. Five customers accounted for 86% of Firefly’s company-wide revenue in 2025, all of it from the US. Its disclosed customers include NASA, the Space Force, the Space Development Agency, the National Reconnaissance Office, along with three defence primes.9

A similar pattern can be observed in China. LandSpace, the country’s leading private launch firm, holds contracts to fly Guowang and Qianfan satellites.10 Guowang is a megaconstellation run by a central state-owned enterprise. Qianfan, another megaconstellation, is backed by the Shanghai municipal government. LandSpace’s Zhuque-3 rocket is also slated to carry the Haolong cargo shuttle to Tiangong, the state-owned space station.11 On 19 August, that rocket became the first Chinese commercial vehicle to land its booster after an orbital launch.12 Most of its contracts are tied to state entities or state-backed programmes.

Even the commercial customers are often the state in disguise. Private satellite firms that buy launches are contracted to sell space-based services to governments. In the US, 82% of satellite manufacturing revenue came from state contracts.13 Many startups position themselves as civil and defence contractors from the outset, or pivot to it eventually.

Emerging launch firms often depend on the state to provide an initial base of demand through anchor contracts. The US is the obvious case study. The Pentagon’s FY2027 budget request provides about $5 billion for 31 launches.14 The US Space Force has also raised the ceiling for National Security Space Launch Phase 3 Lane 1 from $5.6 billion to $17 billion, as the number of expected launches under the contract grew from about 60 to 170 over the next decade.15

Meanwhile, Europe moved later to cultivate a new generation of private launch firms. None yet have proven orbital vehicles. The first will probably be Isar Aerospace’s Spectrum. It is on a €197.8 million milestone-based ESA contract under the European Launcher Challenge.16

States also have a strategic reason to spread launch demand across more than one provider: assured access. National Security Space Launch buys from multiple launch firms so that technical failure or capacity constraints in one firm will not shut off an entire class of payload or orbit. The US National Space Transportation Policy also directs procurement to support a launch industry and maintain multiple avenues for government payloads.17 The surplus also opens an option for surge capacity during wartime.18

During peacetime, this surplus can be sold abroad. These exports finance reserve capacity and give domestic launch firms a higher launch cadence. They can also make foreign space powers structurally dependent on the exporting state’s launch industry. Reliable and inexpensive imports can weaken the economic and political will to maintain a more expensive indigenous space launch capability. When relations sour or interests diverge, foreign launch dependence can stop being benign. The American experience with Russian RD-180 engines after Russia’s annexation of Crimea, and Europe’s sudden loss of Soyuz access after the beginning of the Russia–Ukraine war in 2022, show how an ostensibly commercial relationship can become an assured access problem.19

The state grants access to launch sites

Launch capacity is also fenced-in by access to launch infrastructure. Along with launch vehicles, launch sites also need to grow. In the US, Wallops Flight Facility and Kennedy Space Centre have seen 467% and 252% increases in launches between 2020 and 2025. They could hit their ceilings by 2028–2029.20 Even Cape Canaveral is nearly out of pads after the recent administrative allocations by the Space Force.21 With Vandenberg also running out of room, the country needs a third heavy-lift site.22 The Space Force Chief of Space Operations has warned that they may have to cap launches.23 Range capacity is also constrained by airspace, sea clearance, tracking and common-use infrastructure. New pads alone cannot solve the problem.

A single launch site can become a single point of failure. Blue Origin’s New Glenn rocket exploded during a hot-fire test in May, destroying the rocket and damaging its only orbital-class launch pad.24 I don’t think the company can return to flight before 2027.25 Its first lunar lander mission has already slipped to early 2027, and its second-stage testing has moved onto a NASA stand.26 The delays may also affect American lunar timelines.

Geography has a strong bearing on which orbits a launch site can reach efficiently and safely. Cape Canaveral is not an ideal launch location for southerly polar trajectories. Spent stages can overfly inhabited areas. Floridan launches to polar orbits need constrained corridors or dogleg manoeuvres. Vandenberg is better suited for it.27 Launches from Māhia, New Zealand, on the other hand, enjoy open oceans to the north and the south. Similarly, launches from equatorial latitudes can reach low-inclination orbits more efficiently. Launches from polar latitudes like Esrange are better for polar and sun-synchronous orbits.

Sea launch addresses some of the geographical limitations of fixed pads. A vessel can sail to a desired location and the rocket can launch from it. The Chinese firm Galactic Energy flew the first sea launch of a privately developed rocket in 2023.28 Its Ceres-1S rocket launched from a mobile platform in Haiyang in Shandong. Galactic Energy also flies from Jiuquan, so the sea platform is an alternate launch option when launch sites on land are unavailable. OrienSpace, another Chinese launch firm, followed this up with a sea launch of its Gravity-1 rocket in the Yellow Sea. It flew again this year, from the East China Sea, through some of the busiest shipping lanes in the world.29

These sea launches depend heavily on state assets and support. Galactic Energy’s launch was organised and commanded by the sea launch team of the Taiyuan Launch Centre. Haiyang has operated as a remote base of Taiyuan since 2019, and a second maritime facility is planned near Yangjiang in Guangdong.30 The vessels, the piers, the processing halls and the command structure are state assets made available to private firms.

States can also create geographic redundancy by enabling access to foreign launch sites. Washington and Wellington concluded a Technology Safeguards Agreement (TSA) that clears the path for American launch technology exports to New Zealand.31 Using it, Rocket Lab launches from Māhia, which hosted more orbital launches last year than the home sites of any country except the United States and China.32 A similar agreement was signed with Sweden, which allows Firefly to access the Esrange launch site and even license its rockets out to Sweden.33

Through foreign launch sites, the launch firms gain more than just access. Sweden has committed roughly a billion kronor to military space capability through 2032 and designated Firefly as the primary launch vehicle provider.34 Rocket Lab’s launches from Māhia have also carried Japanese, South Korean and European payloads.35

Access to launch sites, primary and alternate, is allocated by states through treaties and state support. Alternate sites provide the respective state with the ability to retain space access when it is not able to do so from its home base due to war, adverse weather, damaged pads or any other reason. The pads at Cape Canaveral are administratively allocated to certain launch firms. They can be revoked.36 The maritime spaceport at Haiyang was state-built, and the private sea launches were state-commanded. Esrange was built, funded and anchored by Sweden. Māhia needed a treaty and an act by the New Zealand Parliament.37 While the launch vehicles are privately developed, their launch access is still very much authorised by states.

The state tolerates failure

Developing a launch vehicle takes many years. SpaceX took six years to reach orbit, Rocket Lab took twelve, Firefly—eight, LandSpace—eight and Blue Origin—twenty-five.38 Announced timelines rarely stick. Delays are the norm rather than the exception. Throughout development, launch firms can go years without generating launch revenue. In the meantime, the capital requirements are exorbitant: firms must build and maintain factories, test stands, engines and a workforce for over a decade.

Among the companies in this table, first flight attempts failed about as often as they succeeded. SpaceX’s Falcon 1 failed three times before reaching orbit, Rocket Lab’s Electron failed its first flight, as did Firefly’s Alpha, LandSpace’s Zhuque-2 and Virgin Orbit’s LauncherOne. Galactic Energy, OrienSpace, iSpace, Blue Origin and Skyroot succeeded on their first try.39

Launch firms don’t graduate out of failures even after a streak of successes. Despite its track record, SpaceX’s Falcon 9 suffered a failure in its second stage in 2024.40 Rocket Lab lost a payload on its forty-first flight.41 Two days later, Galactic Energy’s Ceres-1 lost its payload on its tenth flight.42 Reliability minimises the rate of failure rather than eliminating it altogether.

