25 things we think will happen in 2025

For the sixth year in a row, the staff of Future Perfect convened in December to make predictions about major events in the year to come. Will Congress pass a tariff bill that makes President-elect Donald Trump happy? Will the H5N1 bird flu become an honest-to-god pandemic? Will the war in Ukraine stop? Will a major sports figure get caught up in a gambling scandal?

It’s fun to make predictions about the future, which is part of the reason why we do it so often. But this isn’t just blind guessing. Each prediction comes with a probability attached to it. That gives you a sense of our confidence (high in the case of, say, Charli XCX’s Grammy chances, less so in the case of Iran’s nuclear plans). And don’t make the same mistake that people seem to make every presidential cycle. Even a probability as high as 75 percent or 80 percent doesn’t mean we’re sure something will happen. Rather, it means we think that if we made four or five predictions, we’d expect three or four of them to come true, respectively.

And as we have every year, we’ll be keeping track of how our predictions fared over the course of 2025, and report back to you at the end of December. You can check out how we did in 2024 here. And we’ve done something new this year in partnering with the prediction platform Metaculus. You can check it out here to see how the community there came down on a number of our predictions — and even compete in a prize pool — and click on the individual questions with links to go directly to them on Metaculus. We’ve also added the Metaculus community’s aggregated forecasts as of December 31 for the questions they’ve taken on. —Bryan Walsh

Donald Trump’s 2024 campaign was perhaps the most pro-tariff of any candidate since William McKinley: He promised 60 percent taxes on imports from China, and 10 percent on everywhere else.

In victory he’s only gotten bolder, calling for 25 percent tariffs on Canada and Mexico, in flagrant violation of the United States-Mexico-Canada Agreement, a free trade deal made by some past president named Donald Trump.

The bad news for consumers and the world economy is that Trump has substantial discretion to impose tariffs as president without consulting Congress. But that discretion isn’t unlimited, and probably doesn’t permit the kind of 10 percent across-the-board tariff Trump promised. Plus, Republicans want a revenue source to help offset the cost of making Trump’s 2017 tax cuts permanent before they expire at the end of next year. This raises the question: Will Congress pass a tariff measure on its own that not only implements Trump’s ideas, but lets them endure under future presidents?

My guess is no. There was a time in the distant past, let’s call it “2015,” when Republicans were the party of free markets and free trade, and some members of Congress haven’t forgotten that. Early reporting suggests that many GOP figures in the House and Senate are hostile to the idea of including tariffs in a tax package. Republicans can only lose three senators and two House members out of their caucus and still pass bills, which gives them very little margin for error, and makes it very difficult to pass legislation that splits the caucus like tariffs.

Two caveats, though. One, I’m predicting about a tariff bill and not new unilateral tariffs from Trump because I think the odds that Trump does new tariffs using presidential authority are nearly 100 percent. Two, the only reason my estimate isn’t lower is that there’s been some bipartisan interest in a “carbon border adjustment,” or a sort of carbon tax that only applies to imported goods. The idea has gotten Republican support because while it does acknowledge that global warming is real, it also sticks it to foreigners. That’s a tariff, and I think the likeliest kind to make it into a tax package (though I still bet against it). —Dylan Matthews

Metaculus aggregated forecast: 7 percent

Something I love and hate about American politics is that big weighty-seeming questions, like “Can Donald Trump fulfill his campaign promise to abolish the Department of Education?” turn out to hinge on much weirder and more technical questions, like “Will the Senate parliamentarian rule a departmental reorganization as ineligible for the reconciliation process under the Byrd rule?”

The Department of Education, whose main duties are administering student loans and financial aid for higher-ed institutions and distributing funds (around $39 billion in 2023) to local schools under programs like Title I (for poor districts) and IDEA (for disabled students), was created in a 1979 act of Congress. Passing a normal bill repealing that act would require 60 Senate votes to break a filibuster, which means winning over seven Democrats to the idea, which isn’t going to happen.

So legislation abolishing the department (already written by GOP Sen. Mike Rounds of South Dakota) would have to pass through budget reconciliation, which lets certain legislation pass with a mere majority in the Senate.

