Why Net Zero Isn’t Moving Fast Enough—and What It Would Take to Reach It by the Late 2030s

A systems view of decarbonization constraints, developed with AI-assisted analysis

As AI hype accelerated, I kept hearing pointed commentary: if these systems are so powerful, why haven’t they solved climate change? For all the techno-optimism, why are we still likely to cross critical climate thresholds rather than avoid them?

Earlier this year, I decided to probe that question more seriously, and in a more grounded way.

It’s worth first explaining why I think I can evaluate AI-assisted work on net zero. I have prior experience with reduced-complexity climate modeling and SSP-based scenario logic. I previously modeled a highly speculative climate-restoration pathway for the Foundation for Climate Restoration, exploring what it would take to use large-scale CDR to return atmospheric CO₂ toward 300 ppm by 2050. That work made one thing very clear: CDR cannot be treated as a life extension for fossil. Even if removals scale, we first must achieve the fastest plausible path to the smallest possible net-zero residiual.

My early-stage product development work helped incorporate the SSP socioeconomic drivers into the analytical foundation of a science- and nature-based carbon analysis product. That included adding technology transfer and economic change, so the product could better assess whether interventions affected total factor productivity or produced local or global economic impacts.

Around early February, I began working extensively with AI, exploring what it could concretely do in the context of decarbonization. Could it map a credible pathway? Could it identify what we’re missing? What would it surface that we don’t already know? Could it reveal root causes that are easy to overlook when we think in terms of targets and ambition alone?

In the beginning, I kept checking and reprompting to verify, as rigorously as possible, the fastest institutionally plausible path to the smallest residual net zero while avoiding additional planetary-boundary exceedance and reducing existing ones where possible.

A note on rigor and editorial control: I used an AI-assisted development process, not an AI-authored one. ChatGPT functioned as a synthesis, stress-testing, and drafting tool. I read every response in full, challenged assumptions, changed bounds, redirected the analysis, checked facts, references, and numbers, and decided what did or did not enter the text. I then reread and re-edited the full playbook, repeatedly using the model to clarify assumptions, stress-test conclusions, and correct errors. Every word, claim, reference, derived figure, and chart was and remains under my editorial control.

At no point was the goal to take unmodified LLM output and run with it. The goal was to use AI to accelerate a disciplined systems-integration process: identifying missing and undervalued pieces, testing climate, energy, finance, infrastructure, fossil-retirement, justice, and planetary-boundary constraints against one another, and determining how the pathway held or failed.

After many long days and nights, aligned with climate justice and planetary-boundary risk, working on energy systems theory, the process did not converge on breakthrough technologies or unprecedented scale. And by climate justice, I mean embedding equity and harm as guardrails for the pathway and this project’s development: avoiding solutions that shift costs, risks, or damage onto frontline and vulnerable communities.

It converged on something more uncomfortable.

It converged on constraints.

Not one constraint, but a whole system of them—interlocking, interdependent processes moving in parallel, each with its own queue, each bounded by limits in capital, infrastructure, workforce, and institutional capacity. What mattered was not simply whether each piece progressed, but whether they progressed together—and fast enough to reinforce one another.

Out of that process came the 2030s Net Zero Playbook.

This work outlines the fastest institutionally plausible path resulting in the smallest feasible residual: a high-ambition path to operational net zero by roughly 2037–2038. But the key insight of the work is not the date.

It’s what determines whether that date is even possible.

Most net-zero pathways still focus on scale: build more clean energy, faster. And to be clear, that absolutely matters. Solar is scaling. Storage is scaling. In some areas, faster than expected. And yet, global progress toward net zero remains uneven—and in many cases, insufficient.

Looking more closely at that gap led to a reframing:

clean growth alone does not guarantee fossil displacement.

You can build large amounts of clean energy, and if you still fail to phase out fossil, the rest of the system will not automatically complete that transition for you.

Ultimately, what determines the timing of net zero is whether a set of interdependent systems moves together:

• electric grid infrastructure: interconnection, transmission, and distribution

• capital allocation and fossil financeability

• adequacy and reliability replacement

• workforce and project throughput

• new demand landing on ready clean systems

• biosphere restoration toward building planetary resilience

These are not independent levers. They are coupled processes. If they remain synchronized, current clean growth rates can translate into real system transition. If they do not, parallel systems emerge, which are more subtle and more problematic.

Clean energy expands, but fossil capacity does not fully exit. Reliability concerns keep legacy assets in place. Capital continues to find pathways into fossil. New demand lands on systems that are not yet ready to serve it cleanly. The result is not transition but extension of timelines, emissions, and risk.

This is why the problem looks less like a question of technological capability and more like a question of system coordination under constraint. To better understand that coordination, we need to examine its dynamics. The system does not change smoothly. It accumulates. And then, under the right conditions, parts of it begin to tip:

• capital shifts decisively toward clean

• grids operate reliably without fossil fallback

• fossil assets move from viable to structurally unfinanceable

• supply chains reorient around clean inputs

But this kind of tipping is conditional. It requires multiple processes to align within a relatively narrow window.

This does not require universal geopolitical agreement or identical national policy. It does require broad functional alignment around the four system rules and sufficient coordination among major power systems, financial regulators, development finance institutions, grid authorities, utilities, and large demand actors to make clean energy the default growth path and close fossil-fallback pathways. This is not an argument that politics does not matter; it is an argument for decomposing “politics” into specific institutional policy decisions that control fossil retirement and clean deployment. Political institutions remain necessary where regulation governs transmission, interconnection, permitting, utility incentives, fossil fuel refinancing, capacity markets, and public-risk finance. The pathway does not bypass politics; it narrows the political task to the decisions that determine whether clean buildout becomes fossil displacement.

This is where AI proved useful, not as a solution engine, but as a structuring tool. By mapping the constraint space, it became possible to understand the problem more rigorously: to identify bottlenecks, stress-test sequencing, and clarify where the pathway breaks.

Because it does break.

Transmission delays can strand clean capacity. Insufficient storage or flexibility can preserve fossil for reliability. If fossil capital remains financeable, it persists. If high-growth regions default to fossil, global timelines slip. Given enough slips, what was delayed by years could be delayed by decades.

In that sense, the challenge is not simply to accelerate everything.

It is to synchronize what matters most, when it matters most.

The window for an accelerated pathway—one that could plausibly reach net zero in the late 2030s—remains open. But it is narrow and conditional. Much of the decisive movement occurs between now and roughly 2030. After that, path dependence begins to dominate: infrastructure choices, capital allocation patterns, and institutional inertia start to lock in outcomes.

So the central question shifts. It is no longer only: can we build enough clean energy?

It becomes: can we coordinate the system quickly enough for clean energy to actually displace fossil systems across regions, sectors, and timelines? What are those constraints that determine whether that ambition translates into reality? This is what this playbook reveals: not a single answer in the conventional sense, but a clearer view of the problem. At this stage, clarity is not trivial. It shapes which efforts proceed, what gets prioritized, and how different parts of the system relate to one another. The playbook is a first pass at that level of clarity.

The use of AI in this work reflects that same framing. It was used extensively to scope, test, and refine the pathway by exploring growth rates, sequencing, and system interactions. Appendices detail the methodology, including how constraint-based reasoning was applied and how the work was bounded by planetary-risk and climate-justice considerations.

If you work in energy, finance, infrastructure, climate policy, or systems analysis, your domain is likely one of the constraint surfaces this pathway depends on. Feel free to leave me feedback; I value your perspective.

Because if this framing is even partially correct, then the question in front of us is sharper than it appears: not just how fast we can build—but whether we can move together, fast enough, for that buildout to actually complete the transition.