The costs of failures are high. In the New Glenn booster explosion, Blue Origin lost its transporter-erector and lightning towers, and suffered damage to its lone orbital-class site—the one it had spent a decade building.43 Three days before the explosion, NASA awarded Blue Origin two contracts worth $188 million to deliver rovers to the lunar South Pole.44 New Glenn is also central to Blue Origin’s role in Artemis.45

States are most willing to carry launch firms through failure when they value the capability the firm is expected to preserve or create. After the explosion, NASA began examining alternative launch options in case New Glenn could not meet the Artemis III timeline.46 But NASA did not remove Blue Origin from the Artemis Programme. The Space Force had awarded Blue Origin a National Reconnaissance Office launch task order only hours before the anomaly. After the explosion, it publicly reaffirmed that the Space Force and NRO remained committed to working with Blue Origin.47 Japan’s H3 rocket has failed twice and lost expensive payloads. The government nevertheless continued the programme, and H3 returned to flight successfully in June 2026.48 LandSpace raised money from the state-backed China SME Development Fund and a Huzhou government fund in 2020. Zhejiang and Jiaxing later backed the construction of its production base. LandSpace’s Zhuque-2 suffered a failure in 2022 and reached orbit in 2023. A state-owned advanced-manufacturing fund invested another 900 million yuan in 2024.49

Where governments saw less strategic need to preserve a capability, they appear to have been less willing to absorb failure. Astra reached orbit in 2021 with a Space Force payload, but went on to fail twice in 2022 with NASA satellites. NASA moved its contracts to Rocket Lab.50 Astra retired its rocket.51 Virgin Orbit had proven air-launch capability, in which rockets are launched from airplanes. It held Space Force contracts for three Space Test Program launches worth $35 million, one of which it had fulfilled.52 It had spent roughly $45 million in the third quarter of 2022 alone.53 The UK government sponsored its flight from Cornwall, UK, which failed. The US had Rocket Lab flying the same class of payloads from two continents. The US also had air launch capability from the Pegasus rocket. Neither country stepped in to preserve Virgin Orbit before it declared bankruptcy. I attribute these decisions, or rather the lack of any, to the fact that both states already had alternatives for the capabilities Virgin Orbit brought to the table.54

The state's tolerance for failure is expressed through political will. Launch firms are judged by the capability they bring and how much the state needs it. The judgement is expressed through regulatory and procurement-related decisions.55 These decisions send signals that private investors and launch customers rarely miss.

The state assigns orbital rights

A launch industry cannot necessarily sustain itself on launch revenues alone. Even SpaceX’s Space segment—which contains its launch business and Starship development—recorded a $657 million operating loss on $4.086 billion in revenue in 2025. The operating losses widened to $1.204 billion in the first half of 2026.56 Some of the losses are from Starship development costs. Taken alone, Falcon 9 is probably profitable by now. The study that put the average space launch at $3,868 per kilogram measured only the flight and did not factor in the development spending. But I see development as a perpetual cost that cannot be overlooked. Just as India needs the next-generation launch vehicle, SpaceX needs Starship, Rocket Lab needs Neutron, Firefly needs Eclipse. Northrop attempted to substitute away the R&D for the Antares rocket. It used imported engines but the Russian supply stopped in 2022. It is one of the dead rockets in the table. To revive it, Northrop is paying Firefly for the R&D.57

The profitability problem spans the rest of the industry. Firefly earned $159.9 million in 2025 and lost $298.3 million.58 Rocket Lab also remains loss-making, with a $198.2 million net loss in 2025.59 LandSpace’s 2025 revenue was 52 million yuan against a net loss of 1.711 billion yuan.60 At the end of 2025, China even changed the listing rules on the Shanghai Stock Exchange. Reusable-rocket companies were excused from profitability and minimum revenue conditions.61

Rather than relying entirely on external customers, launch companies can use their capacity to build downstream space assets. Starlink is the obvious example. SpaceX launched 78 times in the first half of 2026. Seventeen of those were for customers. The remaining 61 launches were for itself. Out of the 1,041 tons delivered to orbit, 908 tons were its own payloads.62 In the same period, Starlink services generated $7,548 million in revenue and made $2,844 million in operating income.

This vertical integration unleashes new models. With reusable rockets, SpaceX has been able to grow and maintain its Starlink megaconstellation at more than 11,000 satellites.63 Being its own customer also changes the perception of risk. An external customer may not seek launches on an unproven Starship rocket but Starlink will.

Megaconstellations could not have been devised with cheap launches alone. Access to massive amounts of spectrum and orbital resources is essential. These are mediated through states and the International Telecommunication Union (ITU) system. Radio frequencies and associated orbits are limited natural resources, to be used so that all countries have equitable access.64 Filings for these systems are submitted by national administrations on behalf of satellite operators, whether public or private. For non-planned satellite services, earlier filings get priority over later ones. The system therefore rewards states that file early and actually have the launch capacity to bring those filings into use.

It is no surprise that the states making the largest filings recently are the US and China. In addition to its authorisation for 15,000 satellites for Starlink, SpaceX has also applied for up to 100,000 Gen3 Starlink satellites and one million satellites in its proposed orbital data centre system.65 Amazon Leo (formerly Kuiper) has filed for close to 12,800 satellites. Around 390 of those are in orbit. Amazon and Blue Origin share a founder. Amazon already has contracts with Blue Origin for twelve New Glenn launches, with options for fifteen more.66 On its own, Blue Origin has filed for 5,408 satellites across low and medium Earth orbit for its TeraWave constellation, and for an orbital data-centre constellation of up to 51,600 satellites.67 New Glenn will have plenty of internal demand when it returns to flight. In late 2025, China filed for more than 2,00,000 satellites across roughly 14 constellations.68 Beijing was meanwhile publicly complaining that Starlink was crowding the orbits.69 LandSpace has a 48 per cent stake in Hongqing Technology. Hongqing has filed for the Honghu-3 constellation consisting of 10,000 satellites.70 Rocket Lab is buying the satellite communications company—Iridium—for roughly $8 billion. Through this, Rocket Lab gets 66 satellites, a positioning business and L-band spectrum. Iridium earned $871 million last year,71 which is actually more than Rocket Lab’s revenue.

Services like Starlink, Amazon Leo, Guowang and Qianfan could not have been devised without state licensing and orbital filings. And in the Chinese cases—even direct state backing. The Federal Communications Commission (FCC) recently approved the transfer of sixty-five megahertz of nationwide spectrum licences from EchoStar to SpaceX through the secondary market without an auction.72 Rocket Lab's purchase of Iridium also relies on FCC approving the licence transfers. The state has a hand in all these businesses.73

LandSpace is a useful case to see all four state decisions in action. Its Zhuque-3 flight on 19 August launched from Jiuquan—a state range. The company has assured demand for its rockets from China’s state-backed megaconstellations. State funds and two municipal governments supported the company through its developmental failures. Its payload was Honghu-03, built by Hongqing Technology, in which LandSpace holds a 48 per cent stake. China has filed with the ITU for 10,000 satellites for the constellation.74

As the strategic returns from orbit-dominating megaconstellations grow, these firms stop behaving like launch service providers. SpaceX has begun turning away some commercial customers seeking Falcon launch services beyond 2028. It is also winding down some expendable Falcon hardware as it transitions to Starship. Falcon rockets will prioritise NASA and national security requirements.75 Foreign customers including India cannot treat surplus American launch capacity as permanently available.

India and the four state decisions

The Indian state has yet to make progress on all four counts:

1. India has not publicly issued anchor launch contracts

2. India has one operational launch range with no other alternatives

3. India has a latent but untested tolerance for failure

4. India has early plans for megaconstellations without the launch capacity to deploy them domestically

1. India has not publicly issued anchor launch contracts

The state has supported Indian launch firms in many ways but not as a customer. Skyroot’s Vikram-1 became India’s first private rocket to reach orbit. It launched from Sriharikota using extensive ISRO support.76 Meanwhile, no anchor launch contracts have been publicly announced. Weeks after the flight, the government announced a scheme to subsidise space launches on domestic rockets.77 This policy move, along with the foreign investment liberalisation and space venture fund,78 betrays an assumption that a ready market already exists. These measures seek to subsidise the development costs for the launch industry and launch-purchase costs for satellite firms. But the space sector has not matured yet. India has a commercial launch market, but not yet one large or dependable enough to sustain a private launch industry. In the comparable cases above, state procurement created the anchor demand on which private demand could accumulate. The foreign entities seeking launches have more competitive launch services elsewhere. Sustained demand for launch needs anchor contracts from the state.

Skyroot’s orbital flight arrives at a time when India’s public launch capability is at a trough.79 Two PSLV mishaps, low launch counts, constrained launch sites and the heavy-lift vehicle scheduled to launch only in the next decade have created a severe launch bottleneck.80 It is likely that the national space programmes will face delays due to the launch bottleneck. Jatan Mehta’s calculation shows that India needs roughly 128 launches by 2030 to meet its national goals.81 That will require an average of twenty-five launches a year. Against that stands the sobering fact that India has had an average of five annual launches across the past decade, never more than seven in a year.82 That gap is the demand that India should have turned into milestone-based capacity development and anchor contracts.