But reconciliation has strict requirements limiting the content of legislation that can be passed that way, and in particular provisions of bills that are only “incidentally” related to the overall level of spending or taxing tend to be struck down by the Senate parliamentarian as contrary to the Byrd rule, the main governing principle behind the reconciliation process.

Rounds’s bill, notably, doesn’t eliminate the Department of Education’s actual functions. It just moves them around. Student loans, for instance, would go to the Treasury Department, and the Department of Labor would get vocational programs. This strikes me as an archetypal example of a change that is merely incidental to the actual level of spending, and that can’t be done with reconciliation.

Will the Senate parliamentarian disagree with me? Possibly. But also, in part because this move is so much more about reorganization than the actual substance of the department’s programs, I am very skeptical that Republicans are going to go to the mat on this one. If they can only win so many fights with the parliamentarian, are they going to prioritize changing the mailing address of the student loan office? I doubt it. —DM

Metaculus aggregated forecast: 4 percent

From the moment the Affordable Care Act was signed into law on March 23, 2010, the Republican Party has been obsessed with repealing it. They even shut down the government over it. Then, in 2017, the dog finally caught the car: Republicans had both houses of Congress and the presidency, and in theory the opportunity to repeal the law.

They didn’t.

Sure, the tax law that year eliminated the individual mandate to get health insurance, but that turned out to not be as important to getting people coverage as the ACA’s authors thought. The rest of the bill — its dramatic Medicaid expansions, rules protecting people with preexisting conditions and letting young adults stay on their parents’ insurance, subsidies for individuals to buy health insurance if their employer doesn’t provide it — remained intact.

Even “skinny repeal,” a bill that zeroed out only a handful of provisions of the law, failed to pass the Senate when John McCain made his famous thumbs-down gesture, but matters had only even gotten to that point because several other senators didn’t want to vote for sweeping Medicaid cuts, like those entailed by simply repealing the ACA in its entirety.

Will they try again in 2025? I’m skeptical. And here, by “repeal Obamacare,” I don’t even necessarily mean repealing all of it. To qualify as repeal, a bill has to do at least three of the following five things:

Yes, Trump’s budgets and those that his past and future budget chief Russ Vought prepared during the Biden years did propose undoing the ACA’s coverage expansions, and then cutting Medicaid still further. I anticipate they will continue to make these proposals. But I am doubtful that with a much narrower House majority than they had in 2017, and an equally narrow Senate majority, Republicans will be able to pass cuts on a scale that they couldn’t get off the ground eight years ago. —DM

Metaculus aggregated forecast: 10 percent

Here are the facts: Jerome Powell’s term as chair of the Federal Reserve expires on May 15, 2026. He has pledged to stay on as chair until that time, though not necessarily to remain as a member of the Board of Governors until his term there expires in 2028. Donald Trump has said he does not plan to fire Powell before that time. Powell has insisted that the president does not legally have the power to fire him before his term is up, and that he will refuse to obey such an order.

In many ways, 90 percent seems too low, because the odds that a 71-year-old man dies in the next year are only 2.9 percent, and I have an easier time envisioning Powell dying in office than acquiescing to a firing.

But I should also fess up to a personal bias here. Jay Powell is, by a wide margin, the greatest chair of the Federal Reserve that the institution has ever had, and perhaps the greatest central banker in any nation of modern times. He prevented Covid from spiraling into a global financial crisis, oversaw an astonishingly rapid recovery of employment and economic growth in the pandemic’s aftermath, and managed a “soft landing” that ended an inflationary episode without having to spark a recession. He is a miraculous figure who we do not deserve, and for my own sanity I need him to stick around. —DM

Metaculus aggregated forecast: 8 percent

Americans have a charming habit of deciding to like the newly elected president as soon as the election’s over, and Donald Trump’s favorability rating has gone from 8.6 points underwater on Election Day to only 1.9 points negative on December 19. (By “favorability rating,” I mean the difference between the share of voters saying they view him favorably minus the share saying they view him unfavorably. Once he’s president, I’ll count this prediction as resolving to “true” if either his favorability or job approval ratings are positive; while similar, these aren’t exactly the same question.)

