To steward responsible AI, investors need to think like campaigners

As companies like SpaceX foreclose on conventional tools of shareholder power, investors must join the movement for tech accountability to mitigate the system-level threats of unchecked AI.  Read the full op-ed at Impact Alpha.

On June 12, 2026, SpaceX completed the largest IPO in history, raising $85 billion on a $1.78 trillion valuation and minting Elon Musk as the world’s first trillionaire. 

In the months leading up to the IPO, the financial press fixated on SpaceX’s questionable fundamentals, whether the speculative promise of orbital data centers justified its trillion-dollar valuation, and the unprofitability of its AI business, which accounts for over 90% of its staggering $28.5 trillion total addressable market. Since the IPO, SpaceX shares have plunged as investors have soured on the company’s relentless AI spending. 

To date, investor discourse has singularly focused on the financial viability of Musk’s AI ambitions. Far less attention has been paid to what Musk’s blueprint for AI development means for the broader viability of our environment, our economy, our society and our democracy.

The dirty gas behind SpaceX’s data centers

To understand those stakes, look no further than greater Memphis, home to SpaceX’s data center complex. 

While Musk has wagered the future of AI on orbital data centers, SpaceX’s near-term value hinges on increasing the scale of its AI infrastructure on Earth. The company currently operates two hyperscale data centers in Memphis (Colossus 1 and 2)  and is constructing a third (Colossus 3) across state lines in Mississippi. 

To investors, SpaceX has touted its ability to build data centers “at scale and rapid speed with lower cost” as a competitive advantage in the AI race. In reality, the source of SpaceX’s speed-to-power is neither strategic nor operational acuity, but rather a willingness to deploy substandard polluting technologies, flout environmental laws, and strip communities of the ability to resist corporate domination. 

To sidestep the time it takes to connect to the grid, SpaceX (formerly xAI) pioneered the strategy of powering its data centers with on-site mobile gas turbines. The trailer-mounted generators are among the dirtiest forms of gas-fired power, emitting not just greenhouse gases but also toxic air pollutants that lead to asthma, respiratory and cardiovascular disease, and cancer. 

SpaceX installed dozens of these turbines in and around Memphis without obtaining pre-construction permits required under the Clean Air Act (CAA). The company powered its data centers with as many as 70 unpermitted turbines for almost a year before it secured the required permits. SpaceX has maintained that under the CAA, it can run its turbines for up to twelve months without permits—a legal interpretation that was outrightly rejected by the Environmental Protection Agency earlier this year. Short-circuiting the CAA process has allowed SpaceX to stand up its data centers at record speed while avoiding environmental review, pollution control requirements, and public comment. 

SpaceX’s gas buildout is far from over. A few months ago, Musk quietly acquired one of the world’s largest mobile gas turbine fleets.  At SpaceX’s first public earnings call, he announced plans to expand the company’s compute capacity from two to ten gigawatts by the end of next year. 

The sacrifice zones of the AI boom

SpaceX’s data center expansion is the latest expression of Silicon Valley’s longstanding adage, “move fast and break things.” 

While the turbines and data centers have catapulted Musk to trillionaire status, their social and environmental costs are being borne by the working class Black communities that live in their vicinity. 

SpaceX’s makeshift gas plant in Southaven, MS, where it is operating 59 unpermitted turbines to power Colossus 2, ranks among the most polluting energy facilities in the country. Increased exposure to air toxics from these turbines could cost up to $44 million per year in the form of premature death, hospital visits, and lost productivity, with most of the burden falling on low-income Black residents in Memphis. In Boxtown, the 99% Black neighborhood abutting Colossus 1, the concentration of nitrogen dioxide has spiked to a level physicians have described as “alarming.” 

SpaceX’s data center expansion was undertaken with little to no public transparency, input, or oversight. The company has been unresponsive to community concerns, despite a groundswell of grassroots opposition and a lawsuit filed by the National Association for the Advancement of Colored People (NAACP) seeking to block the use of unpermitted turbines. A community benefits ordinance directs 25% of SpaceX’s tax revenues to Memphis neighborhoods located near the city’s data centers, but this sum—$3.3 million for 2026—is paltry compared to the billions that VC firms, Silicon Valley moguls, and corporate investors reaped from the IPO.