In developing the launch industry, the state has focused on outsourcing the manufacturing of ISRO-designed vehicles. SSLV technology has been transferred to Hindustan Aeronautics Limited (HAL)—a state-owned enterprise.83 NewSpace India Limited has contracted a HAL-L&T consortium to produce five PSLVs, while a separate plan is in place to industrialise LVM3.84 While these arrangements expand industrial participation, they do not create competing launch vehicles. The industry is producing these vehicles based on ISRO designs. That weakens the demand signal for private investment in alternative launch vehicles since the launch firms have to compete with the government.

China made the same mistake before India. Blaine Curcio told the US-China Commission in April 2025 that Beijing had subsidised the supply of rockets instead of the demand for launches—there was plenty of funding, land, staff and facilities, but almost no purchase orders.85 It has since begun to course-correct. Its constellation programmes and space station missions have now become a steady source of demand for the launch industry.86 China’s privately developed launch vehicles have different builders, different propellants, different launch modes including sea launch.

The status quo in India, on the other hand, will narrow down its rocket fleet. The PSLV and GSLV share a first stage. The Vikas engine is used in both and in the LVM3 core. The SSLV’s second-stage motor is the same HPS3 motor used as PSLV’s third stage, which suffered anomalies on both C61 and C62.87 The concentration creates an assured-access vulnerability. India will have single points of failure in its access to orbit. For key strategic payloads, Indian policy should aim to maintain at least two independently designed and certified launch vehicles. At least wherever mission volumes are big enough to sustain them.

Steady demand from milestone-based funding and anchor contracts will pull the industry up the capability curve. Heavier lift and reusability will see progress. NASA awarded SpaceX its Commercial Orbital Transportation Services (COTS) agreement in August 2006. At the time, Falcon 1 hadn’t succeeded and had a lift capacity of 420 kg. This is in the same broad class as Vikram-1. COTS shared the development risk of Falcon 9; subsequent NASA contracts bought Falcon launches.88 All this eventually produced the Falcon 9.

The Indian space budgets are too thin for anchor contracts at a meaningful scale. The Department of Space received $1.64 billion (₹13,705 crore) in 2026–27.89 Less than a third of the $5 billion that the Pentagon alone has set aside for thirty-one national security launches in 2027.90

The headline budget is not the only binding constraint. Since 2020, the Department of Space has failed to spend even its revised budgetary allocations. It has surrendered over ₹12,000 crore, at a rate higher than Defence and Road Transport. That is close to a full year of funding.91

There are two Indian programmes the state could readily turn into anchor demand. The Earth Observation public-private partnership will produce a twelve-satellite constellation that will launch on Indian vehicles.92 Those launch orders should prioritise Indian private firms. The Space Based Surveillance programme is the larger prize, with fifty-two satellites.93 Reports indicate that SpaceX is being considered to launch the first satellites.94 If Indian strategic surveillance payloads launch on foreign launch vehicles, it would be a missed opportunity to turn national security demand into domestic launch capacity.

2. India has one operational launch range with no other alternatives

Like its launch vehicles, India’s launch sites are also concentrated. India has just one operational launch range at a geographical location that is a hurdle to its small rockets. Sriharikota sits on the east coast with Sri Lanka directly to the south, so polar launches fly southeast and then turn. This dogleg costs precious propellant and payload capacity. Kulasekarapattinam, at the southern tip, is being built to remove this hurdle.95 It is under construction and will likely support launches from 2027. Until then, the SSLV and Skyroot’s Vikram-1 share the First Launch Pad at Sriharikota with PSLV. Which probably explains Skyroot’s announcement to build its own launch pad.96 That will take years but it’s a start. Agnikul built its own pad and mission control at Sriharikota itself.97 It has not reached orbit yet. A third pad has been sanctioned at Sriharikota for the heavy-lift vehicle.98 All of them are on one island under one weather system.

IN-SPACe has invited private firms to operate the complex.99 It should also consider allowing foreign launches from the spaceport while domestic launch capability and infrastructure develop. The ideal of meeting all of the country’s launch requirements using Indian rockets on Indian launch pads is already out of reach for at least a decade. The country has skipped to an inferior substitute, which is to buy foreign launch services from foreign soil. It should consider the intermediate options, in sequence, using Technology Safeguards Agreements: launching Indian rockets from foreign sites and allowing foreign rockets to launch from Indian sites. TSAs are well understood in India and around the world. The US has signed TSAs with New Zealand, the UK, Australia and Norway.100 They turned a stretch of coastline in the host countries into launch infrastructure assets. India's own 2009 agreement with Washington lets American payloads fly on Indian rockets.101 It is worth expanding the agreement under the US-India TRUST initiative framework to enable bidirectional commercial launch access.102 India should also consider TSAs for Indian rockets seeking foreign launch sites. India should separately examine maritime launches as an option.

3. India has a latent but untested tolerance for failure

The Indian state may already have a latent tolerance for failure but has not had an opportunity to prove it. There is strong political capital in favour of the space sector. The Department of Space is under the charge of the Prime Minister's Office. The state waited through Agnikul's scrubbed attempts without visible impatience. But no privately developed Indian rocket has failed. The real test will come with the first failure. The state's response will matter more to the Indian launch industry than the entire price-support scheme.

4. India has early plans for megaconstellations without the launch capacity to deploy them domestically

Reliance Jio plans to deploy a 1,600-satellite megaconstellation and has received an IN-SPACe technical nod.103 The Indian administration may file for it soon. Once filed, the constellation would have to meet the ITU's deployment milestones.

The ITU deadlines force a higher launch cadence. The milestones for Chinese megaconstellations coincide with large domestic launch orders for Guowang and Qianfan. LandSpace, Space Pioneer and CAS Space have been awarded contracts to launch satellites for Qianfan. LandSpace also has launch contracts for Guowang.104

India needs to ramp up launch capacity if it wants to avoid foreign dependence. Assume the Indian administration makes the ITU filing in 2027 and that the 1,600-satellite system falls within Resolution 35. That would put the deadline for bringing the system into use around 2034. The subsequent deployment milestones would require roughly 160 satellites by 2036, 800 by 2039 and the full 1,600 by 2041.105 Reports have Jio targeting service within two to three years, which would put its first satellite launches in 2027 or 2028. Unless a domestic order is placed soon, the first batch of satellites is likely to launch on foreign vehicles. India would then be developing that country’s launch industry instead of using that demand to deepen its own. Anchor purchase orders should be issued soon. It is the best way to ramp up domestic launch capacity before the deadlines approach.

This is where the money going into price-support schemes is better spent on procurement. Indian launch services will initially cost more per kilogram until domestic rockets build cadence, increase lift and achieve reusability. Procurement policy should not chase the occasional cheapest launch. Instead, it should target baseline launch capacity through a competitive domestic launch industry. Along with launches, that procurement buys India cadence, learning, investor confidence, production capacity and independent access to space. As those capabilities mature, heavier lift and lower prices will follow.

Conclusion

A baseline of sovereign access to space that India should aim for is the ability to place strategically necessary payloads into the required orbits, at the required time, without depending on an external state’s permission, industrial priorities or spare capacity. This should not be confused with autarky. India can and should buy foreign launch services when doing so serves its interests. But that should not come at the cost of an independent option for when foreign access is unavailable or politically unacceptable.

India’s foreign-launch dependence has spanned commercial and civil missions and may now extend to strategic constellations. A tightening foreign launch market will therefore affect India more than it ought to affect an independent space power.

The domestic capacity required to avoid that dependence can appear wasteful. A second launch vehicle with similar payload capacity, an alternate launch site or unused production capacity may sit idle while a cheaper foreign option remains available. But its value will become apparent when a crisis hits.

India needs to make this pivot early, in peacetime, because foreign dependence cannot be unwound at the onset of a crisis. Payloads are integrated with launch vehicles through their masses and volumes, mechanical and electrical interfaces, mission profiles, environmental loads and certification requirements. Shifting to another rocket will require redesign and testing. It takes time.

