What role software can really play in helping us reach net zero goals?

While considerable venture capital investments have already flowed into the software sector, the intensifying climate crisis is pushing the need for radical action to the forefront. Software is currently receiving negative publicity as seen as involving too “shy” of an effort. On the other hand, hardware and infrastructure investments, which were traditionally overlooked by asset-light venture capitalists, are gaining momentum. The recent substantial funding rounds of battery makers Northvolt and Verkor or fast charge EV providers such as Elektra serve as notable examples. But does this trend imply that software is losing its relevance in addressing the climate crisis?

A few numbers from the IPCC to size the problem

• Climate change is caused by global warming: global surface temperature has reached +1.1°c above “pre-industrial” levels (i.e. 1850–1900) in the last decade (AR6, 2022 & 2023).
• Global warming is principally caused by the concentration of Greenhouse Gas (GHG) in the atmosphere, measured in part per million (ppm), with carbon dioxide (CO2) being the most significant contributor: in 2019 atmospheric CO2 concentration was about 410 ppm. As a reference, pre-industrial levels of CO2 were around 280 ppm.
• The concentration of GHG in the atmosphere is linked to GHG emissions. And although there are some “natural” sources of CO2 (i.e. decay of organic matter, volcanic eruptions or the respiration of plants and animals), those emissions are mostly anthropic (i.e. caused by human activities). Global net anthropogenic GHG emissions were 59 GtCO2-eq in 2019, about 12% higher than in 2010. And 65% of net anthropogenic GHG emissions were CO2 from fossil fuel combustion and industrial processes.

To give you an order of magnitude, between 1850 and 2019 we emitted 2,400 GtCO2 (half of it in the last 20 years). We have basically consumed 4/5th of our carbon “budget”, i.e. maximum amount of carbon (c.500 GtCO2) that can be released in the atmosphere to limit global warming to 1.5°C, which is the limit we fixed ourselves in Paris in 2015.

Going back to basics

To understand the root causes of anthropic CO2 emissions, a useful tool is the Kaya Equation: it’s a descriptive equation, developed by Yoichi Kaya, a Japanese Economist in 1993, and more recently popularised by Jean-Marc Jancovici. It states that:

Where CO2 is the total anthropic carbon dioxide emissions, POP is the total population, GDP is the gross domestic product (i.e. total economic output), and Energy is the total energy use.

Basically it is saying that the world’s anthropic GHG emissions are a function of:

• The world’s population. The more people they are on the planet, the more we emit.
• The GDP, production or economic output per person, which is linked to the level of wealth. Once again, the more we produce per person, the more we transform, the more we emit.
• The energy efficiency of the economy (= amount of energy used per unit of GDP), which can can be expressed in kWH/$.
• The carbon intensity of the economy (= carbon emitted per unit of energy transformed, which can be expressed in CO2/kWh). It is intrinsically related to our energy mix.

In plain English, it means that CO2 emissions caused by human activity are linked to our Population Growth x Economic Growth x Energy Efficiency x Carbon Intensity.

Coming back to our initial challenge, it is generally acknowledged that to stabilize the level of CO2 in the atmosphere (and stay below the 1.5°c landmark) we must divide global GHG emissions by 3 at least before 2050.

Assuming we are not in favour of an economic decline or a drastic demographic reduction, then it’s fair to assume the world’s population will increase by +1% p.a. and the GDP per person by +2% p.a., which equates to multiplying the population by 1.3 by 2050 and the GDP per person by 1.8 by 2050.

As a result, to divide CO2 emissions by 3 within this time frame and to match the +1.5° trajectory, we thus need to reduce the “carbon content” of our economy by 7 through 1) reducing its energy intensity or 2) its carbon intensity. Basically we need to build more efficient machines and systems and we need to use low-carbon energy.

More on this here as explained by Jean-Marc Jancovici.

Energy efficiency: How to reduce our “Energy intensity”?

The energy intensity of the economy can be reduced by becoming more energy efficient, i.e. using less energy to produce the same output or execute the same function. Basically getting the same or better performance using less energy. Putting aside policy & regulation that can set energy efficiency standards and practices, here are a few examples of energy efficiency drivers:

• Technological improvements: e.g. LED light bulbs use significantly less electricity than traditional incandescent bulbs for the same amount of light output.
• Insulation and energy efficient buildings: properly insulating homes and buildings can significantly reduce the amount of energy required for heating and cooling. Using energy-efficient designs and materials in new construction can also have a significant impact.
• Smart grids and energy management systems: these can optimize the distribution and use of electricity in a power grid, reducing energy waste.
• Transportation efficiency: vehicles can be made more efficient through better design, lightweight materials, and more efficient engines. Public transportation and carpooling can also increase the efficiency of person-miles traveled.
• A more “circular” economy: by promoting repair, reuse, recycling and up-cycling, we can extend the lifespan of existing assets. This in turn increases the denominator of the energy efficiency (i.e. the economic output of a given asset), while the numerator remains constant.

And so on…

How to reduce our “Carbon intensity”?

