By Solomon Assefa and Marina Rakhlin, leading the Future of Climate strategy for IBM Research
Today, IBM along with a dozen other enterprises became the inaugural members of the MIT Climate and Sustainability Consortium (MCSC). The other inaugural members of the MIT Climate and Sustainability Consortium are: Accenture, Apple, Boeing, Cargill, Dow, Inditex Lafarge Holcim, MathWorks Nexplore (Hochtief), Nexplore, Rand-Whitney Containerboard (RWCB), PepsiCo, and Verizon.
Our membership in the MCSC is yet another milestone in our commitment to lead with solutions that mitigate and adapt to climate change. Starting with the creation of the first cross-industry guidelines for voluntary corporate environmental reporting back in the early 1990s to our support for the Paris Climate Agreement in 2015 to our most recent membership on the Climate Leadership Council, our dedication to environmental sustainability has been unwavering for decades.
This past year IBM Research launched a new global initiative called the Future of Climate, with researchers across our worldwide labs focused on developing and demonstrating innovations to enable sustainable hybrid cloud, climate-smart AI platforms, and accelerated materials discovery for carbon capture.
As the leading hybrid cloud platform company, IBM is uniquely positioned to proactively address the challenge that datacentre energy consumption is expected to grow to more than 10% of the world’s electricity use by 2030. A sustainable hybrid cloud solution is one that enables clients to measure, visualise, and optimise the carbon footprint of their workloads running in the hybrid cloud. IBM Research is innovating on technologies that enable coordinated placement of containers to optimise energy efficiency and dynamic scheduling based on the availability of renewable energy. In 2020 IBM received 3,000 patents in the cloud area alone, marking our 28th year of patent leadership.
Climate-smart AI platforms are essential to enable carbon responsibility and business resiliency. As extreme weather events become more frequent and severe, businesses must understand risks, anticipate impacts, and optimise operations. AI-powered platforms for carbon footprint accounting and climate-aware supply chain optimisation all require innovations for accurate impact modelling and prediction. For instance, sophisticated queries of massive geospatial data reveal societal and environmental impact of the COVID-19 pandemic, showing for instance the significant drop in greenhouse gas emissions as a result of regional lockdowns in the spring of 2020. Physics-informed AI now enables us to build AI models that learn the physical laws of a system which can be directly applied to climate. In recent work, we have shown that integrating physics with neural network models can create robust, explainable models that aid in anticipating and reducing climate impacts. We have also trained AI models on physics-based simulated data to generate high-fidelity snapshots of the climate system and build an AI surrogate model of a complex dynamical system – in essence, fast and accurate approximations of computationally-intensive simulators.
Another pressing challenge is to design materials that capture carbon dioxide (CO2) at its emission source, since continued greenhouse gas emissions exacerbate climate change. On average, it takes at least ten years to discover a new material and bring it to market, but we simply can’t wait a decade for new materials for carbon capture to tackle the climate crisis. Thankfully, we can now combine artificial intelligence, quantum computing, and hybrid cloud to accelerate discovery. By applying deep search, AI- and quantum- enriched simulation, generative models, and cloud-based AI-driven autonomous labs, we are super-charging the scientific method to accelerate the discovery of new materials, including complex polymers and materials for carbon CO2 capture and separation. In only a few short months the team has already synthesised a prototype membrane for CO2 capture from flue gas.
Joining the MCSC gives us an opportunity to work hand-in-hand with other industry leaders to define industry roadmaps for achieving sustainability goals and identifying the scientific and technological solutions to execute on those commitments. IBM Research is embracing a model of cross-industry collaboration to tackle the complex and daunting problems posed by the changing climate.
In combining our Future of Climate scientific agenda, the AI expertise across IBM Research including focused work within the MIT-IBM Watson AI Lab and the wealth of experience from our partners and clients, our participation in the MCSC will translate into valuable scientific contributions and industry solutions. Strong collaborations with commercial partners, academia, non-profits, and government are critical for prioritising, validating, and piloting our innovations and enabling a fast path for scale and impact.