The general election for Wisconsin governor is now a week old. But the past seven days of debate between the race’s Republican and Democratic candidates can be paraphrased in just two lines:

“My opponent is a shill for Big Tech who wants to build data centers over Wisconinites’ farms, parks, preserves, graves — and, possibly, Lambeau Field.”

“I know you are, but what am I?”

The Republican nominee, Congressman Tom Tiffany, has been the aggressor in this argument. Tiffany has branded his Democratic rival David Crowley as “Data Center David,” spotlighting the Milwaukee County Executive’s past support for turning Wisconsin into a “data hub” for “the entire globe.” In the Republican’s telling, Crowley is a corporate stooge eager to let Wisconsin’s “lakes run dry” and its “family farms” get “paved over,” just to generate higher profits for his Silicon Valley paymasters.

Crowley has returned fire. The Democrat has insisted that every Wisconsin community should have veto power over any data center project within its territory — and that new computing campuses anywhere in the state must comply with strict regulations. Meanwhile, Crowley has argued that Tiffany is the race’s true corporate shill, noting that the member of Congress voted for tax cuts that benefit Big Tech (paid for with cuts to Medicaid and food stamps), permitting reforms that made data centers easier to build, and a ban on state and local AI regulations. “Congressman Tiffany loves data centers,” Crowley warned in a recent statement.

Meanwhile, it’s not clear how much they actually disagree on what they would do about the issue as governor. 

The Wisconsin governor’s race has quickly become the most prominent flashpoint in the national war over data centers. And it’s not hard to tell which side is winning. On Tuesday, Josh Shapiro, the moderate Democratic governor of Pennsylvania, announced a suite of measures to slow and block “predatory” data center developments, after having previously led efforts to attract them. 

If the public’s views on data centers are increasingly clear, the substantive issues remain unsettled. And the politics are more complicated than polling alone would indicate: Key constituencies in both the Democratic and Republican parties remain invested (literally and figuratively) in hyperscale development. Even as Tiffany and Crowley spar over who would do more to curb the data center menace, their actual policies quietly reflect these tensions.

How data centers became a flashpoint in Wisconsin 

On first glance, Crowley and Tiffany’s back-and-forth on data centers looks odd in multiple respects. For one thing, Republicans do not typically accuse Democrats of trying too hard to spur corporate investment — nor of doing too little to protect the environment. For another, Crowley and Tiffany seem to appear to largely agree about data center policy, which makes their emphasis on the issue sound a bit strange.

Upon closer inspection, however, Tiffany’s emphasis on the subject isn’t hard to understand.

Crowley just barely defeated socialist Assembly member Francesca Hong in the Democratic gubernatorial primary. And data center policy was among the key divides in that contest, according to Crowley himself: Hong championed a statewide moratorium on data centers, while Crowley merely favored imposing tight guardrails on their development. 

What’s more, Crowley really did tout the potential benefits of data center development in the past. And he is closely allied with sitting Democratic Governor Tony Evers, who enacted large tax subsidies for new computing campuses in the state. Thanks in part to that largesse, Wisconsin witnessed a sudden wave of hyperscale (in other words, AI) data center development over the past two years.

Thus, by staking out a more (ostensibly) hardline, anti-data center position, Tiffany hopes to fracture the Democratic coalition — and isolate Crowley from the 76 percent of Wisconsin voters who say that the server farms’ costs outweigh their benefits. 

In reality, however, the distinctions between Crowley and Tiffany’s positions on data center construction are much narrower than the latter’s attacks suggest. And to the extent that Crowley is more sympathetic to hyperscale development, this has less to do with his alleged fealty to Big Tech than his support for organized labor.

Crowley and Tiffany mostly agree about data centers

Crowley and Tiffany’s positions on data centers heavily overlap.

Neither candidate supports a statewide moratorium. Both say that local communities should have the power to veto new computing campuses; that data centers must pay the full price of their energy needs (including the costs of any new power and transmission infrastructure that their consumption necessitates); and that companies should not be allowed to enter into nondisclosure agreements with local governments that temporarily conceal negotiations over new data center developments from the public.

Likewise, as their respective attacks suggest, Crowley and Tiffany were both friendlier to the AI industry in the past, before the backlash to artificial intelligence — and the infrastructure it requires — took off. Crowley really did express a desire for Wisconsin to become a global leader in artificial intelligence. Tiffany, meanwhile, has indeed voted to cut taxes on AI companies and shield them from state and local regulations. And the Republican is aligned with an even more unabashed champion of data centers than Evers, President Donald Trump.

Further, many of the distinctions between Crowley and Tiffany’s data center platforms reflect perennial left-right divides on other issues, rather than any difference in each candidate’s degree of chumminess with Big Tech. 

For example, Crowley wants to require data center projects to finance new renewable energy installations — both so that they can meet their own power needs in a sustainable way and so that Wisconsin’s broader electric grid becomes less dependent on coal. 

Tiffany also recognizes that data center development could be leveraged to fund new renewable energy. But he sees this as a bug, not a feature: After all, new solar and wind farms could eat up Wisconsin’s precious farmland, which Tiffany has sworn to protect. The Republican instead prefers for new data centers to run on coal, thereby propping up that embattled industry.

Crowley is more pro-data center in one way

All this said, Tiffany’s platform is genuinely less pro-data center than Crowley’s in one respect: The Republican has promised to end all taxpayer subsidies for data centers in Wisconsin. 

Crowley’s position is comparatively nuanced. In an interview with the Milwaukee Journal Sentinel in July, the Democrat said, “I do not believe Wisconsin should automatically roll back all data center tax incentives, but I do believe we need a more rigorous review of whether taxpayers are receiving sufficient value for the incentives being provided.”

In response to Tiffany’s recent attacks, Crowley has embraced far more adversarial rhetoric towards the tech industry in general and the AI sector in particular. And he has called for imposing a set of requirements on new data center development so numerous and profound, it is unclear whether many companies would be interested in pursuing new projects under his leadership, subsidies or no subsidies. 

Nevertheless, as of this writing, Crowley has yet to abandon his July position on tax incentives. Instead, he has countered Tiffany’s charges on that front by noting that, in supporting Donald Trump’s tax bill, his Republican opponent effectively “signed over $70 billion in tax handouts to the same companies now steamrolling over our local communities with data centers.” 

This is a fair hit. But it’s also a little evasive. How high the corporate tax rate should be federally — and whether Wisconsin should offer tax breaks to data center developments — are two different questions. 

And Wisconsin’s subsidies to hyperscalers are not small. Under Evers, the state exempted large data center projects from sales taxes. This means that a data center developer need not pay taxes — to either the state or localities — on construction materials, IT equipment, on-site power infrastructure, cooling systems, or electricity. 

Critically, these exemptions do not sunset: Tech companies don’t just avoid sales taxes when initially building their data centers. Rather, they pay no taxes on their facilities’ electricity consumption forever, while also avoiding any levies when updating their servers or other equipment.

All this adds up. According to a report from Wisconsin’s Legislative Fiscal Bureau, the state is poised to forgo $1.5 billion in sales tax revenue during the construction of certified data center projects, and then another $369 million annually after the facilities are fully operating. This is an especially sensitive issue in Wisconsin, where voters are still upset over an effort to lure tech manufacturer Foxconn to the state in Trump’s first term, a costly gambit that failed to deliver on its promised benefits.

To be clear, the Fiscal Bureau’s projections do not mean that data centers are likely to subtract billions of dollars from Wisconsin’s state and local budgets. Even if the hyperscale facilities are exempt from sales levies they must still pay property taxes. And their owners could owe corporate taxes while their construction workers and permanent employees will pay income taxes. 

Their spending also generates business — and thus, income tax obligations — for other Wisconsin enterprises, such as utilities. On net, the projects will almost certainly increase the amount of revenue available to Wisconsin state and local governments. The debate is over whether the projects could have generated even more revenue, in the absence of subsidies — or whether these developments would have gone to a different state, had Wisconsin not made their funders such a sweetheart deal.

The evidence on that question is mixed. On the one hand, tax subsidies of some kind were plausibly necessary to spur hyperscale data center investment in Wisconsin. At present, 38 states provide tax incentives to data center developers. 

On the other hand, Wisconsin’s subsidies are unusually generous. States commonly spare servers and capital equipment from sales taxes, but many do not extend that exemption to electricity. And the indefinite nature of Wisconsin’s subsidies is also far from standard. In nearby Illinois, data center projects enjoy a sales tax exemption for only 20 years. 

Meanwhile, Wisconsin has some structural assets that might make it appealing to tech companies, even if it offered far lower subsidies. Due to its proximity to the Great Lakes, Wisconsin boasts abundant water. And it also has relatively cool temperatures, which limit hyperscalers’ cooling costs. 

Crowley has not ruled out dramatically scaling back the state’s subsidies. But unlike Tiffany, he has thus far declined to endorse abolishing those subsidies in their entirety.

