Buyers of first resort

Words by Neil Hacker
14th September 2021
Issue 5

How do technologies get off the ground? As well as seed funding, many of the best technologies require Buyers of First Resort, which buy products until they improve enough to get to efficient scale.


Research and development can have enormous spillovers for the rest of society. When businesses just make products and sell them, they capture a solid share of the value, or surplus, created for the world. By contrast, a famous study estimates that in the long run, 98% of the value of a new innovation goes to society at large, and just 2% goes to those who made the key contributions to its invention.

Because of this, it is widely accepted that we need institutions like patents, copyright, and trade secrets to protect intellectual property and allow innovators to capture some of the value they create, and for governments to invest in R&D subsidies and tax credits to make grants to academic researchers and even to open up their own labs. What is less well understood is the power of bodies acting as ‘Buyers of First Resort’ – where some entity, public or private, helps to establish the first market for a good or service before it has reached the stage where it could support a customer base of its own.


To the people who worked at Bell Labs, innovation was not an event but a process. It involved the discovery of, say, a new material for transistors, its development – using the new material in transistors – and then a third step many might overlook: actually creating a market for the new invention. The Bell Labs team understood that even something that has the potential to be groundbreaking might flounder on first contact with the outside world. This is where a Buyer of First Resort can help bridge the gap for promising, but immature, technologies.

This concept is similar to what Matt Clifford calls being a venture buyer. For example, Clifford writes, ‘The DoD allowed DARPA to bridge the gap between basic research and commercial application by providing real-world demand ahead of the private sector’s willingness to pay’. It is this idea – of helping to bridge the gap between development and adoption – that being a Buyer of First Resort aims to solve.

Most innovation policies focus on the supply side by funding basic research or giving companies subsidies or grants. These are crucial elements, but replicating the success of a lab like Bell Labs may require supporting every stage of the process – including making sure the product makes it to market.

This can complement other ways we support early-stage R&D. Subsidising early-stage research via grants or tax credits does not imply that any of this will be successfully translated into useful applications. Subsidising only patentable applications through intellectual property institutions and buying finished products may lead to free riding on basic research. Generating an optimal level of research may require an ecumenical approach that works at all points of the process.


Buyers of First Resort work in two key ways that other R&D support systems cannot: speeding up technology transmission and scale-ups, and encouraging firms to apply their technologies in risky ways. Two examples from NASA illustrate how these work: transistors and the commercial orbital transportation service program.

At the beginning of the Apollo missions, NASA used a computing technology called rope core computing – ‘rope’ because it involved people literally threading circuits together with ropes and magnets. Though vacuum tubes operated faster, they were too unreliable for the stresses of the missions they were needed for.

Luckily, a new technology was emerging from Bell Labs: the transistor. The first transistors were dramatically larger than those we have now, measuring around 1 cm squared: we can now fit roughly 15 billion transistors in the same area. Aside from their size, transistors were also not terribly reliable at the time.

Yet NASA realised that this technology would be needed to get to the moon. It began buying as many transistors as they could. In 1962, the US federal government bought every single integrated circuit made in the world. By 1965, this proportion had fallen to 72%, and the market for transistors had become twenty times larger.

The enormous orders from the federal government created incentives for improvements to transistors that other kinds of funding may struggle to create. NASA, acting as a Buyer of First Resort, provided transistor research with the resources and experience to improve while also bringing them into quasi-market conditions, thus requiring the products to provide value. This helped bridge the gap between the early technology, which was too expensive for any selfish buyer, and the point at which they could start their journey to full market adoption.

It is worth noting that transistors – being the better technology – would have almost certainly gained ground against rope core and vacuum tube computing eventually, but this could have taken much longer, delaying many of the subsequent advances in technological progress we’ve enjoyed as a result.

Buyers of First Resort encourage risk taking by sharing some of the risk that the product does not work out, rather than placing it entirely on the firms developing it. NASA’s commercial orbital transportation service program (COTS), which ran from 2006–2013, illustrates this effect. The goal of COTS was to create options to resupply the International Space Station after the space shuttle was decommissioned.