The four state decisions that I argue for are not subsidies for a launch industry. Anchor demand creates industrial capacity. Access to primary and alternate launch sites creates geographic resilience. Tolerance for failure preserves or creates valuable capabilities. Orbital rights create downstream assets and launch cadence. For the state, they are the means to necessary ends: peacetime capacity, wartime responsiveness and freedom from external veto.

Companion tool: India Launch Access Stress Test — see which domestic launch options survive when part of the system fails.

A new edition of The Orrery is published every fortnight. Ashwin Prasad Rao is Staff Research Analyst at the Advanced Military Technologies and Outer Space programme at the Takshashila Institution. The Orrery is personally owned and written; the views are his own.

We have all heard of Starlink. It is a megaconstellation of communication satellites in low Earth orbit, networked to provide internet connectivity to users on the surface. With little reliance on ground infrastructure, this technology shines where terrestrial broadband and terrestrial phone networks fail. And in a world that increasingly relies on constant, highly reliable connectivity for critical applications, Starlink offers solid redundancy from the sky that nothing on the ground can match. I have written about how this technology functions, if you want to gain detailed context:

Risks for India

Starlink is about to begin operations in India. With its large continental landmass and varied geography, India is an ideal candidate for benefiting from this technology. At this point, my colleagues and I felt that it was worthwhile highlighting the very real risks that India has to weigh alongside the benefits as it prepares to embrace Starlink.

I continue to be a believer in global markets and technology access. At the same time, I also feel that a country should be able to wield sufficient control over the flow of information in order to maintain a favourable balance of power. There are various aspects to this information control, but internet connectivity is a major pillar. I do not argue for this so as to enable a surveillance state. Political institutions should safeguard individual privacy, and that is a separate problem to solve. I argue for this from a strategic and tactical perspective.

Wholesale adoption of Starlink will allow SpaceX to leverage the dependence in ways that are contrary to the interests of Indians. The risks stem from the sheer scale of the moat that Starlink enjoys. As an individual exercising significant control in a private corporation, Elon Musk’s erratic personality and behaviour add a risky degree of unpredictability. He seems to wield influence over the US government, which also hints at the US government’s influence over him.

The existing mechanisms India has in place to safeguard against the data security risks from Starlink are woefully inadequate since they are difficult to enforce. The core risks persist.

Ultimately, India should continue to improve connectivity infrastructure and diversify satellite internet vendors to reduce overexposure.

Click here to read the full paper.

I co-wrote an opinion piece about this issue for India’s World.

Here is an excerpt:

… Despite its clock failures, Galileo maintained service because of its built-in redundancy with four clocks per satellite and over 24 satellites in orbit. GPS’s early clocks had very low lifespans, and yet the system survived because the US had over 30 satellites to absorb individual losses. This was underpinned by a domestic infrastructure that could buy, manufacture and launch satellites in larger numbers. The takeaway is that a navigation satellite system is a lot more than a cluster of satellites with atomic clocks that ISRO alone has to build and launch. Instead, it should be viewed as a national infrastructure that demands the same industrial depth, redundancy, and operational rigour as a power grid or a rail network…

… India built a strategically important system without building enough redundancy around it. NavIC uses high-altitude satellites. This design permits fewer satellites covering a smaller patch of sky than global-coverage lower-altitude positioning systems like the GPS. On the flip side, each NavIC satellite carries a disproportionate share of the coverage geometry… With its lean configuration, the loss of even a single healthy satellite can be debilitating. It can take the system from working reliably to barely working at all.

… The main challenge is uneven replenishment. Unsuccessful launches in 2017 and 2025 have disrupted the replenishment of the constellation with new satellites. Analysis of publicly available data suggests ISRO is rationing propellant on its remaining satellites to stretch their functioning lives before replacements arrive. Meanwhile, one of the last three functional satellites is past its ten-year design life… Building domestic launch capacity is not separate from fixing NavIC…

… Today, Indian systems can draw on GPS, GLONASS, Galileo, and BeiDou alongside NavIC. Each of those four foreign constellations is operated by a different geopolitical actor—the United States, Russia, the European Union, and China. Their interests diverge enough that in any plausible conflict scenario, if some of these systems are jammed, spoofed, or selectively denied, at least two are likely to remain available to India.

Click here to read the full article on India’s World.

My chapter was about the frugality of Indian space programme that we hear so much about. India launched Chandrayaan at a cost less than the budget of the movie Gravity. Indian mission to Mars was cheaper than the movie Interstellar.

This ability to achieve successful space missions with shoestring budgets started off as an effective political economy tool to the early Indian governments that had to justify spending crores on space missions instead of addressing health and hunger. Now, it has become bragging rights and often a brand identity for all of Indian endeavors in space.

The India that used this frugal strategy is different from the India we live in today. The country is now more prosperous, more ambitious and more capable. Then, the question that we should ask ourselves is: do we want to hold on to this identity in the present and for the future? Does it afford us the chance to be audacious that our ambitions demand?

These are the questions that I tried to answer in this book chapter. Here are glimpses of the arguments I make:

… ISRO’s frugality is not an inherent trait of the organisation, but a mathematical outcome as a result of a narrow objectives, simpler designs and testing through software simulations…

… This worked well for a world where ISRO only needed to launch a handful of rockets each year. Each launch was a national event. Each rocket was like a handmade product. That world is gone. Today, India needs to launch twenty or thirty times a year at the least. It is not there yet and is actively trying to close the gap. ISRO is changing to make this happen…

… The routine manufucturing and operational responsibilities will be transferred to the private sector. By harnessing the industrial scale of the private sector, the hope is that more of the commercial value of these technologies can be realised. Meanwhile, ISRO will only focus on the cutting edge, the projects that will push the boundaries of Indian technological capabilities in space like the new more powerful rocket or taking Indians to space…

… In other words, ISRO’s job now is to push into territory where failure is likely, timelines are uncertain, and the returns are measured in capability rather than cost-per-kilogram…

… For decades, ISRO’s identity, in the public imagination has been built around the word frugal. Frugal engineering. Frugal innovation. Missions cheaper than Hollywood films. But a research organisation that pushes frontiers cannot optimise for frugality. It must optimise for ambition. And after half a century of being rewarded for keeping costs down, it is not clear that ISRO has the institutional instinct or the sociopolitical permission to spend boldly, and when necessary, to fail expensively…

… None of this is to say the frugal era was a mistake. It was one of the great institutional achievements of post-independence India. A country that transported its first sounding rocket on a bicycle built a space programme that reached Mars on its maiden attempt, landed near the lunar south pole before anyone else, and launched satellites for thirty-six countries at prices no one could match. The engineers who did this deserve more than the word frugal. They deserve the word brilliant…

… Today, India is more prosperous, more ambitious and capable. The next decade asks something different. It asks ISRO to build a space station. To put Indians on the Moon. To develop rockets that land themselves and fly again. These are not tasks that reward the cheapest possible approach. They reward the fastest learning cycle, which means the willingness to build, break, rebuild, and break again, all in public, at great expense, with no guarantee of success…

… The word frugal will have to change its meaning if it is to remain useful. It cannot keep meaning cheap. It must come to mean disciplined—an institution that spends carefully not because it fears spending but because it respects what it is spending on. There is a difference between an organisation that avoids waste and an organisation that avoids risk. ISRO has always been the former. It must not become the latter.

You can buy the book on Amazon here.

If you cannot buy the book and wish to read the chapter, here is my draft of the chapter.

From the article:

It has been 53 years since humans went to the Moon under the Apollo programme. … the two are fundamentally different. The Apollo era was driven by Cold War politics and was a strictly American endeavour. Its goal was short-term missions for symbolic and scientific progress. In contrast, Artemis stands on a web of international collaborations and commercial partnerships. It aims to establish a long-term human presence on the Moon.

The Apollo programme was plagued by massive costs. Artemis is trying to overcome this issue by moving away from government-owned rockets in favour of commercially procured reusable rockets and hardware. This is built on the understanding that long-term lunar exploration can only succeed if it’s economically viable.

… The rush to the Moon mirrors the geopolitical competition on Earth between the United States and China. The Chinese Communist Party views space as a ‘critical domain in international strategic competition’. It aims to supplant the US as the world’s leading space power by 2049.

… By 2030, it has plans to land its astronauts, and by 2035, to construct a station on the Moon in partnership with Russia.