In Wisconsin, banning data centers is not the pro-union position

At this point, you might be asking why Crowley doesn’t simply go “full populist” on data centers. Why give Tiffany any room to get to the Democratic Party’s left on this issue? Why not shore up your left flank by simply calling for a statewide moratorium — or at a minimum, for ending all data center tax breaks? After all, roughly three-quarters of Wisconsinites think the facilities are net harmful to the state. 

In Tiffany’s telling, the answer to these questions is simple: Crowley is a bought-off puppet of the AI industry. But this is almost certainly wrong. 

In truth, Crowley’s refusal to fully disavow data center construction likely reflects his ties to labor, not capital. 

In his attack ad, Tiffany alleges that Crowley’s campaign is funded by “the people building” data centers and then flashes this image:

Tiffany’s claim is true — but in a misleadingly literal sense: The figure cited appears to be a reference to Crowley’s support from Wisconsin’s building trades unions. As of August 13, unions representing Wisconsin electricians, pipefitters, carpenters, construction workers, and other laborers had collectively contributed $412,000 to Crowley’s campaign. The Democrat has received no comparable support from corporate PACs of any kind, much less AI industry-affiliated ones. 

The building trades are strongly supportive of data center construction — and for good reason. A data center serving OpenAI and Oracle in Port Washington, Wisconsin is being built through an explicit partnership with the Wisconsin Building Trades Council. 

Separately, the Democrat has evinced concern for municipalities that do desire data center development, saying that Wisconsin can protect its ratepayers and environment “while also not preempting communities that have done their due diligence to allow for different types of economic development investments.”

For some Wisconsin communities, the fiscal benefits of data center development can be profound. 

Consider Port Washington: In recent decades, the city’s manufacturing base shrank — and its business tax revenues declined with it. As a result, homeowners shouldered more and more of the burden of financing local government. But its new hyperscale data center promises to dramatically change its finances. According to the city’s Mayor Ted Neitzke, the project is expected to add between $1.3 and $1.8 billion in property value to a municipality whose entire existing valuation is scarcely over $1 billion. 

By doubling the tax base, the data center could finance property tax cuts or public service expansions for Port Washington residents. Many of Wisconsin’s other post-industrial towns face similar challenges — and thus, could reap similar benefits from data center development. This may be why neither Crowley nor Tiffany have felt comfortable endorsing a moratorium, despite their increasingly populist rhetoric about data centers and AI.

For most voters, data centers aren’t central

Add it up, and you have two candidates who are balancing the interests of the data center boom’s material beneficiaries against its growing political toxicity — and giving more and more weight to the latter. 

It is unclear whether data centers will remain at the forefront of governors’ race as the campaign enters its homestretch. Although hyperscale facilities are increasingly unpopular in Wisconsin, they also are not a top-tier concern for the vast majority of the state’s voters. In a July poll from Marquette Law School, only 5 percent of Wisconsin voters named “data centers” as the state’s most important issue, putting it behind inflation, “illegal immigration,” health insurance, property taxes, “jobs and the economy,” and housing affordability.

Nevertheless, the fact that Crowley and Tiffany are sparring over who dislikes data centers the most — months after both expressed positive sentiments about the facilities — shows which way the wind is blowing. 

In recent months, two ideas that were once on the fringes of Democratic policy thinking — cutting off financial support and stopping arms transfers to Israel — have gone mainstream.

Nearly half of House Democrats voted to end aid to Israel last month, while roughly 40 percent of the caucus has called for blocking the sale of offensive weapons to the Israeli government until it persuades Congress that it will not use American bombs to perpetrate war crimes. 

Many of the party’s 2028 hopefuls have also embraced these general concepts. Rahm Emanuel, the former Chicago mayor and consummate centrist, recently suggested that America should end aid to Israel and condition arms sales on its compliance with US law. Progressive Sen. Chris Van Hollen has gone several steps further, saying that Israel should be barred from accessing US-made offensive weapons until it “agrees to a time-bound plan to end the occupation and enact a two-state solution.”

If recent elections are any indication, these ideas will only grow more prominent in the coming years. Over the past two months, a slew of Democratic candidates — often with establishment backing and largely progressive voting records — lost to left-wing rivals who promised to take a more adversarial posture towards Israel. Rep. Haley Stevens’s apparent defeat at the hands of the staunchly pro-Palestinian Abdul El-Sayed in this week’s Democratic Senate primary is the latest and perhaps most consequential example of this trend.

The Democratic Party’s burgeoning support for restricting aid and arms transfers to Israel is well-founded. In prosecuting its war with Hamas, the Israeli government has subjected Gaza’s civilian population to grotesque and illegal forms of collective punishment (if not genocide). Meanwhile, in the occupied West Bank, Israel is expanding illegal settlements and allowing Jewish settlers to violently dispossess Palestinian residents. In this context, supplying Israel with money and weapons is contrary to both America’s moral obligations and security interests. It is wrong to subsidize or provision any foreign military’s assaults on human rights. And doing so undermines global respect for both the United States and the concept of international law.

For many Democrats, however, the purpose of cutting off aid and weapons to Israel is not merely to end America’s complicity in its crimes. Rather, the objective is to effect a radical change in Israeli policy — and ultimately, a peaceful resolution of the Israel-Palestine conflict.

Such Democrats believe that the Israeli government will not seriously pursue a two-state solution (much less, a binational democratic state) unless it is forced to do so. Fortunately, in this view, Israel’s reliance on American aid and military technology provides the US with potent leverage over its internal policies: If America ceases to provide the Israeli military with cash and offensive weapons, then Tel Aviv will eventually have little choice but to start upholding its end of the two-state bargain.

But this is probably too optimistic. Ending aid and conditioning arms sales may be preconditions for realigning Israeli policy. But coercing its government into an acceptance of Palestinian statehood would likely require measures far more sweeping — and controversial — than those that congressional Democrats are presently entertaining.

Israel doesn’t need our money 

Of all the Democratic Party’s proposals for disciplining Israel, ending US aid may be the most politically palatable — yet least effective.

The notion that the United States should stop giving Israel $3.8 billion each year has some appeal on the left, right, and center. Which is not surprising: One doesn’t need to be passionately pro-Palestinian to be skeptical of America giving billions in aid to a rich, foreign country. In his Senate race, El-Sayed tried to tap into this isolationist sentiment, describing his plans for slashing aid to Israel — and increasing benefits for US citizens — as an “America First” agenda.

Unfortunately, ending aid would have little impact on Israeli policy. And for the very reason that such assistance is hard to justify, even to those who care little about the Palestinians: Israel is a wealthy nation that can fund its own military.

In 2025, Israel spent about 7.9 percent of its GDP on the military, not counting its use of American aid. To sustain its 2025 level of defense funding — without American aid — Israel would need to dedicate 8.5 percent of GDP to its military. This increase might be slightly inconvenient for Israeli taxpayers. But it’s clearly doable. 

And Israeli Prime Minister Benjamin Netanyahu wants to do it. 

Or so he told 60 Minutes in May, saying that he would like to “draw down to zero the American financial support, the financial component of the military cooperation that we have.” In the view of some Israeli analysts, forgoing US aid would give Israel a freer hand to pursue its security objectives as it sees fit without worrying about a backlash from politicians thousands of miles away. The American Enterprise Institute’s Daniel J. Samet and former IDF Maj. Raphael BenLevi spelled out the logic of this position back in 2023, writing, “As long as aid continues, the temptation to dictate Israel’s policies will be too strong for Washington to resist, while Israeli officials will be inclined to seek approval for every action they take.”

This doesn’t mean that the US should continue funding Israel’s military; revoking such aid could improve America’s global image while saving taxpayers a little money. By itself, however, canceling all military grants to Israel would not force its government to make any real changes in its security strategies, much less to accept Palestinian statehood.

Blocking bombs would have a bigger impact 

Although Israel does not need America’s dollars, it does currently require our weaponry. Between 2021 and 2025, the United States supplied 68 percent of Israel’s imported major arms.

Israel can domestically manufacture many of these weapons — including drones, missiles, air-defense systems, armored vehicles, and aerial, artillery, and tank munitions. But its arms industry is not capable of churning all of these out at the rate required by its intensive, multifront wars since October 7, 2023. 

Meanwhile, Israel is wholly reliant on America for its combat aircraft fleet. They don’t make F-15s, F-16s, or F-35s in Haifa. Critically, this doesn’t just leave Israel dependent on the US for new fighter jets but also for the maintenance of its existing ones. 

“The F-35 is a system that is deeply reliant on just-in-time delivery of spare parts,” Josh Paul, a former Biden State Department official who resigned in protest of American arms transfers to Israel, told me. “And were the US to say we’re not going to provide you those spare parts, it wouldn’t be more than one or two months before Israel’s F-35s would not fly, certainly at the current operational paces.”

Thus, conditionally freezing sales of US military equipment to Israel would place real constraints on its martial capabilities. Although, the scale of those constraints depends a lot on the details of US policy. 

The most prominent proposal for an arms embargo in Congress, the Block the Bombs Act, would freeze the transfer of select classes of weaponry, including bunker-busting bombs and precision missiles. This policy might force Israel to scale back or delay certain aerial campaigns. In the long run, however, Israel has domestic substitutes for these weapons. 