What was different about COTS was that NASA did not want to just start a new program themselves; it wanted private industry to create the launch options in the hope that this would allow NASA to use the private sector’s competition, diversity, and experimentation to get the same output at a lower cost than if it developed transistors in-house. The hope was that this would also stimulate American space companies to build supplies of launch capabilities and eventually become global leaders that served customers all around the world, which would cement America as the leader in space.

This was a risky proposition for private companies: their competitors might be able to copy some of their successful ideas and save the spend on the failed ideas. A private buyer would hardly be likely to indemnify them when things go wrong. Subsidies and tax credits could have helped by covering off much of the basic research, but they would not cover the risks in trying to develop these into successful products.

To overcome this problem, NASA agreed to be the cornerstone user of launch services provided by the companies it contracted with for a pre-agreed number of years after the shuttle was decommissioned, in order to provide reasonably stable market demand for the successful companies. This Buyer of First Resort mechanism was only part of the story – as well as this, the government spent around $800 million on direct funding, along with the private sector’s $1 billion – but it was a vital part of the process.

Because NASA was a Buyer of First Resort, and because it used milestone-based funding that reduced risk for the companies involved, the project was appealing to a wider variety of companies, including, unusually for a contract this size, many startups. One of those startups was SpaceX. While SpaceX did no launches in 2012, it made up over half of all launches into orbit just six years later, and has now become the first private company to send astronauts to the International Space Station.

In late 2008, Musk was quoted as saying, ‘I could either pick SpaceX or Tesla or split the money I had left between them…If I split the money, maybe both of them would die’. Luckily, two days before Christmas it was announced that SpaceX had been awarded the COTS contract, saving the company from the brink of bankruptcy. Not only did COTS create a global power in the space industry, but it led to a rocket – the SpaceX Falcon 9 – that was ten times cheaper than what NASA would have paid to develop the system in-house itself.


Buyers of First Resort have also been a crucial part of efforts to develop vaccines for Covid-19. Operation Warp Speed, the American programme to support vaccine development, acted as a Buyer of First Resort when it used advanced purchase contracts to channel funding to vaccine candidates before they were viable in the market. Other countries such as the UK also bought vaccines well in advance of their being approved, shifting future demand to the point at which it was needed to guarantee a higher level of supply. Although this did carry the risk of paying companies that would fail, the scale of the crisis meant that this was a risk worth taking. Private buyers simply could not take on this risk in the way that governments could do on behalf of society as a whole.

The success of Operation Warp Speed highlights two points. First, that being a Buyer of First Resort often works hand in hand with other forms of investment: private companies poured money into R&D themselves, protected by patents, exclusivity rules, trade secrets, and other direct and indirect subsidies. Early orders complemented these supports. Second, it shows how acting as a Buyer of First Resort can speed up progress in a situation with great time pressure.

Since March 20, 2020, the pandemic-era low point of the S&P 500 index, both the stock market as a whole and the stocks of vaccine companies have risen in value by almost 100%. But these rises represent a gain of around $500 billion for the vaccine companies and $20 trillion for the broader market. This illustrates how the mechanisms we have for subsidising investment without Buyers of First Resort may not have been enough to make the risks worthwhile for the firms that did develop the vaccines.

Another example is Stripe Climate’s work on direct air capture, for which the term Buyer of First Resort was first coined. Direct air capture is the process of extracting carbon from the air and either reusing it or sequestering it back into the ground or some other long-lived medium. Reaching our current climate goals will require us to actively remove the carbon that is already present in the atmosphere, not simply reduce future emissions. Currently, however, direct air capture is very expensive and, in almost all cases, a money-losing venture.

Stripe’s aim in being a Buyer of First Resort is to help companies that create direct air capture technologies to reduce costs until they are profitable in open market conditions. An additional benefit to this model is that, rather than putting investment directly into companies and ‘picking winners’, buying follows market prices. The most capable companies are able to sell more tonnes of carbon, so more resources are able to flow to them in return.