… If China establishes a dominant lunar presence first, there is a profound fear in the West that it could use ‘lawfare’ (legal warfare) to enforce “keep out” zones over the most valuable lunar territory. The Artemis programme is a vehicle to carry not just the US and its partners’ infrastructure but also the values and governing preferences outlined in the Artemis Accords, a set of voluntary guidelines for conduct on celestial bodies.

… The Moon is seen as a massive reservoir of untapped wealth. Some craters near its south pole are permanently shadowed from the Sun and contain water ice. This water is being positioned as the ‘oil’ of the lunar economy. Besides human sustenance, it can also be split into liquid hydrogen and liquid oxygen to serve as rocket fuel. If mining this lunar water and turning it into fuel proves viable, the Moon can become a stepping stone to Mars and Deep Space. Besides water ice, the Moon also holds critical minerals indispensable for modern electronics. Also, the lunar soil has an abundance of Helium-3, a potential clean fuel source for future nuclear fusion reactors.

Many of these resources are geographically concentrated in small regions of the Moon... All of this creates a ‘gold rush’ dynamic to be the first to reach and control these strategic locations and orbits.

It is currently impossible to judge if mining lunar ice or critical minerals will ever be economically viable. There is also no clear or strategic military advantage to be gained from activities in cislunar space yet. However, in case the bets do pan out, the countries that manage it first will greatly increase their wealth and technological competitiveness.

Click here to read the full article.

But the story becomes more interesting when you include the budgetary revisions that happen a year after the allocation. The revised budgets are always lower than the allocations.

Adding the actual spending done by DoS/ISRO makes this even clearer. The department cannot seem to spend its budget fully.

In fact, it has failed to utilise even the revised budget every year in the last seven years. Only in 2019–20 did DoS actually spend all of its allocation, likely because of Chandrayaan–2 mission peak costs. Since 2020, the total unspent amount by DoS totals more than ₹12,000 crore—effectively equal to one full year’s expenditure.

Breaking down the budget

To try and understand these trends at a deeper level, we can look at the funding breakdown. Here is the breakdown for this year’s budget:

Let’s explore each segment one-by-one.

INSAT Systems segment

The INSAT Satellite Systems specifically funds communication satellites (like GSATs) that support TV broadcasting, telecommunications, and VSAT connectivity.

This year, with a mere ₹130 crore allocation, it makes for 0.9% of the total space budget. It is the least among the segments. A few years ago, this was not the case.

Even with absolute numbers, the INSAT budgets are falling.

This is an intended trend. The government has been reforming the space sector since 2019. One of the reforms is to unburden ISRO of its routine tasks. The INSAT satellite systems fall within this basket. They are being transferred to NewSpace India Limited (NSIL), which will take over these operations. NSIL functions on a commercial logic. It will continue to fund them through its own internal resources or loans, based on demand. So systems like the INSAT may even grow in time but it will do so without direct taxpayer money.

Space Applications segment

Space Applications consists of making use of the data we get from ISRO satellites already in orbit. It is the downstream space sector. It includes things like remote sensing, disaster management support, agriculture, and urban planning using satellite imagery.

This segment has seen a stable but flat funding of around 13% of the total budget. The services are continuing but there is no big government-led expansion in the application infrastructure.

This segment operates like a service provider. 79% of its spending is revenue expenditure. Revenue expenditures fund the day-to-day operations. We can think of it as the ‘maintenance’ budget. It does not create assets or reduce debt like the capital expenditure does, which is harder.

Most of the space applications budget is used for running the National Remote Sensing Centre (NRSC) and Space Applications Centre (SAC), paying scientists, and maintaining the Bhuvan and Bhoonidhi data portals. The small share of capital allocation is for upgrading data centres and purchasing high-resolution data processing units. As a result, you see fuller utilisation of the allocated budgets every year. The revisions stay nearly the same.

The lack of growth here indicates that the government wants to focus more on other infrastructure creation activities. Private players are expected to take over application development in due course.

Space Sciences segment

Space Sciences is the exploration arm. It funds planetary missions like the Chandrayaan series, Aditya-L1, Venus Orbiter and astronomy. This is a priority sector for ISRO, under the new space policy. ISRO is being unburdened of INSAT and application-related responsibilities so that it can focus on space sciences and other strategic missions.

The government’s intent shows. It is the fastest-growing vertical in the 2026–27 budget. It has seen a 4x increase and occupies around 4% of the total budget. But this is still too low considering this is now a core function of ISRO.

New project approvals for missions like Chandrayaan-4 and Shukrayaan are driving the increase in budget. Chandrayaan-4 aims to bring back moon rocks. Shukrayaan will be India’s first mission to orbit Venus. There are other projects on the horizon too like the Chandrayaan-5.

Space Sciences is a capital-heavy segment. 75% of the funds are for building hardware. Most of this year’s budget increase is for capital expenditure (from ₹98 crore revised budget in 2025–26 to ₹428 crore in allocated budget in 2026–27).

In 2025–26, the Department of Space projected a requirement of ₹876.14 crore for space sciences. However, the Government allocated only ₹371 crore. This year’s rebound to ₹569.76 crore could be because of last year’s cuts.

Despite the financial constraints, this segment has performed well. The successes of Chandrayaan-3, Aditya-L1, XPoSat and Astrosat highlight this. However, with actual spending consistently falling short of allocations, execution is lagging. Considering ISRO’s new mandate and the renewed global focus on lunar exploration, the spending here needs to increase.

Space Technology segment

The Space Technology segment is the largest and the most critical component of ISRO. It consumes over 76% of the budget. It covers the heavy engineering and much of the core research activities that form the bedrock of India’s ability to access and participate in outer space.

The technology budget includes:

• The development of launch vehicles. Launch vehicles are space rockets like LVM3, SSLV, PSLV and the upcoming NGLV.

• The human spaceflight programme like the Gaganyaan mission and precursors to the Bharatiya Antariksh Station, which will be India’s first Space Station.

• The development of Earth observation satellites like Oceansat and NISAR as well as navigational satellites for NavIC.

• The infrastructure maintenance of ISRO centres like VSSC, LPSC, and the spaceport at Sriharikota (SDSC-SHAR).

Evidently, developing rockets and human spaceflight infrastructure remains with the government. This is in line with the new space policy as well. This part of the space technology infrastructure is foundational. It supports all the other space sector activities in the country. 2026–27 budget has a higher capital outlay of around 46.5% compared to the previous years where it hovered around 40% or lower. The Space Technology segment is the main driver of the capital increase. It is being used to build the third launchpad at Sriharikota, the cryogenic engine testing centre, human-rating facilities for Gaganyaan and NGLV. The revenue expenditures cover the fuel and operations for the rocket launches planned for the year.

The Space Technology component sees the most ambitious allocations but also the steepest cuts. It suggests that big-ticket hardware development (rockets/human rating systems) faces significant delays, leading to surrendered funds. The actual spending (and revised budgets) are consistently a lot lower than the allocated budgets.

Slow pace of technology development has led the Finance Ministry to slash the budget at the revision stage for key projects. For instance, the Gaganyaan budget was cut from ₹1,200 crore (allocated) to ₹847 crore (revision) in 2024–25 due to delays in hardware realisation. The semi-cryogenic engine budget was cut from ₹190 crore to ₹115 crore. GSLV Phase-4 was cut by over ₹100 crore. So, this is a recurring pattern. Some of these delays are uncontrollable. There might be supply chain bottlenecks in international components. But the rest are controllable within ISRO and its vendors. The repeated mid-year cuts represent pauses in crucial R&D activities. It means a slowdown in indigenous innovation and delay in Indian space ambitions.

This is keeping in mind this segment’s successes that justify its high funding. The successful commercial deployment of 72 OneWeb satellites using LVM3; developing new capabilities like the SSLV to be transferred to the industry, the RLV (Re-usable Launch Vehicle) as a critical step towards reducing launch costs represent clear headway. There is the Gaganyaan progress also like the Test Vehicle (TV-D1) flight that demonstrated the Crew Escape System and the qualification of the LVM3 rocket for human rating.

India needs more government spending in Space

Drawing on Parliamentary Committee reports, I have added the ISRO projections below. The allocations are always lower than projections. This is not unique to space. There is always a gap between what the branches of the executive ask and what the Finance ministry can afford based on broader priorities and fiscal consolidation.

