Originally published on Cait Brumme’s Substack on September 16, 2026.

As the battle over AI acceleration and AI deceleration produces surprising agreement amongst the fiercest frontier lab competitors, we appear to be in the middle of relearning a lesson the market thought it had already learned, and then discarded: trust compounds into real, durable value, but only for the companies that do not cash it in early.

Twenty or so years ago, that lesson was given a name: ESG. Environmental, social, and governance. (In venture, a variation was responsible venture). It was meant to encourage companies to take a broader view of what drives shareholder value over time: thoughtfulness around environmental impacts, around impacts on employees and communities, and then risk management. The acronym became shorthand for a subsequent wave of frameworks, disclosures, and scorecards that, if we are honest with ourselves, overwhelmed the initial objective: to give CEOs and boards a shared lexicon and shared measures for thinking about enterprise value creation holistically.

However, as checklists proliferated, ESG eventually became a target, first for greenwashing critics (not progressive enough), then for a broader political backlash against the values that had gotten programmed into the construct (too progressive). Many companies quietly dropped the label. Some dropped the substance along with it.

That retreat, in form and in substance, was a mistake, whether we call it ESG or not; and AI is about to force us to relearn why.

Strip away the reporting frameworks and the acronym, and the core idea was simple: a company’s license to operate is not granted by its Delaware incorporation, or by its investors, alone. It is granted by the trust it earns from the people and systems it depends on, such as employees, customers, regulators, communities. That trust sits outside the P&L, until it does not: e.g Meta’s recent $18B settlement for social media addiction for teens. It is an input to long term value, even when there are no easy answers to the real tradeoffs underneath it: innovation against risk, growth against durability, quarterly earnings against consumer safety; what a company owes its investors against what it owes everyone else its decisions touch.

What has made this core concept fragile rather than enduring is two-fold.

First, trust is a compounding variable. It takes time to convert from cost into durable advantage, and particularly in the fast-moving world of AI, the temptation to make short-term tradeoffs against it is enormous. In many ways, this is no different from the challenge corporate partners face investing in innovation with high upfront cost and an uncertain payback period.

Second, ESG was easily co-opted by interests it was not necessarily designed to serve. Somewhere in the years leading up to the backlash, it became less about that mechanism and more about the language: the statements, the commitments, the pledges. When the language became a political flashpoint, it was easy to discard the whole idea along with the words that had come to represent it.

AI is forcing the conversation back. The public backlash is demanding that tech earn its trust first, because it has yet to see enough commitment to the hard questions underneath the hype: how big is the risk to security, are companies making genuinely ethical decisions under pressure to grow, and will this actually make people’s lives better, not just more efficient.

The market incentives to date do not instill confidence in fixing this. Valuations are moving fast enough that near-term growth gets rewarded well before durability is tested, and plenty of capital is being extracted long before anyone finds out whether the underlying trust was ever earned. That dynamic is real, even where the intent behind it is not nefarious. Interestingly, a handful of vocal leaders combined with a growing public backlash against what people can actually see and control, like data centers in their own communities, seems to be converting into a potent forcing mechanism, pushing those with enormous power in this industry to at least commit to bring more care to how they do the very hard work of getting this right.

Biotech has been here before, and it offers a few insights. I had the opportunity to hear Professor Sharp at MIT describe the early controversy over recombinant DNA research in the 1970s, where public concern about a poorly understood technology was real and legitimate, even when it was not always technically precise. Scientists themselves raised the first alarm, calling for a voluntary research moratorium in 1974 before anyone forced their hand, and the following year convened at Asilomar to hash out safety guidelines in the open, with lawyers and journalists in the room alongside researchers. But the more telling moment came at the local level. When Harvard sought to build a recombinant DNA lab in Cambridge, Massachusetts, the city council did not simply defer to scientific consensus. It held public hearings, empaneled a citizen review board, and made the university’s own researchers explain their work to their neighbors before construction proceeded. As Sharp described, that type of real engagement is what earned durable public trust for the field: not just experts agreeing among themselves, but a willingness to be questioned by people with no scientific training and treat their concerns as legitimate rather than an obstacle to route around. The lesson is not that public concern about a new technology should be dismissed as fear or ignorance; or that technologists should treat public as idiots. It is that engaging with it honestly, with something closer to compassion than condescension, is what clears the way for the innovation to advance.

Here is where I gain hope in my work with those building the technology of the future: A trust crisis this visible is not only pressure on the founders and investors building this technology, but also permission for many. Many if not most of the founders we work with believe reliability and accountability are worth building for, they just have not always had the capital partners or the markets that rewarded patience over speed. A moment like this hands them both a mandate and a cover story: customers, regulators, and the public are now asking, loudly, for exactly what many builders already wanted to build: important companies, solving important problems, for the good of humanity.

---

^{Authored by Cait Brumme, CEO MassChallenge}