But how long do presidential “honeymoon” periods last? Not very long, as it turns out. Back in 2022, FiveThirtyEight’s Geoffrey Skelley and Jean Yi crunched the data and found that Obama, Trump I, and Biden alike all saw their approval ratings dip below 50 percent by the end of the first year (Trump was never even above 50 when he started!):

The two exceptions on the chart are Bill Clinton, who saw a curious fall and then recovery over 1993 that I don’t really understand, and George W. Bush, whose first year included 9/11. I think the odds of another 9/11 are mercifully low, and the trend appears to be toward lower approval for presidents in their first year in recent times.

Moreover, Trump is unusually loathed by a huge segment of the population and is promising massive tariffs that I suspect will prove unpopular once they start raising the prices of everyday household items. So I feel pretty good predicting Trump will be below water at year’s end. —DM

Metaculus aggregated forecast: 17 percent

“Still friends” is obviously a subjective category but I like the prediction markets guru Nathan Young’s proposed definition: if one or the other publicly and unambiguously disparages his counterpart at least three times. Luckily for us, Trump and Musk are not subtle or taciturn men, and when they dislike someone they have a tendency to scream that loudly many, many times, so I don’t anticipate it being hard to decide where they stand at the end of 2025.

The list of one-time Trump allies who eventually came to denounce him is too long to include in full here, but let us briefly remember, say, 10: Anthony Scaramucci; Mike Pence; John Kelly; John Bolton; HR McMaster; Stephanie Grisham; Alyssa Farah Griffin; Betsy DeVos; and of course Michael Cohen.

It does not seem like an ambitious prediction that Musk will eventually join their ranks. His role as the head of the new Department of Government Efficiency seems guaranteed to put him on a collision course with Trump’s Cabinet officials and with congressional Republicans, and probably also with his cochair Vivek Ramaswamy. Trump might side with Musk each time — but he’s always been more pragmatic about spending policy than the cut-happy Musk seems, and there are ripe opportunities for conflict.

What if Musk wants to slash Medicare and Social Security, which Trump has promised to defend? What if he wants defense cuts and Trump wants a tougher posture toward China? What if Musk pushes for reconciliation with China, with whose government he is extremely close (Ramaswamy once called Musk “a circus monkey” working for Xi Jinping)?

I won’t predict the exact inciting episode that causes Trump and Musk to fall out. But I feel like I know how these guys operate, and I think it’s more likely than not that they will fall out. —DM

Metaculus aggregated forecast: 35 percent

Last year, I correctly predicted that more than 40,000 Americans would be killed by cars in 2023 (according to the National Highway Traffic Safety Administration’s estimates, which are released with a lag the following year). Since the 1960s, the US has seen rapid, dramatic progress in cutting its car fatality rate, and 2007 was the last year that over 40,000 Americans were killed by our car-dependent transportation system — until the Covid-19 pandemic.

You would think that fewer people driving would mean fewer car crash deaths, but not so in America, where our dangerously designed roads lead to more speeding and death when there’s less traffic. Ever since, America’s rate of death by cars has sat at levels that should honestly be humiliating for such a rich country.

These numbers are slowly starting to come back down. NHTSA recently estimated that for the first half of 2024, car crash deaths were down 3.2 percent from 2023. If the same trend from 2023 carries over to the second half of 2024, total 2024 car fatalities will come in at a hair under 40,000. It’s far from guaranteed, because car crash patterns vary significantly across different seasons. And that number would still be nothing to write home about — but in a country so thoroughly built around automobiles, getting deaths back under 40,000 would be a milestone worth celebrating. —Marina Bolotnikova

Metaculus aggregated forecast: 81 percent

What a difference a year makes. In December 2023, Israeli Prime Minister Benjamin Netanyahu was incredibly unpopular, his image severely damaged by his government’s total failure to anticipate the deadly October 7 attacks by Hamas. Polls indicated his Likud party might win only 17 of 120 seats in Israel’s Knesset. Israel was on its way to becoming an international pariah because of the destructive way it was waging its war in Gaza, and Israelis were furious about the government’s failure to rescue the hostages held by Hamas, even after a November 2023 deal to bring some home. Oh, and Netanyahu was only a few months removed from massive street protests and was facing corruption charges.

