London Climate Action Week Recap with CO2 AI (2026)

Nearly 50,000 people poured into London last week, to attend more than 700 events during London Climate Action Week 2026. Ran under the theme "Climate cooperation in a fragmented world", it was clear that fragile geopolitics, energy security, and an accelerating regulatory landscape were pulling business leaders in different directions at once. 

The week was also, quite literally, a reminder of why we were all there. A severe heatwave swept London, forcing some events to downsize and giving every conversation about resilience a sharper edge. And yet the message across the city was consistent: the climate conversation has moved decisively from ambition to implementation. As we head toward COP31 in Antalya, leaders are less interested in pledges and more interested in what actually works — at scale, with measurable returns.

Team CO2 AI was on the ground all week, hosting and participating in several events. Here are our key takeaways.

What we heard across the week

A few themes echoed from the main stages to the side rooms:

• From promises to implementation. This was the "delivery between COPs" edition. The energy in the room was about execution, turning commitments into operational reality.
• Energy security and the clean economy are now one conversation. Investing in and innovating the clean economy was framed less as a moral choice and more as a competitiveness and resilience play for Europe.
• AI moved to center stage. Across summits and panels, artificial intelligence was the recurring thread alongside an honest reckoning with its risks, governance gaps, and the discipline needed to deploy it responsibly.

These macro themes set up the two conversations we hosted at LACW this year.

Beyond the pilot: making AI actually work for decarbonization 🤖

Most companies have run the AI pilot. Some have run several. And yet, when you ask sustainability leaders, the dashboards aren't driving decisions and the net-zero targets haven't moved. That tension was the premise for our invite-only executive dinner, co-hosted with Accenture, where we brought together sustainability and enterprise leaders for an honest conversation on enterprise AI adoption.,

We were joined by Toby Martin Hughes, Global Director – Data, Sustainability, R&D & AI at Reckitt, who shared what this journey has actually looked like inside a global organization. Over dinner, five clear patterns emerged:

1. Measurement is eating the work 📊

Leaders described spending far more time wrangling spreadsheets to measure than on the sustainability initiatives that actually move the needle. The mechanics of data have quietly become the job which is exactly the problem the work was meant to solve.

2. Internal hesitancy around AI is itself a barrier 🚧

In several organizations, caution about adopting AI is visibly limiting the impact sustainability teams can have. Some leaders are actively trying to work around their own organizations to make progress.

3. Home-grown tools can get in the way 🛠️

IT teams are often protective of the platforms they've built in-house. That creates real friction for sustainability teams that want to adopt purpose-built AI and move quickly.

4. There's a genuine appetite for guidance 🧭

Leaders don't want theory. They want to be shown the way by partners who have actually implemented AI solutions and can speak to what real deployment takes.

5. Governance is the confidence gap 🔒

There was clear recognition that output from general-purpose LLMs, without proper governance, is risky. Leaders simply don't feel comfortable putting ungoverned AI output in front of their boards or into the public domain. This is precisely why a governed, auditable, sustainability-specific AI layer matters.

The throughline from the evening: the bottleneck to AI in sustainability isn't ambition or even the technology. It's trust, governance, and the organizational confidence to move from pilot to production.

Beyond compliance: turning carbon data into real reductions 🎤

That same theme carried onto the stage at the Sustainability LIVE Leadership Summit, where our CEO and co-founder Charlotte Degot joined the Beyond Compliance: Turning Carbon Data Into Real Reductions panel in The Enterprise Theatre, alongside senior leaders including Rupert Taylor, Head of Advocacy and Industry Positioning at Symrise, and Dr. Phil Brown, Head of Sustainable Innovation at Circularise.

The premise was one we hear constantly: companies have spent a lot of time and effort building the machinery to report their emissions, but reporting a number is not the same as moving it. The panel dug into how leaders can close that gap and turn carbon data from a compliance output into a lever for actual reductions. A few points worth sharing:

• Compliance-grade data and decision-grade data are not the same. A footprint accurate enough to satisfy a disclosure requirement is often too coarse to tell you where to cut. Real reductions depend on granular, supplier- and product-level data that points to specific actions and not an annual aggregate.
• The data has to earn trust before it can drive decisions. Leaders won't redirect budgets or renegotiate supplier contracts on numbers they can't defend. Governance, auditability, and a single source of truth are what turn an emissions figure into a decision the business will actually act on.
• AI is what makes reduction work at scale. Manual measurement keeps teams stuck in the reporting cycle. Applying AI to the data foundation is what frees leaders to model scenarios, prioritize the highest-impact levers, and treat decarbonization as an optimization problem rather than a disclosure exercise.

Looking ahead — what next?

If New York Climate Week was about proving the ROI of sustainability, London was about operationalizing it. The appetite is clearly there: leaders want AI, they want measurable impact, and they want to move faster.

‍