By contrast, were the US to choke off military supplies more broadly — leaving the Israeli air force bereft of fighter jets and component parts — the implications for Israeli security strategy would be far more profound.

In either scenario, however, it seems unlikely that the Israeli government would respond to an embargo by agreeing to cede territory or autonomy to the Palestinians, in service of a two-state solution. 

The reasons for this are twofold. First, although an embargo would substantially diminish Israel’s military capabilities, those capacities would remain formidable. As already noted, Israel domestically produces a great deal of military hardware — so much, it actually exported more arms than all but six countries between 2021 and 2025. 

What’s more, the Israeli government has already launched an effort to achieve “munitions independence.” This entails a $110 billion investment into its arms industry, including initiatives to increase Israeli production of aerial munitions, heavy bombs, and artillery shells.

At the same time, Israel is also developing a plan for expanding its domestic raw materials industry, so that it enjoys sovereignty over the inputs necessary for arms production.

Given all this, even without access to American weapons, Israel would still be able to militarily dominate the West Bank and Gaza. 

“Israel’s dependence on US weapons is more about its position in the region; its capacity to deter Iran and so forth, not in the Palestinian context,” Paul said. “Cut off Israel’s access to US weapons, and it would still have overwhelming military superiority over any Palestinian resistance group.”

Further, in the face of a US arms embargo, Israel would not only remain capable of sustaining its subjugation of the Palestinians; it would also likely retain the will to do so. 

Israeli opinion on politics in general — and the Palestinian issue in particular — has been drifting rightward for decades. In a 2025 Gallup poll, 27 percent of Israelis supported a two-state solution while 63 percent opposed one. And opposition to Palestinian statehood is even more pronounced among Jewish Israelis, who dominate the nation’s politics. In a Pew survey from last year, only 16 percent of Israeli Jews agreed that peaceful coexistence with a Palestinian state was possible — down from 46 percent in 2013. Support for a binational, democratic state of Israel-Palestine — the “one-state solution” favored by many on the American left — is all-but non-existent among Israeli Jews, with only 1 percent backing the policy in a 2025 poll.

Israeli political leaders share its electorate’s opposition to Palestinian statehood. Netanyahu’s leading rival in Israel’s upcoming election, former IDF chief Gadi Eisenkot, recently told an interviewer, “I have never said two states for two peoples, because I understand the depth of this religious-national-social conflict, and anyone who talks about ‘peace now’ and two states doesn’t understand the conflict.”

Israeli attitudes toward Palestinian sovereignty do shift with events; after the horrors of Hamas’s October 7, 2023, attacks, skepticism of a two-state solution among Jewish Israelis spiked. It is therefore conceivable that Israelis might become more open to Palestinian statehood, were the people of Gaza and the West Bank to unify behind conciliatory political leaders, who were prepared to accept the existence of a Jewish state within Israel’s legal borders. But that is an unlikely prospect in the near or medium term — and one that Israel’s criminality makes evermore remote. 

Absent such a development, a US arms embargo is unlikely to realign Israeli politics fundamentally. Historically, when great powers have tried to force weaker states to shift their policies through coercion, the latter often resist, even at a high cost. Iran’s refusal to submit to US demands on its regional policies and weapons programs, amid years of devastating sanctions and months of bombardment, may be today’s most salient example. 

The Iranian regime’s recalcitrance may partly reflect its authoritarian character. But democratic states have also routinely bucked the strong-arming of foreign powers. For example, in the 1980s, New Zealand barred the United States from docking nuclear-powered or armed ships at its ports. The US tried to coerce the Kiwis into reversing this policy by suspending America’s commitment to defend New Zealand from military attack in 1986. But New Zealanders remained undaunted; indeed, support for the anti-nuclear policy grew over the ensuing years.

What getting tough on Israel would really look like

This is not to say that no amount of sustained American coercion could shift Israel’s political calculus on its conflict with the Palestinians. But the scale and duration of US pressure would likely need to be much greater than what congressional Democrats are presently contemplating. 

Specifically, such a campaign would likely require an arms embargo more far-reaching than the one proposed by the Block the Bombs Act, along with sweeping economic sanctions, and the withdrawal of American diplomatic protection at the United Nations. 

The Biden administration created the architecture for such a sanctions policy when it froze the US-held assets and visas of Israeli settlers who’d perpetrated acts of violence in the West Bank, along with those who had materially assisted them. These sanctions could be expanded to cover all involved in settlement expansion, including construction companies, banks, financial institutions and government agencies that facilitate the colonization of Palestinian territory. Should this prove insufficient to change Israeli policy, America could apply sanctions to those populating and serving already-existing Israeli settlements.

For reasons previously noted, Israel’s first response to all of this is liable to be resistance. To the extent that these measures were politically controversial within the US — and opposed by the Republican Party — Israel would probably seek to undermine and wait out a Democratic government that implemented them, rather than making substantial concessions to Palestinian autonomy. 

America’s pressure campaign may therefore need to be sustained for a prolonged period — and enjoy a measure of bipartisan support — before having any transformative impact on Israeli policy. 

This could be challenging. As El-Sayed’s “America First” message indicates, many Americans favor cutting off aid to Israel because they want to disengage from Middle Eastern conflicts and focus attention on domestic problems. But the case for blocking arms sales, sanctioning large financial institutions — and, more broadly, expending political capital and energy on a protracted fight over a foreign country’s internal policies — cannot rest on mere isolationism.  

An effective pressure campaign would also require long-term resolve across multiple administrations, who may be tempted to re-engage with Israel if the US runs into new challenges abroad in which Israeli intelligence, military assets, diplomatic ties, or technology might be useful. When Joe Biden took office, he initially sought to treat Saudi Arabia as a “pariah” over the regime’s appalling record on human rights, only to come calling to its leaders in person, in order to secure its cooperation with regional issues and the global energy supply. Since then, President Trump has drawn the two nations even closer.  

In any case, for now, the idea of not merely cutting off aid and select “offensive” weapons to Israel — but rather, forbidding the sale of all military goods to its government, sanctioning all involved in settling the occupied West Bank, and withdrawing US diplomatic support at the UN — remains too radical an agenda for most congressional progressives, much less their Republican colleagues.

Even if Democrats can’t change Israeli policy, they can reduce American complicity

None of this means that ending US aid to Israel — or restricting the sale of offensive weapons to its military — aren’t worthy objectives. Such measures could constrain Israel’s capacity to bomb civilian infrastructure and launch wars of choice. They just might not be sufficient to singlehandedly overwhelm Israeli resistance to Palestinian sovereignty.

Regardless, the congressional left’s Israel agenda would improve America’s moral health and global image. And this should arguably be US lawmakers’ priority. 

“The United States played an enormous role in creating Israel’s conventional military advantage as a player in the Middle East,” Daryl Kimball, executive director of the nonprofit Arms Control Association, told me. “And in the face of the devastation that the Israeli government has wrought in Gaza and now on the West Bank — which involves, by any objective measure, serious war crimes — the United States needs to exert its leverage, so as not to be complicit in such behavior. Whether that changes Israeli policy regarding a two-state solution, or its ongoing treatment of Palestinians in occupied territories, is ultimately a separate question.”

New Yorkers will soon enjoy a 30 percent discount on their meat, seafood, milk, and bread — so long as they shop at a city-owned store.

Mayor Zohran Mamdani touted such bargains this week at a press conference detailing his plans to launch five public grocery stores, in a bid to lower New Yorkers’ food costs. Under the proposal, the city would own each supermarket and dictate its pricing and labor practices, while private grocery companies would handle the day-to-day operations.

Mamdani’s proposal has been getting people worked up from the moment he unveiled it on the campaign trail for his 2025 mayoral run. For many progressives, state-owned supermarkets embody one of their movement’s highest ideals — the prioritization of public needs over private profits. For conservatives, meanwhile, “Mamdani Marts” represent an attack on free enterprise (if not, the first step on the road to Soviet breadlines).

While stimulating, these grand philosophical arguments have often overshadowed more banal but pressing questions: Would public grocery stores do more to advance Mamdani’s stated goals than other things the city could do with the same money? Has New York’s mayor discovered a way to sell groceries more efficiently than Costco does — or an approach to nutritional relief more effective than simply giving poor people more cash and food?

The answer to both of these questions appears to be “no.” And that should concern more than just New Yorkers. Mamdani’s vision is already inspiring imitators, with San Francisco and Boston both exploring their own public grocery store proposals. And in national discourse, the mayor’s policy is widely understood as an experiment — one testing a bold new theory of how governments can make food more affordable.

Unfortunately, although public grocers have merit in some circumstances, the theory underlying Mamdani’s specific plan is not merely unproven but incoherent.

Grocery socialism makes more sense for small-town Republicans

Public grocery stores have historically aimed to solve a problem that New York City does not have — a total absence of private supermarkets.

In 2018, the small town of Baldwin, Florida lost its only grocer. With only 1,600 residents — and a median income of $44,271 — the community no longer generated enough business to be worth the local IGA’s while. And no other national chain rushed in to fill the gap. 