The most visible spike in projection was in 2020–21. ISRO projected ₹24,686 crore, probably because of Gaganyaan and Chandrayaan-3 development. The allocated budget was nearly half this amount. After all, it was the pandemic year. Since then, DoS seems to have rationalised its projections. The allocations are still lower but by much smaller margins. This indicates better alignment between ISRO’s planning and the Finance Ministry. But this does not mean the space programme is getting all the money it needs.

Indian space budget is ~0.05% of GDP. Other dominant space powers spend orders of magnitude higher than what India spends and this is not counting the commercial funding.

The previous ISRO chairman has said that the Indian space sector’s budget needs to be a lot higher. This is true. The private sector cannot do the things ISRO is tasked to perform. Planetary exploration, lunar missions, and human spaceflight aren’t just for scientific prestige. They come with geopolitical logic and are hard necessities for a space power today. Nations around the world have renewed focus and accelerated spending on their national space programmes in pursuit of these goals. If India fails to keep up in these technological races, it will have to bear political costs. It will lose its seat at the table that will decide the norms, rules, and benefit-sharing of humanity’s endeavours in space.

Bridging the gap between intent and outcomes

Another consistent trend we notice across all the segments and have already discussed is that the actual spending is significantly lower than allocations. Project delays are leading to unspent funds being returned. This high surrender ratio is pervasive within the broader category of scientific departments that includes space, biotechnology as well as science and technology.

This is a state capacity problem and not one of political will. The government has been consistent in allocating over ₹13,000 crore over the last few years. The political intent and support show. The leadership wishes and expects ISRO to achieve its objectives. The bottleneck is ISRO’s absorption capacity. This is particularly evident in ISRO’s capital expenditure. A large portion of the unused budgets often comes from the capital side which includes infrastructure, equipment and facilities—all critical for long-term growth. In fact, DoS’ surrender ratio is higher compared to sectors like Defence or Road Transport which also have capital-intensive activities. Those ministries have used their capital budgets more aggressively in recent years.

This issue can be addressed by positioning the government as an anchor customer for space technology. Not as an investor, not as a subsidiser, but as a customer. Since the new space policy has liberalised the sector, it is no longer a government monopoly. There are over 300 startups and traditional companies in the country alongside ISRO. Most are nascent with few avenues for funding and fewer avenues for early commercialisation. There is enormous scope for the government to act as a customer through service contracts with these companies.

The anchor customer model

An anchor customer relationship is where the government commits to buying services (launch, imagery, analytics) from a private firm at meaningful scale and over multiple years. This allows the firm to justify upfront capital expenditure and attract private capital. The secret ingredient that unlocks these things is the credible demand.

This is not entirely new to the Indian experience. The first instances of government acting as a customer in the space sector began with component contracts between ISRO and industry. Vendors supplied components and subsystems to ISRO’s missions. They were build-to-print contracts. ISRO designed everything. The private sector built it according to these specifications. ISRO absorbed all the risk as well as all the credit or blame. The vendor was rewarded for delivery and not for market success. They had no incentive to innovate on architecture, business model or other downstream services.

The more recent instance of the sector moving closer to anchor customer model comes from NSIL. The space reforms have positioned NSIL as ISRO’s commercial arm. ISRO has transferred the ownership of its Earth observation and communication satellites to NSIL. Other government ministries, foreign, and domestic customers approach NSIL for space technology services. NSIL provides these services by leveraging ISRO assets and by outsourcing the work to Indian industry whenever it can. For example, the Ministry of Defence is paying ₹3,000 crore to NSIL for the GSAT-7B communication satellite to fulfil the army’s needs. NSIL will pull in the industry to build and operate the asset. Another instance is NSIL using the HAL-L&T consortium to build PSLVs. This PPP model is structurally very close to an anchor-customer model, but not quite there yet. Because the demand is still not credible enough. It is coming from outside and can be inconsistent. For instance, foreign customers may prefer to launch via SpaceX over NSIL since SpaceX offers more reliable launch schedules or cheaper costs.

In a true anchor customer model, private firms design and own the satellites, launchers, analytics platforms and offer them as services to the government. Government buys outcomes, not ownership of hardware. If the government commits to a multi-year off-take, it will have de-risked capital expenditure, and the private firm will attract private equity or debt that it otherwise would not be able to. At the end of it, the firms will own their IP and can go on to export the same service globally. Many globally successful space companies like SpaceX, AWS, and Northrop Grumman emerged from this model.

India is already in the early stages of doing this systematically. There is currently an Earth observation-PPP project where a consortium of companies are investing over ₹1,200 crore to design, build, own and operate a constellation of satellites (government as an anchor customer de-risking capital expenditure). They will use this asset to sell imagery and analytics to government and armed forces in India. Though they call it a PPP, this is functionally the government acting as an anchor customer for EO services. In another instance, the Telangana government has signed a contract with a private company to digitise farmlands to replace manual crop surveys. As a result, these companies have attracted new private investment (government as an anchor customer unlocking private investment).

Clearly, the anchor customer model is taking shape in the country. But the instances are few and far between. Government should fund more contracted service payments to Indian firms that can grow and add to India’s space prowess.

A lot of that has changed in recent times. The bipartisan consensus in Washington that favoured India is now less predictable and more contested under Trump. We’re facing US economic pressure that, in some sectors, is as consequential as anything China throws at us. Right now, India faces among the steepest US tariffs, steeper than China’s.

Meanwhile, the Indo-Pacific remains as fragile and volatile as ever. Several neighbours face acute instability. Political crises, conflict, or debt stress often afflict them simultaneously. These circumstances limit opportunities for long-term economic engagement. Meanwhile, the war and sanctions have severely constrained Russia, inflicting long-term damage to investment and access to technology. Sanctions have forced Russia to redirect trade and procurement toward a smaller set of partners, especially China.

Amidst all this, the Free Trade Agreement with the EU is announced—at a crucial time when India seeks new economic and technological options. To assess its value amid geopolitical volatility, we must determine whether the FTA’s foundation supports a strategic superstructure that secures India’s interests.

The FTA sets a framework for predictable regulatory cooperation. This groundwork enables credible partnerships in critical sectors without duplicating efforts. Here are some examples.

Many countries want to accelerate their space programs to match the launch cadence of the US and China. India and France could anchor a consortium with Brazil and Indonesia. India provides low-cost launches. Brazil offers equatorial launch sites that can reduce costs. France supplies advanced satellite payloads. Indonesia has an acute communication and maritime surveillance requirement across 17,000 islands. The EU deal matters here because it pushes India and France toward clearer, more predictable rules around trade and IP—reducing friction for joint projects. In an aggressive but realistic timeline, this group could launch a jointly owned earth-observation constellation within three years and a communication megaconstellation in five. Operational capability, shared sovereignty, strategic independence.

When it comes to AI and compute, India and several European countries share the same strategic vulnerability in terms of access, along with the UAE and Japan. France brings open-weight AI models, the UAE offers deployment capital, Japan provides semiconductor packaging expertise, and India contributes engineering depth and market demand. The FTA aims to de-risk collaboration by clarifying IP, services, and digital-trade disciplines over time, enabling capability-building now as the broader agreement phases in. These four nations could establish shared compute infrastructure within 18 months and jointly develop open alternatives to Compute Unified Device Architecture and Electronic Design Automation tools.

Gathering ideas from a collective brainstorming exercise at the Takshashila Institution, I’ve written a paper about this with my co-authors. We have argued that sectoral plurilateral blocs, like the examples above, provide shared resources, collaboration, and resilience, making them our best bet for overcoming geopolitical coercion.

Our paper outlines 12 such sectoral blocs spanning semiconductors, fintech, counterterrorism, cyber-fraud, disaster relief, migration, pharma and more. By participating in these blocs, India reduces single-point dependencies, creates alternatives, and gains more flexibility to navigate a world where old certainties have weakened. Each bloc delivers specific benefits, such as secure access to technology, improved crisis response, or greater economic autonomy.

Link to the paper: https://takshashila.org.in/content/publications/20260125-Sectoral-Plurilateral-Blocs.html

In my latest Op-Ed for The Print, I argue that while these setbacks are serious, they are statistical anomalies for a workhorse with a 95% rate of success over three decades. The real challenge is not a 𝘣𝘳𝘰𝘬𝘦𝘯 rocket. It is the managing of a mature technology system during a critical transition in India’s space policy.