Fast-forward to December 2024, and polls suggest Netanyahu’s Likud party would win 25 seats if elections were held today, more than any other party. Israel has all but destroyed Hezbollah, by far its most capable opponent, and has isolated Iran, arguably its most existential threat. After the sudden fall of Syria’s Bashar al-Assad, Israel has even captured territory formerly under the Syrian government’s control. And President Joe Biden, who at least occasionally pushed back against Netanyahu, is about to be replaced by President-elect Donald Trump, who has signaled that he will happily give Israel a freer hand in Gaza.

Netanyahu has been prime minister of Israel for roughly 17 of the past 28 years. Every time it seems like he’s in an unwinnable position, he seems to find a way to wriggle out of it. I have every expectation that will continue in 2025. —BW

Metaculus aggregated forecast: 75 percent

Americans got pretty upset about inflation in the aftermath of the Covid pandemic, but we’ve got nothing on our Argentinian friends:

Here in the US, we’re such babies that we’ll complain about 6 percent to 7 percent inflation. In Argentina, double-digit annual inflation rates were normal even before the pandemic. Annual inflation hit triple digits and started what looked like an exponential climb, still ongoing when left-wing Peronist Alberto Fernández left office.

Javier Milei, a chainsaw-wielding self-described anarcho-capitalist who named his dogs after the libertarian economists Milton Friedman, Murray Rothbard, and Robert Lucas, initiated shock therapy upon taking office this year, eliminating price controls and subsidies for things like fuel and food, as well as massively devaluing the peso. That made prices surge even more massively at first (which you can see in the chart above). But since then they’ve been subsiding. The price surge had the salutary effect of easing the government’s debt burden, and the nation’s budget went into surplus for the first time since the 2008 financial crisis.

This has come at a considerable cost, with poverty and unemployment spiking, and the economy as a whole in recession for much of the year. But now that the recession is over and the country is seeing both cooling inflation and a growing economy, a sadly unusual combo down in Buenos Aires.

That said, I don’t think we’re going to see the country get down to a 30 percent annual inflation rate in 2024. The 12-month inflation rate in November 2024 was 166 percent, down 27 points from the month before. If the rate keeps falling at that pace, the country will hit the 30 percent mark in five months. But I think progress against inflation will slow as the initial shock of Milei’s policies subsides and pressure for wage hikes intensifies in a country that’s finally growing again. The IMF anticipates annual inflation hitting a low of 45 percent next year, and I think that’s a reasonable guess. —DM

Metaculus aggregated forecast: 55 percent

There will be a ceasefire in Ukraine (75 percent)

The war in Ukraine is just short of its third anniversary. The very fact that Ukraine has continued to fight this long defies most early prognosticators, many of whom expected the government in Kyiv to collapse not long after the Russians invaded. (An exception there, as Future Perfect readers know, is the State Department’s perspicacious Bureau of Intelligence and Research.) But the longer the war goes on, the more Russia’s sheer size and willingness to sacrifice unbelievable numbers of soldiers has outweighed Ukraine’s ability to fight back, even with the material support of the US and European allies.

President Biden has mostly been a steadfast ally, but he’ll be leaving office on January 20, replaced by Donald Trump, who has made no secret of the fact that he has little interest in continuing to support Ukraine. Both sides are still fighting hard to gain and protect territory, but it seems clear that’s being done by both Ukraine and Russia to put themselves in the best possible position before expected peace talks. Exactly what form that will take is difficult to predict, and a ceasefire doesn’t mean a permanent peace. But I would be shocked to not see a durable pause in the fighting some time in 2025. —BW

Metaculus aggregated forecast: 44 percent

Iran gets nuclear weapons (30 percent)

For the purposes of this prediction, “getting nuclear weapons” means producing enough fissile material to fuel a nuclear weapon. Actually producing a usable nuclear weapon — including miniaturizing a warhead enough to fit on a missile — might take another several months to a year or more, and thus probably falls outside the 2025 time frame.

Iran is already on the brink of sufficient enrichment — estimates are that it would only take about a week for Iran to enrich enough uranium for five fission weapons. So the question here is primarily one of international politics. Iran had a terrible 2024. It directly attacked Israel with missiles twice, only to see both salvos largely neutralized by missile defense systems, while Israel’s own retaliatory attack on Iran was far more successful. The Lebanese militia Hezbollah, Iran’s most powerful proxy, was all but annihilated by Israel, which continues to operate in parts of southern Lebanon. And the return of Donald Trump brings a president into office whom Iran has been accused of trying to assassinate.