So, the municipality opened its own grocery store, the Baldwin Market.

This attracted national media attention. And not without reason. A deeply conservative small-town trying its hand at socialism was a fun story. And it turned out that Baldwin wasn’t an aberration: Other aging rural communities in red America had turned to government grocers once private supermarkets had left them in the lurch.

Progressive policy thinkers took inspiration from these examples. In the ensuing years, proposals for using public grocery stores to combat urban food deserts — impoverished neighborhoods without convenient and affordable supermarkets — began to circulate. The merits of such plans can be debated. But their logic was straightforward: Where the market fails to provide residents with healthy food options, the government should step in. 

When post-COVID inflation sowed outrage over food prices, however, some on the left reconceived the purpose of public supermarkets: In their telling, such stores weren’t merely a means of eliminating food deserts, but also a way to make groceries more affordable.

Mamdani’s plan is principally animated by the latter goal. With more than 1,100 grocery stores and 10,000 bodegas, most of New York City is well-provisioned by private food vendors. And although parts of the municipality are under-served, the mayor is not actually concentrating his stores in such areas; his planned East Harlem location lies just blocks from an Aldi, Costco, and many other grocers. 

In a report detailing its policy, the mayor’s office makes clear that its primary aim is to drive down New Yorkers’ food bills. Yet there’s little reason to think that public grocery stores are a cost-effective way to do that.

New York City probably won’t provide groceries more efficiently than Costco

Without question, a government store can make groceries more affordable for its own customers: All it has to do is charge below-market prices.

And yet, if that store manages this feat by operating at a loss, then its shoppers’ savings will come at the broader city’s expense.

That might not be too troubling, if one pictures rich taxpayers footing the bill. But for a city like New York — which faces a structural budget deficit — revenue is a scarce resource. A tax dollar spent on public grocery stores is one that can’t be spent on nutritional assistance for low-income people. And all else equal, the latter will do more to enhance affordability: If you have a limited pool of food subsidies — and want to minimize the number of New Yorkers who can’t afford groceries — then you should spend your funds on the poor, not on whoever happens to show up at your city-owned store. 

Thus, for Mamdani’s supermarkets to be cost-effective, they can’t just subsidize low prices with taxpayer money. Rather, they need to deliver groceries more efficiently than private retailers do.

Public ownership can unlock efficiencies in certain contexts. For example, America’s private health insurance system generates massive administrative redundancies. By one estimate, if the US replaced its sprawling insurance industry with a single government payer, it could cut our health system’s annual administrative costs by $500 billion. In other words, a Medicare For All program could theoretically provide more healthcare-per-dollar than the current system does, by eliminating excess bureaucracy.

But there are no comparable inefficiencies in the grocery sector. True, a government retailer could generate some savings by declining to take a profit. But margins in the industry are slim; the average profit rate among food retailers was just 2.1 percent in 2025. Therefore, Mamdani’s stores can’t finance 30 percent discounts merely by dispensing with shareholder returns. 

Meanwhile, the city actually wants its stores to be less cost-efficient than private grocers in some respects. According to its policy brief, the government grocers will pay their workers higher wages than most retailers and may also favor “local and regional suppliers,” rather than automatically contracting with whichever food producers offer the best rate. 

The city suggests that its stores will have some cost advantages: Unlike private competitors, they will pay no rent or property taxes on their facilities. But these are merely additional subsidies, not actual efficiencies: By providing free real estate to its grocers, the city is forgoing revenue that it could otherwise collect and spend on nutritional assistance or other public goods. 

It is not yet clear how much money the city is prepared to lose each year on these stores. If Mamdani’s celebrity — and heavily advertised bargains — lure large masses of people to his markets, their steeply discounted goods will quickly sell out. In that circumstance, to avoid long stretches with empty shelves, or some complex rationing system, the city would need to rapidly restock money-losing items, compounding the stores’ operating losses.  

In theory, there is one way that government stores could benefit consumers throughout a city without being especially efficient enterprises in themselves: They could force other retailers to cut prices by accepting lower profits or discovering new efficiencies. But with grocery margins already thin, squeezing private markets further could lead to closures, thereby reducing shoppers’ options. Further, Mamdani himself insists that his stores will go out of their way to avoid harming private competitors (which, in New York, consist primarily of the sorts of small businesses that Mamdani has promised to help). 

In any case, five stores won’t substantially impact pricing citywide. And so long as each government grocer operates at a loss, scaling up the program will only deepen its costs — and thus, the tradeoff between funding public supermarkets and anti-hunger programs.

Aldi for all

At present, Mamdani’s grocery store experiment looks fairly cheap. The city estimates the stores will require $70 million in capital costs. As already noted, it’s unclear precisely how large each store’s annual operating budget will be. But even if each supermarket loses many millions each year, it still won’t make that big a difference, in the context of the city’s $125 billion budget.

Still, the opportunity costs of Mamdani’s policy are real. New York City currently plans to spend $3.1 million next year on a program called Get The Good Stuff (GTGS), which essentially gives SNAP recipients up to $10 off each time they purchase fruits and vegetables from 25 participating grocery stores. With the money slated for its public supermarkets, New York could dramatically expand the program, while also investing more money into its food banks. Alternatively, the city could modestly increase cash transfers to its lowest-income residents. 

To be sure, none of that would offer much benefit to middle-class shoppers. And the mayor is quite reasonably concerned with increasing grocery affordability for the typical New Yorker. But the city can advance that aim — while actually increasing its revenue — by easing zoning and permitting rules that currently make it difficult for large, low-cost retailers like Costco and Walmart to operate in many parts of the city. Happily, to Mamdani’s great credit, his administration’s paper on public grocery stores does briefly mention the need for permitting reforms.

Maybe public grocery stores are an end in themselves

All this said, there may still be a reasonable case for Mamdani’s stores or others like them. For example, a city’s residents might come to see such retailers as a kind of attraction, public amenity, or community space, akin to a park or library.

Further, public grocers could theoretically increase the public’s faith in the government. While the grocers are unlikely to be efficient in reality, they may appear to be. After all, their low prices will be far more visible to consumers than their operating costs. If Americans come to associate the public sector with cheap, ostensibly well-run supermarkets — rather than dreary DMV lines — that could aid the left’s broader efforts to expand the government’s remit. (Although, if the public grocers end up being characterized by overcrowding and empty shelves, the policy could further erode Americans’ confidence in the public sector.)

As a means of maximizing affordability, however, Mamdani’s program makes little sense. It will mostly just transfer income from the city’s broad population to the small subset of New Yorkers who happen to live near — or work at — one of its five stores.

Where private markets fail to provide any grocery options, public stores may be the best way for municipalities to meet their residents’ nutritional needs. But in other contexts, governments can make a bigger difference by simply putting more money in poor people’s pockets.

The AI industry’s investors and critics don’t agree on much. But many in each camp share at least one basic conviction: America’s top labs are about to make a killing. 

Capital markets have signaled their faith in Anthropic and OpenAI’s impending hyper-profitability, valuing each at nearly $1 trillion. Many of Silicon Valley’s progressive adversaries also expect the labs to grow filthy rich but fear the implications, warning that AI-induced automation could transfer vast sums of money from ordinary workers to a handful of giant tech companies. Sen. Bernie Sanders’s call for nationalizing the top AI labs rests partly on that concern. 

But recent advances in Chinese AI call all of this into question.

Over the past two months, Chinese companies have released three AI models that are nearly as powerful as America’s frontier systems — and radically less expensive. 

In June, Beijing’s Z.ai debuted a model that performed nearly as well as Claude and ChatGPT’s second-tier systems on independent benchmarks. Weeks later, another Chinese firm, Moonshot, unveiled “Kimi K3,” a model that allegedly outperforms all of its American rivals except for the very latest versions of Claude and ChatGPT. Finally, just days ago, Alibaba launched a preview of Qwen3.8 Max, which purportedly outclasses even OpenAI’s most advanced systems, while trailing only Claude’s Fable in its capabilities. (Disclosure: Vox Media is one of several publishers that have signed partnership agreements with OpenAI. Our reporting remains editorially independent.)

These developments don’t merely threaten America’s AI giants with stiffer competition in the race for superintelligence. Rather, they raise a more harrowing prospect: that the AI race’s ultimate rewards will be far smaller than anticipated. In a world where new advances can regularly be leapfrogged by cheaper upstarts, hoarding the technology — and its profits — will be harder for any one company to do.

In other words, building a machine God might not be as lucrative as it’s cracked up to be. AI, it turns out, may “want to be free.”

How AI was supposed to pay off

To see how China’s new models threaten Anthropic’s profit expectations, we must first examine why those expectations have been so high.

This is not entirely self-evident. After all, AI labs aren’t much like the hyper-profitable tech giants of the 2010s. Facebook and Airbrb were relatively capital-light businesses with ultra-low marginal costs (adding a profile to Facebook or listing to Airbnb costs the companies virtually nothing). And once each gained a foothold in their respective markets, network effects enabled them to retain formidable positions without needing to constantly upgrade their products.