To impose their will on Memphis’ most vulnerable communities, Musk and SpaceX have exacted a sweeping capture of our regulatory and political systems. Emails obtained through public records requests suggest that Mississippi regulators were pressured to fast-track air quality permits for the Southaven facility. As head of the Department of Government Efficiency (DOGE), Elon Musk gutted the EPA, weakening the very agency tasked with protecting frontline communities from corporate polluters. Last year, Musk secured a $200 million deal with the Pentagon to use Grok. In June, the EPA and Department of Justice moved to intervene in the NAACP lawsuit in defense of xAI, arguing that keeping Grok online—dirty, unpermitted turbines be damned—is a “matter of paramount national security.” Taken together, these maneuvers systematically erode the democratic levers communities rely on to resist unchecked corporate power.

Memphis’ Black neighborhoods have for decades been “sacrifice zones” subjected to extractive development, environmental racism, and policies that subordinate their rights to the profit motives of corporations. Just three years before SpaceX broke ground on Colossus 1, energy companies seeking to build the crude-oil Byhalia Pipeline through Boxtown referred to the route as the “path of least resistance,” a sentiment echoed by a long line of industrial companies that have located polluting facilities in southwest Memphis. 

Today, these same neighborhoods are being transformed into the first sacrifice zones of the AI boom.

SpaceX’s assault on shareholder power

There is growing acknowledgment among institutional investors that securing long-term investment returns requires safeguarding the environmental, social, economic, and political systems that underpin long-term market performance. Unchecked AI development is a key driver of system-level threats, including climate change and environmental pollution, deepening racial and socioeconomic inequality, and democratic erosion.  Such threats are materializing in real time in places like Memphis.

For universal owners like pension funds and endowments, stewardship is critical to decreasing exposure to system-level risks. Proxy voting, engagement, and litigation are important tools for exercising shareholder voice and curbing externality-driving corporate behavior.

But SpaceX is foreclosing on these tools altogether.

By its own admission, the company’s bylaws “preclude the ability of shareholders to influence corporate matters.” SpaceX’s dual class share structure concentrates unprecedented control in the hands of Musk, who holds 82% of voting power, dictates the outcome of all shareholder votes and director elections, and has sole authority to remove himself as CEO, CTO, and chairman of the board. None of the directors responsible for board recruitment and executive compensation are independent.  SpaceX’s bylaws prohibit shareholder class actions, require that all shareholder claims be adjudicated in the Texas Business Court or in mandatory arbitration, and establish a 3% ownership threshold for shareholder proposals and derivative lawsuits—measures enabled by the company’s redomicile to Texas. 

Collectively, these tactics constitute what SOC Investment Group has characterized as the “near abandonment of the accountability mechanisms that have been associated with the term ‘corporate governance’ for the past half-century.”

SpaceX’s assault on shareholder rights has alarmed a number of institutional investors. Some, like Danish pension fund AkademikerPension, are blacklisting SpaceX entirely due to governance objections. Others, such as CALPERS and the New York City Pension Funds, have expressed concerns but have nonetheless become passive holders of the company as it has joined major stock indices.

Scaling the Musk blueprint

The Musk blueprint for AI development is fast becoming the industry’s own. 

AI companies seeking to emulate SpaceX’s speed-to-power are increasingly turning to dirty, makeshift gas generation—as evidenced by Meta’s El Paso data center, which will be powered by 800 mobile gas turbines. Big Tech is strategically siting its hyperscale facilities in poor and Black neighborhoods in the South, where weak environmental laws, lack of zoning regulations, and racialized patterns of systemic disinvestment and political exclusion make it easier for companies like SpaceX to impose their will and harder for frontline communities to resist. But perhaps the most telling endorsement of the Musk blueprint comes from Anthropic, which is renting up to $45 billion of compute from SpaceX even as it stakes its reputation on being the safe and ethical frontier AI alternative.