I explore why this moment feels heavier than past failures and outline a constructive path forward: transparency, a clear return-to-flight plan, and keeping the long game in sight.

From the article:

… For decades, it maintained an unbroken streak of successes. It earned a global reputation as one of the most reliable launch vehicles. It carried the weight of India’s most visible scientific missions like the first lunar probe (Chandrayaan-1), the first Mars orbiter (Mangalyaan), and the first solar observatory (Aditya-L1). Few rockets have demonstrated this level of mission flexibility.

The point here is this: The PSLV is a mature piece of technology. It is a platform designed, optimised, and perfected over three decades. But even mature technology is not immune to entropy. Systems can drift when suppliers change, components age, and processes evolve. However, ISRO has proven it knows how to build a reliable launch system, how to run missions repeatedly, and how to recover from setbacks. It is reasonable to believe ISRO can fix PSLV again…

… it must prioritise three decisive actions.

… it should commit to a clear cadence of communication: What is known, what is still being tested, what hypotheses have been ruled out, and when the next update will come. Silence creates a vacuum that gets filled with speculation. It is bad for commercial confidence.

… The industry needs to see a checklist. By when will the failure investigation be closed? By when will corrective actions be completed? What additional verification tests will be done on the ground before flight? What are the “go/no-go” criteria for declaring PSLV ready again? This restarts the rest of the ecosystem and commercial plannin

… keep the long game in sight. As ISRO moves toward higher-risk missions, failures become more likely because the frontier is harder. That is exactly why a culture of openness and process discipline must begin now.

Click here to read the full analysis here: theprint.in/opinion/isro-was-set-to-move-on-to-a-new-era-pslv-failure-has-stalled-indias-space-policy/2830706/

(𝘕𝘰𝘵𝘦: T𝘩𝘦 𝘩𝘦𝘢𝘥𝘭𝘪𝘯𝘦 𝘢𝘯𝘥 𝘴𝘵𝘳𝘢𝘱 𝘸𝘦𝘳𝘦 𝘤𝘩𝘰𝘴𝘦𝘯 𝘣𝘺 𝘵𝘩𝘦 𝘱𝘶𝘣𝘭𝘪𝘤𝘢𝘵𝘪𝘰𝘯.)

I wrote about it for the Hindu’s Text and Context today:

Link to the website version: https://www.thehindu.com/sci-tech/technology/how-does-satellite-internet-work-explained/article69924030.ece.

I write for the Indian Express in this opinion piece:

… we must first stop thinking of “space” as a single destination. A more useful framework is to think of Earth as our home, the immediate space around us as the farm that sustains it, and deep space as the great, unexplored wilderness.

The farm (Earth-Space)

The farm is the region around Earth, known as the Earth-Space, where our satellites operate. This farm sustains life in our home. It is not food which is grown here, but data. Modern life depends on this data for everything — GPS for navigation and financial transactions, satellites for weather forecasting and broadcasting, and constellations for the global internet.

The farm is not tended to by humans. Uncrewed technology like satellites and probes are able to do this job far more cheaply and safely. The potential for the data they generate is endless and priceless.

The astronauts we send to this farm through missions like the Axiom-4 and Gaganyaan are not mere passengers. They are pioneering specialists trained to survive beyond the safety of our home and skilled in performing complex tasks and experiments that require human ingenuity and dexterity. Their experience and learning inform their successors who will dare to venture out even further into the wilderness.

The wilderness (outer space)

The Moon, Mars and beyond are fundamentally different from our home and even the farm. They are full of risks and opportunities. Today’s technology cannot use this region for sustenance at home. The costs of accessing the resources here are either too high, take too much time or both. With scientific advancement and technological progress, perhaps our farms can expand to these regions. Or even form a new settlement in the wilderness and go on to build another home and another farm around it. Therefore, the exploration and discovery of the wilderness is necessary to realise the long-term opportunities.

As a part of exploring this wilderness, uncrewed missions like Chandrayaan and the Mars rovers are sent first. They are the “drones” that scout out the region. They map the terrain, identify risks, and locate resources. They go first because it is dangerous and unknown.

However, you don’t build a settlement in a wilderness with scouts alone. You need pioneers. Humans are adaptable, can problem-solve in real-time (the “Apollo 13 moment”), and make intuitive leaps that a pre-programmed rover cannot…

Winning hearts and minds

The human element extends beyond the practical reasons. Great endeavours are powered by human emotion. The story of Shubhanshu Shukla venturing into the cosmos on behalf of his nation captures the imagination of millions. The stories provide a human face to a highly technical and challenging endeavour. They become focal points for national pride and aspiration.

This public imagination is not just a feel-good emotion. It is a strategic asset. Like India-Pakistan cricket matches, it helps societies surpass domestic squabbles. It creates a groundswell of popular support that translates into political will…

The long game

As technological progress accelerates, the question will shift from whether humanity will travel to other planets, to who will travel and write the rules. In this future, human spacefaring capability is no longer just symbolic; it is central to Indian national power and international standing.

The nations that are physically present in the “wilderness” will be the ones who shape the future of humanity in space. They will set the precedents for law, commerce, and ethics beyond Earth. Gaganyaan is India’s entry ticket to having a seat at that table. It is a declaration that India intends to be a rule-maker, not a rule-taker.

Read the full article here: https://indianexpress.com/article/opinion/columns/as-axiom-4-and-shubhanshu-shukla-lift-off-so-does-india-10088066/

While we often think of space as a vast, infinite frontier, the area that is of most consequence and increasingly contested is Earth-space—the part that envelops our planet.

I discuss these ideas and more in my journal article in 𝘿𝙚𝙘𝙮𝙥𝙝𝙚𝙧.

Thrilled to be in incredible company in this latest issue, which features contributions from Paul Krugman and Najeeb Jung.

Download the full PDF here: https://asiaresearch.co/decypher-journal-em-powered-authority-in-a-fragmented-world/

A few months later, the Trump administration has repealed the framework. Unsurprisingly, the reasons stated are about innovation and diplomatic relations. Yet, the same thinking remains, in other forms. I suspect that we haven’t seen the last of such framework for AI control by the US.

My colleagues and I have written a short issue brief about the implications of this move.

From the brief:

India was unfavourably placed in tier 2 of the AI Diffusion Framework. This placement meant that there were some thresholds on the number of AI chips that India could procure. However, the hard caps on the number of chips were far higher than the current demand in the Indian compute ecosystem. In other words, India could have gotten all the chips it wanted in the next few years under the framework.

Additionally, the framework had many second-order effects against Indian prerogatives. The framework would have undermined India’s goal of maintaining a level of strategic autonomy in its technology ecosystem. The possibility of Indian Global Capability Centres (GCCs) being denied access to advanced AI chips as country thresholds were reached or having to seek authorisations for purchasing chips is not an ideal situation. This could have led to tier 1 countries becoming the favoured destination for building cutting-edge technologies over tier 2 countries. These constraints would have led to a loss of talent and investment from India to tier 1 countries.

Given these realities, the government’s emphasis on building sovereign compute via GPU procurement warrants reconsideration. Indian companies are now free to procure all the chips they need, and the government could focus on other priorities. Union Minister of Electronics and IT, Ashwani Vaishnaw’s, recent comments about Indian efforts towards developing indigenous GPUs by 2029 could be seen as a response to US attempts at maximising control of the technology.

Despite the repeal of the framework, it is clear that this is an era of export controls for AI chips. The geopolitical climate will undoubtedly continue to dictate the ebb and flow of AI chip access. In its aspirations for strategic autonomy, India has a reason to develop a baseline capability in making its own advanced AI chip ecosystem. However, India is heavily reliant on foreign-made software and machinery. This makes the prospect of India rapidly developing these capabilities on its own highly unlikely. The extent to which the US and its allies will help India in this regard will not only depend on India-US relations but also India’s own position in the AI value chain.

Read more at: takshashila.org.in/blogs/us-ai-diffusion-framework-repealed

Satellite internet is not a new invention. We have operationalised it decades ago. However, the satellite internet megaconstellations of today mark a new age of space-based internet. They consist of hundreds and thousands of satellites in the Lower Earth Orbit, working together to beam a global internet.