Building a state-of-the-art AI company is a much more involved — and astronomically more expensive — endeavor. To get to the frontier, Anthropic and OpenAI have sunk (at least) tens of billions into semiconductors, data centers, power plants, and other capital investments. Staying at the cutting-edge, meanwhile, compels them to perpetually churn out evermore costly models.

To put a new Claude model through its initial training — in which it spends months digesting the internet and sussing out statistical patterns within its text — can now cost hundreds of millions of dollars. And such foundational computation is only the beginning. A truly superlative model requires several additional months of fine-tuning. Armies of contracted experts — such as computer scientists, physicians, and mathematicians — tutor the models, grading their answers and guiding them towards better ones. Then the AI systems complete millions of rounds of practice, in which they learn through trial and error how to solve countless problems. This arduous process, known as “post-training,” compounds the costs of a single model’s development. 

All of which raises the question: Why would investors expect businesses with a cost-structure this challenging to be not merely profitable, but massively so?

There are (at least) two answers. The first (and most obvious) is that the market for superintelligent machines is liable to be vast. Frontier AI systems promise to reduce costs and improve performance in myriad white-collar sectors. And Anthropic’s soaring revenues indicate that firms do, in fact, find Claude useful. A company like AirBnB has earned billions by revolutionizing a single industry; imagine then what a technology that remade virtually all industries might be worth.

Of course, plenty of technologies are valuable but not massively profitable to produce. After all, in well-functioning markets, competition should eventually erode individual firms’ margins, even if the underlying technology continues generating huge value. 

But this is where the second answer comes in: Frontier labs’ immense costs are a burden, but they’re also a safeguard against competition — or, in industry parlance, a “moat.”

Startups may be able to afford to build or acquire more rudimentary models, many of which are “open source.” But, the thinking goes, they won’t be able to deliver Claude Fable-level performance without raising giant amounts of capital. And what investors will be willing to pour hundreds of billions into an AI pipsqueak that’s light-years behind Google, Anthropic, and OpenAI?

Alas, the Chinese AI labs’ rapid progress — and the way it was achieved — suggest that Anthropic’s moat may be shallower than previously thought.

How Moonshot swam Anthropic’s moat

The existence of powerful, Chinese AI systems is neither new nor surprising. Xi Jinping’s government has made vying for global AI dominance a key economic goal. And China’s DeepSeek, which also has stunned US companies with its lower-cost competitive models, surpassed ChatGPT as the most-downloaded free iPhone app more than a year ago.

The latest models, however, have dramatically narrowed the gap in capabilities between frontier American systems and their Chinese rivals. Just as critically, they’ve done so in a manner that other, relatively underfunded AI upstarts might be able to emulate.

Alibaba and Moonshot needed to invest massive resources to train their base models. But they allegedly found a low-cost way to refine those models into near-frontier systems: Just ask Claude.

Or, more specifically: Engage Claude in 16 million conversations, using 24,000 fake accounts. In each of those exchanges, ask the model to not only answer countless difficult questions but also, walk you through its reasoning, step by step. Then take all of this data and feed it into your own model as study material, training it to respond to the world’s most challenging queries as Claude would. 

Through this process — known as “distillation” — an AI lab can replicate virtually all of a frontier model’s capacities, without sinking vast sums into human experts and post-training computing runs. 

China’s AI labs have not admitted to using distillation. But OpenAI and Anthropic both reportedly uncovered Chinese distillation attempts earlier this year. And some of the new models appear to display tell-tale signs of distillation in conversations with ordinary users; Kimi K3 has routinely identified itself as “Claude.”

Chinese AI companies are hardly alone in using distillation to catch up with frontier labs. Earlier this year, Elon Musk admitted in court that xAI enhanced Grok’s capabilities by running distillation techniques on Claude and ChatGPT. Nonetheless, China’s latest models appear to demonstrate that distillation can help take a second-tier model to the frontier’s threshold.

America’s frontier labs have tried to defend themselves against such imitators. But this is technically difficult when distillers can assemble massive networks of bots, each asking an inconspicuous number of questions. And legally, it is difficult for America’s AI giants to argue that distillers are stealing their intellectual property. After all, in a sense, China’s copycats are merely doing to Anthropic and OpenAI what those companies did to journalists, coders, lawyers and other specialists: Feeding their public-facing outputs into a model, which then replicates their capabilities by discerning underlying patterns within the text.

Oh, and China’s giving these models away

The new Chinese models would have caused Silicon Valley enough headaches, if they merely provided stiffer competition, while demonstrating the power of distillation. 

What makes Kimi K3 and Qwen3.8 Max especially threatening to the American AI giants’ profitmaking potential, however, is that they are officially open source — meaning that the models’ parameters can be downloaded for free. (Alibaba and Moonshot have not yet released these parameters, but they say they will shortly.)

In other words, any company or hobbyist with enough computing power will soon be able to run a near-frontier Chinese model on their own hardware, modify that model to better serve a specialized purpose, and then sell access to their new version — without paying Alibiba a single yuan.

As Kimi and Qwen grow more capable, their market-share is likely to grow, at American AI giants’ expense.

For many of Anthropic and OpenAI’s potential customers, that proposition may be hard to turn down. Most businesses don’t need the world’s smartest AI, just one competent at their enterprise’s core tasks — compiling legal research, answering IT queries, writing working code, etc. A model that produces outputs 90 percent as good as Claude’s — at roughly one-sixth of the cost — will sound pretty good to many corporations.

Further, open source models aren’t just cheaper than frontier systems, but potentially more secure. If you run an AI on your firm’s own servers, then you don’t need to entrust sensitive data to Anthropic, Google, or OpenAI.

All this had led much of corporate America to embrace open-source models, even before the latest versions narrowed the capabilities gap. In a Linux Foundation survey, 63 percent of organizations reported using open-source AI systems.

And increasingly, those models are Chinese. According to Sequoia Capital, one of Silicon Valley’s premier venture capitalist firms, a majority of American AI startups now use open-source Chinese systems. As Kimi and Qwen grow more capable, their market-share is likely to grow, at American AI giants’ expense.

What’s bad for OpenAI is good (and/or catastrophic) for humanity

All this said, it is still entirely possible that OpenAI and Anthropic will justify their colossal valuations. In many highly competitive economic domains, having access to the world’s very best AI model will remain highly valuable. And America’s frontier labs still outperform all their peers. 

But it’s increasingly plausible that selling state-of-the-art AI systems will prove to be a low-margin undertaking. In a world of ubiquitous, near-frontier open source models, the AI sector’s big winners probably won’t be its top labs, but rather, its chipmakers and cloud computing providers. 

For ordinary people, a future where superintelligence is dirt cheap — and rival AI companies are constantly rising and falling, rather than consolidating into mega-corporations — would look somewhat different than the cyberpunk dystopia that the left’s been dreading. 

And not entirely in a good way. For one thing, in that reality, mitigating AI’s biggest risks would be immensely difficult. Having a handful of firms monopolize control over frontier AI systems is bad in many respects. But it does make those models easier to regulate, as the Trump administration’s decision to temporarily block Claude’s Fable in the name of cybersecurity demonstrated. 

By contrast, if recipes for ultra-powerful AI models are published all over the internet — and anyone with modest technical skills can modify them at will — then systems willing to help their users hack government bureaucracies or engineer bio-weapons are liable to proliferate.

From another angle, however, the “AI becomes almost free” scenario may look like capitalism at its finest: Retrospectively, such a development would mean that a small number of extremely rich people bankrolled the creation of an immensely useful technology, under the expectation of massive profits, only to see competition erode their returns — and disperse that tech’s benefits across a wider group of businesses and consumers. 

Granted, in the case of AI, this process might also generate a super-virus that kills us all. But hey, no system is perfect.

Abdul El-Sayed is among America’s most prominent proponents of Medicare-for-all. 

The frontrunner in Michigan’s Democratic Senate primary literally wrote the book on that policy (or at least, a book on it). In El-Sayed’s view, Medicare should cover “all necessary healthcare” for every American — without co-pays, premiums, or deductibles — and be “accepted everywhere.”

Alas, despite his many years of advocacy, El-Sayed has seemingly failed to persuade his wife of that last point: According to a recent report from the Washington Free Beacon, El-Sayed’s partner, the psychiatrist Sarah Jukaku, does not accept Medicare as a form of payment at her private practice.

This bit of gossip is of little importance, in and of itself. The Free Beacon’s story does nothing to refute the case for El-Sayed’s candidacy or his healthcare plan (his wife’s business is, well, her business). As hit pieces go, it’s weak tea.

Nevertheless, the tension between El-Sayed’s healthcare proposals and his wife’s business practices is real. And it is illustrative of a major challenge facing anyone who wishes to reform our nation’s misbegotten healthcare system: To meet the medical needs of all Americans, reformers will need to defy the interests of most doctors — and in many cases reduce their compensation. 

And that won’t be easy. Few people feel a deep fondness for insurance companies. But El-Sayed is far from the only American who loves a physician.

The biggest obstacle to Medicare-for-all

To understand why Medicare-for-all would be bad news for many doctors — and how the Free Beacon’s story illustrates that point — we must first dwell on one fundamental fact about America’s healthcare system: It’s a rip-off. 