Put that all together, and the Iranian regime finds itself in a very insecure place, and may look to nuclear weapons as a way to level the playing field. At the same time, relatively new Iranian President Masoud Pezeshkian has made overtures to the West and seems to understand that the only path to economic relief for his country is a new deal that limits the nuclear program in exchange for easing economic sanctions.

The Iranian regime’s number one priority is its own survival, and my best guess is that they will decide that the risk of going full speed on a nuclear program isn’t worth it, at least for another year. (There’s also the possibility that accelerating its nuclear work could lead to a military intervention by Israel or the US that would stop the program in its tracks.) So I think on balance that Iran won’t join the nuclear club in 2025 — though it’s not a prediction I make with a great deal of certainty. —BW

Metaculus aggregated forecast: 8 percent

Science and technology The World Health Organization (WHO) will declare H5N1 a pandemic in 2025 (25 percent)

First off, some math. While it may feel as if infectious disease pandemics have become a regular occurrence, they still remain highly rare. Since 1918, there have been five influenza pandemics: the Spanish flu of that year, the 1957 Asian flu, the 1968 Hong Kong flu, the 1977 Russian flu, and the 2009 swine flu. That gives a naive percentage of about 5 percent for any given year.

But there’s evidence that outbreaks of new infectious diseases are increasing, as the Covid pandemic amply demonstrated. And the H5N1 avian flu has been infecting a growing variety and number of animals, and more recently, people. On December 18, California, where 34 human cases of the virus have been detected, became the first state to institute a state of emergency over bird flu. New research suggests just a single mutation could be enough to potentially increase the virus’s ability to spread from person to person, which would be a prerequisite to becoming a pandemic. (Right now, H5N1 only rarely seems to be able to spread between people, and only in very limited fashion.)

So why am I mostly pessimistic about H5N1’s ability to truly break out, which for the purposes of this prediction would mean the WHO officially declaring it a pandemic, which would require sustained human transmission over multiple regions? Not because we’re doing a great job containing it — we definitely are not. Rather, it’s personal experience.

I’ve been covering H5N1 since it began really spilling over in Southeast Asia in 2004. I’ve been to backyard chicken farms in Indonesia and virology labs in Hong Kong. I’ve watched this virus as closely as any other subject I’ve covered in nearly a quarter-century as a professional journalist, and I just don’t think H5N1 has it. Call it a hunch, and one I hope will hold true. —BW

Metaculus aggregated forecast: 26 percent

A major lab will formally claim it has achieved AGI (30 percent)

For precision, let me clarify that by “major lab” I mean any of the following companies:

OpenAI Anthropic Google (including DeepMind) Microsoft Nvidia xAI Meta/Facebook Mistral Databricks World Labs Safe Superintelligence Hugging Face Scale AI Magic.dev Amazon Apple Netflix IBM

This is a purposefully broad list and includes companies that haven’t made it a priority to be on the bleeding edge of deep learning (like Netflix) and ones whose primary business isn’t in developing their own models so much as hosting or enabling models that others create (like Scale or Hugging Face). But, you know, I thought Nvidia wasn’t in this race until it dropped a massive model in October with impressive benchmarks, so a lot of things can change quickly in the world of AI.

Artificial general intelligence (AGI) is a vague term, and there is a large and growing literature in which AI researchers seek merely to define it, let alone to predict what it would look like or mean. That said, most definitions rely on an analogy to humans: an AI will be generally intelligent if it can do everything a human being can do, as well as a human being can, including meta-tasks like learning to complete new tasks.

This idea itself has holes in it. Different human beings can do different things — I cannot do everything, say, Katie Ledecky can do.

Luckily for us, the prediction here doesn’t require us to know what AGI means. It just requires a major firm to claim to have achieved it, accurately or not. One OpenAI staffer took to X this past year to claim that the firm’s models had already gotten there (though, importantly, the company itself has not made claims that grand).

So if the bar is that low, why do I think we’ll make it through the year without a company making this claim? Mostly because a) this is a young field where firm reputation matters a ton and being discredited by a premature AGI announcement might make the difference between a company ending up like Apple and ending up like Atari, and b) this is the kind of technology where premature claims can be discredited really, really fast.