Given the penetrative, foundational and integrative ability of the internet, satellite internet technology holds immense potential for military strategy and economic productivity. Its inherent dual-use nature also creates security and regulatory challenges. It is imperative that we understand this technology before we go about using, investing or governing it.

I have written a paper, a first in a series planned around satellite internet megaconstellations, where I explain how it all works and what it is being used for. Give it a read: takshashila.org.in/research/satellite-internet-explained

From the article:

…Even among key allies exempt from export controls, this policy sets a concerning precedent of unilateral American executive action that could potentially deny them their strategic priorities in the future… Countries, including U.S. allies, will become more motivated to develop independent supply chains and sovereign AI innovation to circumvent U.S. technological constraints. This could fragment the global AI ecosystem that the U.S. currently dominates…

The policy framework disadvantages India, positioning it less favourably than anticipated…The framework evaluates India solely by its technological capabilities in AI, while ignoring the strategic value of the India-U.S. partnership. It also disregards the substantial social capital built between the two countries... Instead, it serves as a stark reminder of the U.S.-led technology denial regimes that restricted India’s access to nuclear technologies for three decades following its nuclear tests.

There is a disconnect between America’s export control policies and its strategic partnership with India regarding Indo-Pacific cooperation… It could compel India to hedge against its technological and economic dependence on the U.S.

Read more at: https://www.thehindu.com/opinion/op-ed/implications-of-the-ai-diffusion-framework/article69210897.ece

… The space sector is highly capital-intensive, and the Rs 1,000 crore proposed by the government is insufficient to satisfy the funding needs of space startups.

… The government would be better served by using the Rs 1,000 crore fund to enable the purchase of more such services from the private space sector in India.

Having the government as an anchor customer provides a stable source of demand for space-based services, reducing uncertainties in the market. It will boost the industry and increase investor confidence in the Indian space sector, which can draw in more investments from home and abroad. It will also induce healthy competition among companies, enabling only the best to win government contracts. Finally, such an approach incentivises innovation, efficiency, and overall success among Indian space startups.

… NASA’s Commercial Orbital Transportation Services (COTS) programme is an excellent example of how this government procurement model yields favourable outcomes for all stakeholders. In 2004, the US government announced that its space shuttles would be retired due to high costs and safety concerns. The COTS programme was started in 2006 to replace the space shuttles.

Under the COTS programme, companies had to achieve certain milestones to avail payments, ensuring the funds were linked to performance and progress. The risk was shared between the companies and the government, rather than the government bearing all the risk with upfront investments.

The programme was a success. SpaceX’s Falcon 9 rocket and Orbital Sciences Corporation’s Cygnus spacecraft were developed under COTS. SpaceX completed a cargo delivery demonstration in 2012, restoring the American ability to resupply the International Space Station (ISS).

Read the full article here: https://theprint.in/opinion/govts-rs-1000-cr-fund-for-space-sector-has-drawbacks-it-should-bet-on-outcomes-not-ideas/2222314/

… The government established NewSpace India Limited (NSIL), a central public sector enterprise, as ISRO's commercial arm. One of NSIL's main functions was to build and launch satellites based on market demand. This approach has led to a decrease in the launch of ISRO-built satellites, as NSIL now seeks end customers prior to actual launch.

… The deployment of these satellite constellations is driving demand for launch services globally. Even ISRO's LVM3 rocket carried 72 OneWeb satellites across two launches in 2022 and 2023, which marks the first instance of Indian rockets launching satellites for space-based internet…

… ISRO’s priority must be to drive costs down. Despite being praised for their low-cost launches, ISRO faces stiff global competition from companies like SpaceX, which have already implemented and tested reusable launch technology. This technology enables SpaceX to offer more competitive prices per kilogram of payload…

… For the foreseeable future, it is crucial for the government to assume the role of the anchor customer, providing stability and certainty to the burgeoning space ecosystem. This support will not only catalyse the expansion of space launch industries but also foster local development and spur innovation within the sector.

Read the full issue brief here: https://takshashila.org.in/blogs/demand-for-space-launch-services-in-india

From the article:

Space Shuttle’s retirement created an opportunity: When the US retired the Space Shuttle, Nasa had no ready substitute spacecraft…

A burgeoning American space industry cultivated robust capacities in LEO that has led to a flurry of private spacecraft capable of taking crew and cargo to orbit. Their pursuit of a competitive advantage spurred the development of reusable spacecraft, slashing launch costs by a staggering 95% compared to the Space Shuttle era. These advancements have done wonders for space accessibility, opening wider doors to Nasa and many space and satellite companies globally. We can create similar opportunities in India…

…The transfer of technical knowledge from Isro to the private sector for LVM3 production will spread crucial space expertise. These capabilities could then strengthen India’s supply of launch vehicles to meet growing global demand for launch services.

Read more at: www.livemint.com/opinion/online-views/private-sector-participation-in-lvm3-production-could-mean-a-space-sector-boom-11718129556130.html

From the article:

…Cryogenic and semi-cryogenic rocket engines offer much higher energy density and specific impulse than traditional solid- and liquid-fueled engines. Specific impulse measures the engine’s efficiency in converting propellant to thrust. It is essentially the rocket’s ‘mileage’. Semi-cryogenic engines are the most popular rocket engines used today. Compared to solid-fuelled variants, semi-cryogenic engines allow for higher variability of thrust—the ability to throttle.

They also offer a higher thrust-to-weight ratio than liquid-fueled engines. Additionally, propellant storage and handling are simpler in semi-cryogenic rockets than in liquid-fueled ones. This is also true for cryogenic rockets that use liquid hydrogen, which needs to be stored in heavy insulation tanks. Semi-cryogenic engines use denser liquid kerosene, which can be stored in a smaller volume, reducing the overall weight of the engine and the launch vehicle…

There are key strategic advantages, too. All spacecraft, after being launched, must pass through a precise point on the opposite side of the globe, called an antipodal point, before completing their first orbit. Potential adversaries could blockade this antipodal point of well-known spaceports such as Sriharikota or the upcoming Kulasekarapattinam, thereby cutting off India’s access to orbit. However, the capability to launch from any location enables India to overcome this vulnerability…

…Heavy launch missions have significant advantages over dedicated ones. For instance, multiple smaller payloads can be launched together on heavy launch vehicles, which makes these missions significantly cheaper…

…Heavy launch vehicles can reduce the need for expensive efforts to reduce the size and weight of satellites. This lowers barriers and opens access to space to more players, empowering the entire space economy.

Read the full article: www.firstpost.com/opinion/agnikuls-first-test-flight-time-for-private-sector-to-be-wings-of-indian-space-industry-13779865.html

From the article:

The guidelines describe the process of this authorisation for the different areas of space-based technologies such as satellite communication, rocket launch services, and earth observation. Yet key areas such as scientific missions, and positioning, navigation, and timing services, which enable GPS and air traffic control, are missing from the guidelines…

Space-based services range from telecommunication services to military applications to cybersecurity, and fall under different ministries and departments under the Government of India. Though IN-SPACe is a ‘single window agency’ for authorising space activities, private players must seek separate clearances from relevant ministries even after getting a provisional authorisation from IN-SPACe.

…Ultimately, India needs a space regulation Act that addresses the gaps in the guidelines. Its formulation should be a consultative process with inputs from all stakeholders. Together, these legislation can democratise the space governance ecosystem, and provide the necessary legal clarity that the private sector expects before undertaking risky ventures. The path to a burgeoning private space sector in India starts with the right regulatory environment.

Read the full article: https://www.deccanherald.com/opinion/new-guidelines-could-hamper-indias-promising-space-industry-3035235

Each of these components plays a vital role in making AI tick. The variety within each component type offers trade-offs between performance and efficiency. As a General Purpose Technology, AI will shape techno-national strategies worldwide. All stakeholders need to understand these specialised chips and support the development of alternatives to ensure a more diverse and resilient AI chip ecosystem.

Toward this end, I have co-authored a paper that provides an understanding of these various AI chips, their characteristics, limitations, and what makes them suitable for certain AI applications and not others. It also provides an overview of the global market landscape and identifies some single-vendor dependencies and monopolies. Ultimately, it emphasises fostering open-source technologies and ecosystem diversity.

Click to read the full paper here.