The US spends about twice as much per person on medical goods and services as other wealthy countries. And yet, all that money does not actually buy us much more care. Compared to our peers abroad, Americans are less likely to see a doctor, secure a long hospital stay, or access a timely appointment for medical treatment. On the other hand, we do have the privilege of paying radically higher healthcare prices.

To take just one telling example: In the United States, a coronary bypass surgery will typically cost more than $89,000; in Australia, it costs just $17,741.

Such exorbitant prices are the chief obstacle to any version of universal healthcare. Even with one-third of working-age Americans uninsured or underinsured — and thus, consuming too little medical care — the bill for America’s health sector ran to $5.7 trillion in 2025. 

In El-Sayed’s vision, Americans would consume vastly more medical services than they do today: The uninsured would suddenly have access to every doctor in the country, while everyone else would see their co-pays and deductibles drop to zero, encouraging them to schedule far more doctors’ visits.

This would be a costly proposition in any country. At America’s current healthcare prices, it would be prohibitively expensive. There is simply no way to realize anything approaching the left’s healthcare ambitions without slashing the amount of money that Americans pay per medical service. 

Doctors will pay a price for universal healthcare

Medicare-for-all advocates are aware of this fact. And they’re typically eager to talk about one source of America’s high healthcare prices: The inefficiencies of our private health insurance model. 

In America’s byzantine system, each insurer needs its own teams of auditors, claims reviewers, and myriad other specialists, while every major healthcare provider needs a horde of administrators to navigate the idiosyncratic rules of all these different insurance companies. Americans pay dearly for this bureaucratic bloat. By one estimate, our system’s administrative costs are $500 billion higher than they would be if the insurance industry was consolidated into a single public insurer. 

And yet, as large as that figure may seem, it still represents a fraction of America’s excess healthcare costs. The primary cause of our nation’s exorbitant medical prices is simpler than administrative redundancies: our healthcare providers charge exceptionally high rates.

Hospitals are the biggest culprits on this front. But physicians are also part of the problem.

According to a 2026 study from economists at the University of Chicago, Stanford, and the US Census Bureau, American physicians earn about twice as much as Canadian ones — and four times as much as Swedish doctors. 

Critically, this does not merely reflect America’s greater wealth or wage inequality. It is true that educated professionals of all kinds — financial analysts, lawyers, software engineers, etc. — earn more in the US than they do in other rich countries. But American doctors don’t just earn unusually high absolute incomes — they also occupy an atypically rarified place within their own country’s class hierarchy. About 42 percent of American specialty physicians are in the top 1 percent of their nation’s income earners. Among Canadian specialists, that figure is just 27 percent; for Swedish ones, it is 7 percent.

The main driver of these disparities is straightforward: America imposes fewer price controls on its healthcare sector than other nations do. 

And this is where Jukaku’s practice reenters the picture. 

The public parts of America’s insurance system — Medicare and Medicaid — pay rates that are only modestly above international norms. It is when American doctors bill private insurers — or the rich consumers of boutique medicine — that they really make bank.

As a result, top clinicians like Jukaku often decline to take Medicare. If you’ve got affluent patients beating down your door, accepting Uncle Sam’s rates just doesn’t pay.

Unless the government forces doctors and hospitals to swallow steep pay cuts, however, Medicare-for-all won’t pencil out. According to a widely cited 2018 analysis by the economist Charles Blahous, if a single-payer system kept provider payments constant, national health spending would rise by $3.25 trillion over a decade, even with administrative savings taken into account. By contrast, if all providers were forced to accept Medicare’s rates, health spending would actually fall by $2.05 trillion over the same period.

Soaking physicians is tough politics

Thus, there is a clear conflict between progressives’ healthcare ambitions and medical providers’ material interests. 

Yet the left is often reluctant to acknowledge this reality. El-Sayed tends to portray insurers as the sole economic beneficiaries — and political defenders — of America’s inequitable healthcare system. The fact that hospitals and doctors also profit off the status quo’s dysfunctions does not feature prominently in his rhetoric. To the contrary, El-Sayed suggested in 2020 that doctors like his wife are actually underpaid, even though American psychiatrists earn far higher salaries than their counterparts abroad. 

To be fair, progressives aren’t alone in eliding providers’ culpability. Virtually all Democratic politicians do the same. And not without reason. Politically speaking, it is one thing to denounce the greed of private insurers — the faceless bureaucracies standing between Americans and their desired treatments. It’s quite another to call for reducing the wages of doctors, men and women who perform laudatory work and enjoy widespread admiration. 

Precisely for this reason, however, reformers must grapple with healthcare providers’ investment in the current system. The American Medical Association (AMA), the lobby representing our nation’s physicians, was instrumental in killing past attempts to move toward single-payer. And at least some segments of the medical profession would surely mobilize against any contemporary Medicare-for-all bill that imposed substantial cost controls on the healthcare sector. What’s more, in doing so, they would be able to draw on a resource the private insurance industry lacks — the public’s trust.

How to make healthcare less expensive right now

There is no easy answer to the problems all this presents. But part of the solution is to chip away at providers’ payment rates where progressives already have the power to do so. This would not only help drive down costs for existing healthcare in the short term, an urgent priority all its own, but also would smooth the path to universal coverage in the long run.

That project can take many forms. One would be state-level payment regulations. In Maryland, hospitals receive the same rates, no matter whether their patients pay with Medicare, private insurance, or cash. And their budgets are also fixed, so that they aren’t able to milk fees out of unnecessary care. Rhode Island, meanwhile, caps the growth of its hospital reimbursement rates at the pace of overall inflation. Other states could follow their lead. 

But policymakers should also address the supply constraints that undergird American doctors’ high salaries. US physicians’ ability to command high pay doesn’t just reflect America’s weak cost controls but also a persistent shortage of working doctors. The US has roughly 2.7 physicians for every 1,000 of its residents; the average among comparable countries is 3.9, according to a Kaiser Family Foundation analysis.

In this context, forcing down doctors’ pay might seem perilous. After all, doing so would reduce young people’s incentive to pursue a medical career, potentially deepening the shortage. 

In reality, however, there is no dearth of qualified people who want to practice medicine in the US. We just don’t let many of them do so.

This is partly because American policymakers consciously sought to restrict the number of doctors in the country, beginning in the 1980s. As Robert Orr of the Niskanen Center explains, the US government issued a report in 1981 warning of an imminent “physician surplus” and recommending “immediate action to curtail both the domestic training of physicians as well as the admittance of those trained outside of the country.”

The report’s argument rested on false premises; it failed to anticipate that Americans’ demand for healthcare would rise sharply as they grew wealthier. Nonetheless, its recommendations were largely implemented: Federal support for medical-school scholarships was pared back while funding for residencies has been capped since 1997. 

At the same time, policymakers maintained high barriers to the immigration of fully-trained foreign doctors: Even physicians with years of experience, and credentials in nations with high medical standards, are typically required to complete a multi-year residency before being able to practice in the US.

Ending the federal freeze on residency funding will require congressional action. But states can immediately make it easier for foreign doctors to practice within their borders. In fact, Tennessee established a pathway for such physicians to ply their trade in the state, without having to repeat a residency, in 2023. And many states subsequently enacted similar reforms.

Removing the bottlenecks on America’s doctor supply won’t eliminate the political hurdles to Medicare-for-all. But it would put downward pressure on doctors’ salaries, reduce the risks of capping physician pay, and make the left’s vision of healthcare abundance more feasible. After all, you can’t actually eliminate the care rationing that so many Americans resent by extending insurance coverage or enacting price controls alone. No matter how we pay for our medical services, we can only deliver as much care as our health sector’s resources allow. 

Don’t hate the doctor, love the sick

In saying all this, I don’t mean to convey disdain for the medical profession. Like El-Sayed, some of my best friends are doctors! In fact, my mother, father, brother, and sister-in-law are all physicians. And they all have contributed far more to American society than I ever will. My brother spends his workdays providing lifesaving treatments to cancer patients; I often spend mine sitting at a desk in my pajamas, arguing about politics on the internet. 

Physicians deserve to be well-paid for their strenuous labor. But if we want healthcare in America to be universally affordable and widely accessible, we will need to pay many of them a bit less.

The Democratic Socialists of America say they are fighting for “a government by, for, and of the working class.” 

Yet the organization’s membership and candidates consist largely of educated professionals. As of 2021, 80 percent of DSA members over 25 were college-educated, while 28 percent earned more than $100,000 a year (putting them in roughly the top 15 percent of America’s income distribution), according to a DSA internal survey. Among the group’s most prominent politicians, meanwhile, is an Ivy League graduate student, a lawyer, and the well-educated son of a Hollywood director and famous academic. 

In recent weeks, as the DSA has racked up victories in Democratic primary elections, the apparent tension between the group’s self-conception and demographics has attracted critical scrutiny.