And while SpaceX’s brazen assault on shareholders is unprecedented, it is not the only AI company seeking to attenuate investor influence. A December stewardship report by Boston Common Asset Management found that the semiconductor, electronics, and platform companies that sit at the foundation of the AI ecosystem are largely unresponsive to investor engagement on responsible AI. Alphabet, Meta, Palantir, and Oracle all concentrate voting power in the hands of founders. Coinbase and Dell have redomiciled to Texas in order to shield themselves from shareholder claims, and Meta is contemplating a similar move. Andreessen Horowitz, one of the largest financiers of AI startups, is openly urging its portfolio companies not to incorporate in Delaware. 

In short, Big Tech is taking a cue from Musk. As the externalities of the AI boom compound, AI companies are insulating themselves from community accountability on the ground as well as shareholder accountability in the boardroom.

Finding common cause with Memphis

For institutional investors whose portfolios span entire markets and geographies, place-based data center fights like those in greater Memphis often seem too granular to merit sustained attention. 

But the fact is that system-level investors seeking to protect shareholder influence and long-term portfolio value face the same strategic threat as frontline communities fighting to safeguard public health and democratic voice: the unprecedented concentration of wealth and power among AI companies. 

In their relentless pursuit of profits and impunity, AI firms led by the likes of Musk are engineering a wholesale capture of our economy and democracy. Beyond undermining shareholder rights and building dirty data centers in our backyards, they are buying our elections and supercharging the surveillance state. AI’s loudest voices warn of AGI-induced mass unemployment, further eroding the bargaining power of workers desperate to keep their jobs. 

From permit hearings to proxy voting, tech companies and their billionaire boosters are dismantling corporate accountability mechanisms that are central to curbing AI’s negative externalities. In doing so, they are destroying the environmental, social, economic, and political systems that underpin human flourishing as well as long-term value creation.

For system-level investors, then, the stakes could not be more clear. Mitigating the system-level threats of unchecked AI requires confronting the entrenched power of Big Tech.

New stewardship tools for responsible AI

As tech companies foreclose on shareholder rights and conventional stewardship tools, system-level investors need new strategies for curbing AI’s negative externalities. This requires reimagining stewardship as a collective exercise in power-building where investors bring their capital and influence to existing campaigns that are confronting the concentrated power of AI companies.

Data center campaigns have become ground zero in the fight for AI and tech accountability. As reporter Jasmine Sun recently put it, the widespread resistance to data centers is about more than just local water, energy, and environmental impacts. In Memphis and elsewhere, data center fights are about a “profound violation of little-D democracy” where “a small group of Silicon Valley billionaires [have the] ability to impose their vision of the world onto everybody else without their consent.”

As deployers and allocators of capital, investors could bring a unique source of leverage to data center campaigns. Consider, for example, the $1.25 billion per month compute deal between SpaceX and Anthropic. For SpaceX, the Anthropic revenue helps buoy investor confidence in its lossmaking AI business. For Anthropic, compute from SpaceX is critical to meeting surging consumer demand and securing the capacity it needs ahead of its own trillion dollar IPO. Investors could condition their participation in Anthropic’s impending IPO on the company flexing its commercial relationship with SpaceX to pressure Musk into meeting frontline demands, including ending the use of dirty turbines. 

But investors cannot initiate these types of bold actions in a silo. To be effective, these strategies must be forged in partnership with community groups as part of a coordinated campaign that challenges the power of AI companies across multiple domains, from local policy and permitting to the economic and capital markets.

This is what it means for investors to think like campaigners. Rather than narrowly focusing on the bilateral relationship between shareholders and issuers, investors must see themselves as part of a broader movement, bringing their unique sources of leverage to bear in campaigns to countervail the growing power of AI companies and tech billionaires. 

The stakes extend far beyond Memphis, SpaceX, and Elon Musk. As the AI industry increasingly follows the Musk blueprint, system-level investors will need to foster new forms of collective action to ensure that companies do not continue to shift the costs of unchecked AI development onto workers, communities, and society at large. 

Next
Next

Director Voting: A Key Lever For Investor Stewardship