If a nuclear fusion company claims to have achieved net energy gain, it is very difficult for me, a non-nuclear physicist, to tell if they’re bluffing. It’s not like I can use the nuclear reactor. But an AGI would presumably come with text, video, audio, and other interfaces that average consumers could try out and use, and it’d be immediately clear if some AI firm claimed to have gotten there when they hadn’t. —DM

Metaculus aggregated forecast: 20 percent

EVs will make up more than 10 percent of new car sales in the US by the end of Q3 2025 (65 percent)

I’m not really a fan of private cars (see my other car-related prediction above), and I wish our solution to climate change was to just have fewer of them. But this is America, so we have to work within the maddeningly car-dependent cage of our own creation. Electric cars are obviously better for society in most respects (though not all — their heavy weight means they’re more dangerous to pedestrians, cyclists, and anyone outside the vehicle), so I grudgingly have to welcome the EV transition that’s finally picking up.

By the third quarter of 2024, EVs made up 8.9 percent of new car sales in the US, according to an analysis by Kelley Blue Book. There’ve been reports that electric car sales are slowing, but given their consistent past growth rates, plus the fact that interest rates are coming down, I think we’ll hit 10 percent by the same time this year without much trouble. Donald Trump’s promise to do sweeping tariffs could throw a wrench in all that, but given my colleague Dylan Matthews’s prediction about the unlikelihood of that happening, I won’t calibrate my prediction around it too much. —MB

Metaculus aggregated forecast: 67 percent

Bitcoin’s price will at some point in 2025 breach $200,000 (70 percent)

The digital gold rush probably still hasn’t reached full frenzy, believe it or not. Bitcoin recently topped $100,000 in value for the first time, but we are going to have to think bigger. I think it’s going to double its value in the next 12 months — and I’m not the only one.

This is not an endorsement, to be clear; I own no bitcoin. When I read Warren Buffett still believes Bitcoin is a fad, that people do “stupid things” and this will be the latest trend in fiscal speculation to end in failure, I take it seriously. The case for bitcoin remains muddied, at least to me.

But Buffett has also compared bitcoin to gambling and, well, the gambling business is booming. Even if Bitcoin is a questionable long-term investment — don’t forget it dropped below $20,000 in late 2022 — people can still get a kick out of the continued accumulation of value, and that’s the basis of any bubble. It helps that Donald Trump and the crew he’s bringing back to Washington have gone all in on the crypto craze; they are likely to do whatever they can to stoke the speculation further.

Bitcoin just saw its value more than double over the course of 2024. I think it’s more likely than not it’ll repeat the feat with those winds at its back. —Dylan Scott

Metaculus aggregated forecast: 20 percent

Elon Musk is still the richest person in the world (55 percent)

My source here is the Bloomberg Billionaires Index. Since the 2012 inception of the Bloomberg list, the occupant of the top spot has changed five times. In 2013, surging Microsoft shares enabled Bill Gates to beat out Mexican telecom mogul Carlos Slim, who led the list at its outset. In 2017, Amazon shares put Jeff Bezos ahead of Gates. The massive rally around Tesla led in January 2021 to Elon Musk deposing Bezos. But Louis Vuitton chief Bernard Arnault overtook him in October 2022 in part because Musk had to sell much of his Tesla fortune to finance his purchase of Twitter. But by May of the following year, Musk was back on top.

“Will Elon Musk still be the richest person in the world throughout 2024?” is actually two separate questions: one, will he be dethroned by anyone in the next year; two, will he live through the year? Normally the latter wouldn’t be a concern for a 53-year-old man with access to the best health care money can buy, but Musk is, uh, not the most stable person on earth. So I’d put maybe a 5 percent probability he loses the title by way of the Grim Reaper.

Five switches in the top ranking over 12.5 years of the Bloomberg ranking existing implies, naively, a 40 percent chance that the top rank will switch in any given year. There’s tons of fluctuation within the top 10 even in a given month, as these net worths are hugely dependent on stock returns.

Jensen Huang of Nvidia, currently ranked 12th, gained $73.4 billion in 2024 alone, by far the biggest part of his $117 billion fortune. It's sobering to return to the 2013 article on Gates overtaking Slim, which notes at the

Открыть больше