In the New York Times, Thomas Edsall argues that most of the DSA’s members and supporters in the electorate are “in no way working class.” The libertarian commentator Robby Soave, right-wing pundit Batya Ungar-Sargon, and liberal economics writer Noah Smith have all voiced similar sentiments. 

There are many cogent ways for socialists to respond to such criticism. One would be to note that an organization doesn’t need to have a uniformly working-class membership in order to fight for working-class interests. Another would be to observe that not all college graduates are well-heeled professionals, even as pundits often use educational attainment as a stand-in for class in political analysis. 

But some socialists prefer a simpler retort: The DSA’s critics are simply mistaken; in reality, the group’s membership is more or less entirely “working class” — because that term describes everyone who must sell their labor in order to meet their needs. 

This conception of the “working class” has deep roots on the left. As Jacobin’s Ben Burgis explains, orthodox socialists subscribe to a “two-class map of capitalist society,” in which “Some people are capitalists who own their own ‘means of production’ (like factories, farms, offices, or grocery stores) and other people work for the capitalists.”

Many contemporary socialists (including Burgis) find this dichotomy reductive, and endorse more nuanced portraits of America’s class structure. But the orthodox, “two-class” model informs a great deal of the modern left’s rhetoric, among both socialists and progressives. And the notion that lawyers and day laborers both belong to the “working class” is widely held among anti-capitalists. 

There’s something to be said for this idea. Educated professionals and manual workers surely do have some shared interests, due to their common status as laborers. For example, both benefit from the existence of a well-funded unemployment insurance system, in a way that the Elon Musks of the world do not. 

For the most part though, orthodox socialism’s “map” is badly out of date. And its expansive conception of “the working class” does less to illuminate modern America’s economic divisions than to obscure them.  

The long afterlife of Marx’s failed prophesies

Karl Marx and Frederich Engels laid the foundations for the modern left’s conception of class in The Communist Manifesto of 1848. 

In that book, the claim that there were just two important classes under capitalism — capitalists and workers — rested on a set of grim predictions about where industrialization was headed. 

According to Marx’s analysis, technological progress was in the process of obliterating “all distinctions” between different strata of working people, as automation was driving virtually everyone’s wages down to “the same low level.” The middle class of small manufacturers and shopkeepers, meanwhile, was poised to “decay and finally disappear,” as more productive, large-scale enterprises left them in the dust.

Society was therefore splitting into “two great hostile camps” — the capitalist and working classes, whose interests were impossible to reconcile. 

Under these conditions, a binary conception of class makes perfect sense. If all workers will soon be condemned to low wages, then there’s little point in dwelling on the (temporary) distinctions between them. Likewise, if shopkeepers and small-time capitalists have gone the way of the triceratops, one can more comfortably cast all business owners as members of the ruling elite.

You can’t navigate modern capitalism with a 19th-century map

But we don’t live in the future that Marx foresaw. 

Rather than melting all workers into a uniform proletariat, capitalism cultivated a kaleidoscopically diverse laborforce — one segmented by countless fine-grained distinctions of pay and prestige. And instead of immiserating ordinary people, industrial development brought unprecedented mass prosperity (albeit, with the aid of many socialistic reforms). 

As a result, old-school socialists’ “two-class map” makes for a poor guide to modern capitalism’s terrain. And anyone trying to impose Marx’s cartography onto the landscape will encounter several problems.

For one, in a world where some workers earn as much in a year as others do in a lifetime, the idea that there is only one fundamental class division — that between laborers and capitalists — becomes unconvincing. 

In the United States, some convenience store owners make $70,000 a year, while many surgeons earn upward of $600,000. It would be odd to say that the former are members of society’s dominant class (since they live off business income), while the latter belong to its subordinate one (since they live off salaries). After all, the minimart owner and Musk do not have especially similar material interests. Nor, for that matter, do the impoverished farmworker and high-earning neurosurgeon (much less, the superstar athlete with a $40 million salary). Yet this is what orthodox socialism’s binary would suggest. 

Ultimately, in a modern capitalist society, it makes a lot more sense to see class as a multidimensional continuum than as a binary.

A second, related difficulty is that the border between labor and capital has grown fuzzier since Marx’s time. Today, American workers often own both a home and some share of their nation’s financial assets, like a 401(k) filled with stock from major corporations. Many billionaires, meanwhile, choose to work. 

Socialists are well aware of this reality, which is why they define a person’s class status by their dependence on labor: The typical white-collar worker may own a little stock, but not enough to quit her job and live off dividends. Unlike a billionaire CEO, such a professional must work in order to meet her needs. 

This is a sound distinction. But it does not actually resolve the challenge of distinguishing the working class from the ruling one. This is because, in an advanced economy, whether someone “needs” to work is partly subjective. Millions of Americans survive on less than $25,000 a year in market income. If upper-middle-class families were willing to accept a similarly austere living standard, many would be able to live entirely off their investments.

Of course, Americans shouldn’t be content to live in poverty. But it is harder to say precisely what standard of living qualifies as “enough.” In Marx’s time, a relatively good life for a wage worker might have meant reliable access to food and shelter for one’s family and a little time and money left for simple leisure. Today, the list includes high-quality education for one’s children, access to advanced medical treatments, modern home appliances for cooking and cleaning, Internet-connected devices for entertainment, and often personal transportation, like a car — all of which are typically (and justifiably) treated as basic necessities.

This raises the question: At what level of wealth does a person stop needing to work — even if they must keep clocking in to achieve their desired lifestyle? The socialist writer Matt Bruenig has proposed drawing the line at about $1.4 million (or roughly 20 times America’s average wage): Once you’ve accrued that much in housing and financial assets, your decision to sell your labor becomes voluntary and you cease to be working-class. 

This is a reasonable cut-off. But it’s still fairly arbitrary. Ask an American with a net worth of $1.4 million whether they need to work, and they’re liable to say yes. And typically, they will be correct — at least, in the sense that they could not sustain their way of life without earning a paycheck. You could say that their present lifestyle is too lavish — and that $1.4 million in assets is all that’s required to generate the amount of income one truly needs. But it’s not obvious why this threshold doesn’t actually lie at $1.6 million, or $800,000, or $500,000.

As a result, the only way to draw the upper boundary on who “needs” to work — without making arbitrary judgments about what standard of living is necessary — is to place it extremely far up the class hierarchy: The superrich generally don’t need to choose between sustaining their lifestyles and abstaining from work. But once your definition of “working class” becomes capacious enough to include low-end multimillionaires, it loses all plausibility. 

Ultimately, in a modern capitalist society, it makes a lot more sense to see class as a multidimensional continuum than as a binary. An individual’s degree of economic freedom and power is determined by a host of factors — above all, their household incomes and net worths, but also the market value of their skills (which shapes their future earning prospects) and familial wealth and connections. Even one’s eligibility for social welfare benefits can matter; after all, the largest group of Americans who don’t need to work to meet their needs are Social Security recipients.

Precisely how we should weigh these different sources of advantage is a tricky question. And implicit disagreements over it shapes our discourse about “the working class.” 

Take Graham Platner, the former Democratic Senate candidate in Maine. To his champions, Platner was a working-class candidate, due to his lack of a college diploma, years of service in the military, and low-income, physically demanding job farming oysters (along with his gravelly voice and tattoos). To his detractors, meanwhile, Platner was a pseudo-proletarian due to his upper-middle-class childhood, prep school education, and receipt of steady financial help from his father and mother, a lawyer and entrepreneur. Meanwhile, Platner probably would not qualify as “working class” under the orthodox socialist definition for multiple reasons: For one, he does not need to work in order to get by, thanks to his disability benefits and family support. For another, he is the owner of his oyster business, rather than a salaried worker who answers to a boss. 

In any event, whichever factors one emphasizes, it’s clear that a person’s class privilege increases continuously as their income, net worth, familial wealth, and “human capital” rises. There is no single point on any of these scales past which a person’s class position is abruptly transformed. The gap between “the 99 percent” and the billionaire class matters — but so does that between the top 15 percent and the working poor.

Of course, in ordinary political discourse, we inevitably need to boil down reality’s intricacies into rough categories. No politician will ever promise to fight for the “bottom six deciles in combined SES score, nor will any labor activists ever chant, “The 30th to 60th percentiles in household income, united, will never be defeated.” 

But a taxonomy of classes with several subcategories (such as “the poor,” “the working class,” “the middle class,” “the upper-middle class,” “the rich,” and “the superrich”) will capture a bit more of reality than a taxonomy with only two. 

“Working-class” rhetoric can camouflage professional-class interests

All this said, even if orthodox socialism’s “two-class map” is analytically flawed, it could still be politically beneficial. Perhaps, when affluent professionals decide that they belong to the same class as the working poor, they become more inclined to fight for the latter’s interests. 

Some of the DSA’s own activities lend credence to this theory: Despite its largely white-collar membership, the organization advocates for many policies that would principally benefit America’s most disadvantaged. More broadly, in the wake of Occupy Wall Street, many college-educated millennials came to identify with “the 99 percent.” And as this cohort has grown more influential in Blue America, the Democratic Party has embraced more ambitious anti-poverty policies, such as a guaranteed monthly income for working-class families. 