The carbon intensity of the economy can be reduced by emitting less carbon (in the form of CO2 emissions) for each unit of economic activity. This is linked to our energy mix. Here are a few examples of reduction drivers:

• Shift to low-carbon or zero-carbon energy sources: we could rebalance our energy mix from fossil energy towards renewable energy sources like wind, solar, and hydroelectric power, or nuclear energy (fission).
• Improvement in Carbon Capture and Storage (CCS) technologies: there is an in upward momentum to promote technologies (e.g. Post-combustion capture) that can capture CO2 emissions produced from the use of fossil fuels in electricity generation and industrial processes, preventing the CO2 from being released into the atmosphere.
• Improvement in Carbon Dioxide Removal (CDR) technologies: these are techniques to remove CO2 that is already present in the atmosphere and to store it durably in e.g. geological reservoirs.

And so on…

So where do we stand on all of this?

We are not progressing fast enough.

The global energy efficiency has only decreased by 2% p.a. between 2010 and 2019. And carbon intensity only by 0.3% p.a. As a comparison point, we would need to decrease carbon intensity by 7.7% p.a. to meet our net zero goals by 2050 (AR6, 2023).

Source: IPCC, AR5, 2014

Projected CO2 emissions from existing fossil fuel infrastructure without additional abatement would exceed our remaining carbon budget: it is clear that if we continue on the same path, it is likely that warming will exceed 1.5°c during the 21st century.

We need to act today, not wait for tomorrow.

We are playing against the clock. Basically, we need mitigation options that are feasible and effective today: the more we wait, the more our set of options becomes constrained and less effective. And the more we also become ‘locked-in’ to high emissions infrastructure, with level of risk increasing across the board.

We need breakthrough innovation and increasing investments in Hardware & Deeptech.

We need hardware solutions (either new technologies or existing technologies becoming scalable and commercially viable) to enable some of the biggest transitions needed.

The unit costs of several low-emission technologies are quickly falling (i.e. 85% for solar energy, 55% for wind energy, 85% for lithium-Ion). We need more investments in the space to accelerate large scale deployment.

There could be some “silver lining” but we should not count on it.

Several start-ups are developing new sources of low-carbon energy that could completely reshuffle the climate equation. For instance, energy made of fusion (vs. fission in today’s nuclear plant) has the potential to provide a nearly limitless, clean source of power. Helion, backed by OpenAi Sam Altman, claims to be 5 years away from its first fusion plant and has lined up Microsoft as its first customer. Yet, we have been 5 years away from fusion for 30 years… It is not yet a practical or commercially viable source of power. The primary challenge is that it currently takes more energy to create the conditions for a fusion reaction (extremely high temperatures and pressures) than the amount of energy the reaction produces.

So what role can software play here?

Although it is very clear that software alone will never solve the climate crisis, is software useless in helping us reach net zero goals? At Samaipata, we do not think so. On the contrary we believe:

1. Software is needed to act today. Hardware can take slightly longer to realise its full potential, as it typically entails longer development and deployment cycles vs. its software counterparts. Energy efficiency software or small meters can be installed today in old buildings, whereas it might take time for cost-efficient low-carbon materials to come to market for residential construction.
2. Software can help to realise rampant optimisation potential within existing supply chains (i.e. optimisation of processes and existing assets, retrofitting, optimisation of market participation etc.). Recent advances in AI could also be fairly interesting in that regard: at the end of the day, the climate crisis is an optimisation problem: it’s about handling trade-offs. It boils down to allocating and using our resources in the most efficient way. And optimisation lies at the core of AI: it typically refers to the process of finding the best solution from all possible solutions. Large Language Models (LLMs) or other future Machine Learning paradigms will provide us with new insights.
3. Software can help us truly measure and analyze climate data & related risks, which we are still far from truly understanding. It’s the necessary step to make sure we are focusing investments and efforts into the right direction, away from “green-washing” fads. cf. the recent scandal on Verra-certified carbon credits. As an example, blockchain technology can be very useful in solving some of the traceability and transparency challenges we are facing by “tokenising” carbon credits and integrating down to the actual project.
4. Software can help us change incentives and behaviours: individual choices, culture and life style changes can be powerful source of demand-side mitigation. And here, social networks could play a key role here in promoting climate consciousness and eventually new social norms. We have seen interesting consumer fintech apps or social apps in this space leveraging the same mechanisms, which have contributed to people’s addiction to their phones, to encourage involvement in initiatives that positively impact the climate.
5. In any case, as the market matures, the hardware will provide the infrastructure but software will be what enables these systems to run efficiently and intelligently. Net-zero CO2 energy systems will couple both hardware and software layers. And conversely, software can help us make better hardware decisions.

All in all, we acknowledge software alone won’t be enough to solve this challenge; but we believe that it can play a critical role and that the opportunity to leverage its speed and quick RoI should be seized urgently. As such, as software investors at Samaipata, we are very excited to back early stage founders addressing the climate challenge. Beyond carbon credit monitoring systems and carbon accounting solutions (which have already attracted a high level of VC investments including ours as per our investment in CarbonMaps or Retraced), we are truly convinced that software can continue to “eat the world”.

Appendix

More data points and equivalence from the IPCC

As a reference to the equation described above, approx. 28 GtCO2 = 1ppm. The relationship is not linear though as C02 concentration varies depending on temperature, pression, altitude, etc.

As another reference, the link between the variation of temperature and ppm can be approximated with:

Where PPM = current concentration in CO2 and PPM0 = Pre-industrial concentration in CO2.

‍