Stripe Climate is a slightly strange example: although it is indeed a strong example of a Buyer of First Resort, the undertaking is being done by a profit-making firm. Its motives are not necessarily transferable to other companies: Stripe seems to be taking this route as a form of socially responsible investing, rather than with the expectation of financial profits in the future (although it may have a plan for this we can’t see).

Looking back at Operation Warp Speed, the advance market commitments to buy vaccines created vast amounts of value for society, but capturing that value would have been essentially impossible for any private company. Part of this is because the government has different goals than most private firms, and can do things that are socially beneficial even if it will not be able to recoup all or most of the benefits as a private company would have to. It also comes from the government’s ability to ‘get in all the cracks’ of the economy: when it does recoup value created, it can do so through taxes that fall all across the economy, not just on a specific business or sector.


Buyers of First Resort often work best in situations where there are relatively clear objectives and few overlapping interests. COTS wanted to create a launch vehicle that could get to the International Space Station; Operation Warp Speed wanted to develop a safe and effective Covid vaccine at scale; Stripe wants to buy carbon that has been removed from the atmosphere. All of these have a narrow focus, and it is likely that experts in the area will have some idea as to which firms might be more likely to benefit from the additional help of a Buyer of First Resort. Indeed, all of these schemes only acted as a buyer for a small subset of all the options that could have been bought.

In his article “The Use of Knowledge in Society”, the economist Friedrich Hayek showed how the price mechanism aggregates the knowledge of all participants in a society. No single person or group could ever hope to know the desires and uses of all people and of all goods, but they can be reflected in the price.

Buyer of First Resort schemes, like the prizes that gave us a simple and practical way to measure longitude in the 1700s, only work in cases where we think that society’s needs are so obvious that we can ignore this. Most of the time this isn’t the case, and other ‘outcome-blind’ research funding mechanisms like patents or R&D tax credits will perform better. These don’t require the government to ‘pick winners’ and guess where the future of the economy and technology are going. But sometimes the goal is obvious, and the question is just how to get there.


We’ve looked at some areas where being a Buyer of First Resort has served the purpose of some combination of de-risking, speeding up, or pooling market information, and all to great effect. You might now wonder which other areas this idea could be applied to now or in the future.

One area where a similar strategy has been proposed is in the sectors of nuclear fission and fusion. The suggestion is that the Department of Energy applies a similar strategy to COTS, where it would share costs and be a Buyer of First Resort for newer types of reactors that may start off as more expensive, but could eventually be used to supply the world with heat and electricity.

De-risking is particularly important in these areas because electricity is a commodity. Convincing end users to pay a higher-than-market price for the sake of the long-term gains we would get from these technologies would be incredibly tricky. Targeting a new nuclear energy rather than making a broader push for buying green energy would help to narrow down the focus, so a few projects that are likely to be successful would be targeted and would have a stable expectation of demand if they did work.

How far will the marginal dollar go? Bob Mumgaard, the founder of Commonwealth Fusion Systems, has suggested that it may take on the order of $10 billion to push fusion above the break-even point, where it would create more energy than we put in. This sounds a large amount but is almost 100 times smaller than the US expenditure on energy per year. Why haven’t investors just filled in the gap? As with the earlier examples, they have to some extent. But if the Department of Energy acted as a Buyer of First Resort, it would help place semi-market pressure on the firms involved, bringing a reality check to the ones that can really make progress – and the ones that can’t.


Investment into scientific and technological research often has enormous spillovers for society. These positive externalities have inspired enormous efforts to support research, to try to get closer to a level that is optimal for society. All of the standard mechanisms are important, and governments should continue to create patents, give out grants, and tax returns from research lower than other sources. However, when we know what the goal we have in mind is, governments and altruistic private organisations should also consider acting as a Buyer of First Resort. In some cases, direct support for useful products is the best way to foster good new ideas.


Neil Hacker is pursuing a MSc in Computer Science at Imperial College London. You can follow him on Twitter here.

Image by SpaceX on Unsplash.