The main fiscal barrier to American social democracy is that our middle class in general — and upper middle-class in particular — pays much lower taxes than their peers in Western Europe.

And yet, socialists’ expansive conception of “the working class” can be put to more dubious uses. Specifically, it can help relatively well-off professionals disguise their class’s particular interests as those of all working people.  

The left’s push to cancel all graduate student debt is arguably a case in point. During Joe Biden’s presidency, the DSA criticized the White House for attempting to forgive only $10,000 of most borrowers’ student loans (a plan that was ultimately blocked by the Supreme Court). Instead, the socialists called on Biden to cancel all student debt, including the balances accrued by the graduates of elite medical schools and MBA programs. 

On its face, the idea that the US government should transfer tens of thousands of dollars to young doctors and Harvard MBAs would seem wildly regressive. The median US physician can expect to earn more than $6.5 million over the course of their lifetime, while the median graduate of a top-five business school can expect to earn more than $8 million. Virtually all Americans have worse prospects than this. And in the inflationary conditions of the Biden era, increasing young professionals’ disposable income was liable to push up prices for everyone else. In other words, blanket student debt forgiveness would have likely transferred purchasing power from low-income workers to early-career physicians and management consultants.

Many socialists would struggle to defend such a policy, were it presented in these terms. But by eliding the class distinctions between low- and high-income borrowers,  the DSA was able to convince itself (if not the public) that across-the-board forgiveness was in the interest of “the working class.”

The socialist framework is similarly unhelpful in the realm of tax policy. 

In recent years, as socialist and progressive politicians have grown more dependent on professional class support, they’ve become more averse to endorsing broad-based tax increases.

From Zohran Mamdani to Chris Van Hollen, the Democratic left tends to reserve its tax hike proposals for the superrich. And yet, the reason that America lacks a comprehensive welfare state is not that our billionaires pay unusually little in taxes; to the contrary, our top rates are comparable to those of some Scandinavian countries. The main fiscal barrier to American social democracy is that our middle class in general — and upper middle-class in particular — pays much lower taxes than their peers in Western Europe. Simply maintaining our existing programs, like Social Security, will likely require a major tax increase on higher-end salaried workers. 

In this context, binary conceptions of America’s class structure — which pit an undifferentiated “working class” against the capitalist elite, or the 99 percent against the 1 percent — can help educated professionals evade responsibility for bankrolling a more egalitarian economic system.

Check your class privilege

Ultimately, the DSA’s ability to serve working-class interests hinges less on its own demographics than its policies and politics. When an organization mistakes the preferences of its most privileged members for those of all working people, however, its campaigning and policymaking is liable to suffer. 

Perhaps, before this century’s end, AI will bring about the calamity that Marx prophesied — and Ivy League lawyers and delivery drivers will find their skills equally devalued by an all-powerful ownership class. Until the robot apocalypse, however, educated professionals and low-wage workers will live in very different circumstances — and harbor distinct economic interests. To safeguard the wellbeing of the disadvantaged, well-off egalitarians need to mind these gaps, not paper over them with obsolete class categories. 

President Donald Trump blames immigrants for virtually every problem in American economic life. In the president’s telling, the undocumented have been taking American-born citizens’ jobs, lowering their wages, driving up their costs, and bankrupting their welfare programs.

These claims are largely false.  

Most studies suggest that immigrants do not typically reduce wages or job opportunities for the American-born — and actually pay more in taxes than they receive in social welfare benefits (this is especially true of undocumented immigrants, due to their ineligibility for Social Security and Medicare).

But there is at least one way that immigration really can hurt American-born people economically: When migrants come to a locality in large numbers, they can push up housing costs for longtime residents. 

The president called attention to that fact earlier this week. In a Truth Social post Sunday, Trump wrote, “Fed Reserve working paper suggests Biden illegal immigrant wave drove up home prices 30%.”

Here, the president was garbling the findings of a recently published Dallas Federal Reserve study, which mostly contradicts his narrative about immigration’s economic harms. 

Still, that paper did find that unauthorized immigration pushed up both rents and home prices during the Biden era (albeit by much less than Trump claimed).

This is a real problem. Yet, the solution is not to close America’s borders. Throttling immigration would make America less prosperous, while doing little to make housing more affordable.

Fortunately, there is a way to secure immigration’s considerable economic benefits while limiting its more modest downsides: We can simply make it easier to build more homes. 

What the Dallas Fed study actually shows

The spike in unauthorized immigration under Biden was genuinely massive. 

Between 2021 and 2024, seven million unauthorized immigrants settled in the United States. Put differently, America added 1.75 million unauthorized migrants per year during this period. For context, between 2000 and 2019, the US added an average of 0.1 million annually.

The extraordinary scale of the Biden-era wave provided researchers at the Dallas Fed with a unique opportunity to gauge unauthorized migration’s material consequences. Drawing on court records that revealed where immigrants had settled, the economists examined how local economic trends varied between places with more or fewer migrants, controlling for other variables.

As the president suggested, the immigration spike had some adverse impacts on local housing markets. Between early 2021 and early 2024, unauthorized migration pushed up the average American municipality’s home prices by 6.6 percent (not 30 percent, as Trump claimed) and its market rents by 4.3 percent.

In some contexts, rising home prices aren’t entirely unwelcome. Immigrants have helped spur the economic revival of many declining localities, which had previously suffered large drops in population, and, thus, in tax revenues and job opportunities. When such communities become more economically vibrant, property values tend to rebound — a welcome development for local homeowners. 

Nevertheless, many American cities currently suffer from perennial crises of housing unaffordability. And the Fed’s research suggests that the Biden-era immigration surge made this problem worse, accounting for 30 percent of the overall increase in home prices during the period. 

That’s significant, even as other factors played a collectively bigger role, among them tight post-pandemic labor markets, the soaring cost of construction materials, a jump in demand for floor space due to remote work (as middle-class families sought larger homes to accommodate home offices), and America’s pre-existing housing shortage.

This finding is consistent with past research into immigration’s impact on housing costs. And it is also intuitive, given the inherent and artificial constraints on housing construction in the United States.

After all, cities can add new residents much faster than they can build new housing. When a large number of immigrants show up in a city all at once, demand for homes will inevitably grow quicker than supply.

In a healthy housing market, this would merely be a short-term problem. Eventually, rising home values would encourage more building until prices fell back towards their pre-surge trajectory. Recent trends in Austin, Texas, are a case in point. Between 2021 and 2025, rents in that city fell by 4 percent, even as its population increased, thanks to a boom in housing construction.

In most US cities, however, land-use rules make it functionally impossible for housing supply to keep pace with demand. Bans on the construction of apartment buildings, large minimum lot size requirements, and other rules effectively cap the amount of housing that can be erected, no matter how much the population rises. Under these conditions, the Trumpian vision of immigration — in which American-born people and new arrivals are locked in a zero-sum struggle for scarce resources — becomes a little more true.

Slashing immigration is not the answer

It would be a mistake to respond to this by throttling immigration rather than fixing our housing markets. 

Immigration is still economically beneficial for Americans on the whole. In truth, the Dallas Fed’s study does more to undermine nativists’ arguments than to support them. 

According to that paper, the Biden-era migration wave did not take a fixed set of jobs from American-born people. Rather, it added roughly as many new jobs as new workers to local economies. 

Which makes some sense: Immigrant workers do not just provide goods and services, but also consume them. Thus, although they increase an area’s labor supply, they also boost its labor demand, sparking job growth.

Further, contrary to nativists’ expectations, unauthorized immigration had no statistically significant impact on local wages during the Biden era and drove down government transfer spending per capita in impacted localities, likely because undocumented immigrants are ineligible for many social welfare programs.

Meanwhile, we know from previous research that immigration benefits economies in myriad ways, not least by fostering a more efficient division of labor and, thus, higher productivity. This is partly because immigrants are far more willing to move in response to shifting economic conditions than American-born workers, which makes them uniquely effective at filling gaps in local labor markets. 

More critically, as the American population ages in the coming decades, our nation’s ratio of workers-to-retirees is set to fall. That will make supporting seniors’ pensions and health care cost more onerous for the young while also depressing economic growth and boosting deficits. Welcoming more working-age immigrants into America is the simplest way to mitigate these problems.

In any case, choking off immigration would do little to resolve America’s housing crisis. After all, even during a historic surge in unauthorized entries, 70 percent of the increase in home prices — and 80 percent of the rise in market rents — was due to factors other than immigration, according to the Dallas Fed’s report.

Thus, if we want to maximize Americans’ prosperity, we need to find a way to reconcile robust immigration with affordable housing. 

This will probably require avoiding another abrupt surge in unauthorized migration, like that which transpired under Biden. Rapid, disorderly spikes in irregular immigration are politically toxic and economically disruptive. Even in a well-functioning housing market, a sudden jump in population will generate short-term price spikes. 

Above all, however, we must pare back barriers to homebuilding. Until we do, housing will remain unaffordable in much of America, no matter how few immigrants we take in — and policymakers may struggle to sell voters on high levels of legal immigration, no matter how many facts they have on their side.