Predictions are a terrible business. If you get it right, the events leading up to the prediction render it less bold. If you get it wrong, you’ll be reminded of your gaffe 10k times a day (i.e., Twitter). The purpose isn’t really to be right, in fact, but to catalyze a conversation. Every year, we make predictions. We start by holding ourselves accountable. Here are our predictions for 2024, followed by our predictions for 2025.
Power Couple: OpenVidia
Since the launch of ChatGPT in November 2022, investors have added a staggering $8.2 trillion to the market valuations of tech’s Big Six firms: Alphabet, Amazon, Apple, Meta, Microsoft, and Nvidia. For context, the 2024 federal budget was $6.8 trillion. Companies that referenced AI during their earnings calls registered a 12% increase, on average, in performance, compared to a 9% increase for those that didn’t mention it.
The AI ecosystem is settling into three layers: applications (Duolingo, Netflix, Tesla), AI models (Anthropic, Gemini, OpenAI), and infrastructure (AWS, Google Cloud, Nvidia). Two companies dominate. OpenAI has doubled its annualized revenue to $3.4 billion in the past six months. And its ChatGPT accounts for 56% of premium LLM subscriptions, i.e., people pulling out their credit cards. Over the past 12 months, Nvidia has reported $96 billion in revenue — 4x its 2022 total. I look at peer-reviewed research to evaluate whether a technology is enduring: Nvidia chips are cited in 19x more research than those of its competitors combined. For two companies to dominate a technology this early is extraordinary.
The AI Company of 2025: Meta
No business is better positioned to register progress in AI than Meta. Nine out of 10 internet users (excluding China) are active on Meta platforms. The company has access to more unique human language data, i.e. raw training data, than Google Search, Reddit, Wikipedia, and X combined. In terms of compute, Meta has purchased more Nvidia Hopper GPUs (advanced AI hardware) than any U.S. company other than Microsoft, giving it unmatched AI training and deployment capacity.
Palindrome: Service-as-a-Software
So far, the benefits of AI have accrued to existing players. The next set of winners will be firms that capitalize on service-as-a-software, i.e., taking human-intensive services and putting a thick layer of AI on top to scale with less labor. This is a fancy way of saying there will be more consumer-facing AI applications. The real cabbage, however, is in routinizing back-office functions (e.g., accounting, compliance, customer service, etc.).
Technology of 2025: Nuclear
AI’s chokepoint is energy. A ChatGPT query demands 10x the energy of a Google query. The majority of the 10 most valuable companies in 1980 and 2024 were/are in energy and tech. However, the construction of acres of data centers and the energy investments needed to power them reflect a deeper convergence. AI is accelerating Big Tech’s transformation from an industry that sells computers into an industry that sells compute. In a knowledge economy, compute is energy.
Wind and solar are great, but they lack the scale and reliability of nuclear power. One nuclear reactor produces the equivalent of 800 wind turbines, or 8.5 million solar panels. Nuclear is also carbon-free: 48% of the clean energy in the U.S. comes from nuclear. Nuclear power may be the worst-managed brand in history. Every energy source has tradeoffs in emissions and externalities. I believe nuclear energy represents the best trade. If you gathered all the used nuclear fuel produced by the U.S. in the last 60 years, it would occupy only 10 yards of a football field. (Note: Do not go anywhere near that field.)
Get Used to It: Drones
Radar, jet engines, nuclear power, GPS, and blood banks were all developed during wartime. There’s something about war, and the potential loss of a civilization, that inspires creativity. At the outset of the war in Ukraine, Russia’s defense budget and standing army were 10x and 5x the size of Ukraine’s, respectively. Drones are the premier technological innovation birthed by the conflict.
Drones provide constant surveillance capabilities and enable precision strikes at a fraction of traditional costs. A successful drone strike can yield a 100,000% return (e.g., $400 drones routinely destroy $4m tanks). 3D printing, AI, and micro-cameras have converged to shape the latest David vs. Goliath sequel. Using drones for last-mile delivery of Comtrex and commuters, search and rescue missions, and monitoring and maintenance in manufacturing and agriculture should reap substantial gains.
Musk Bids for Warner Bros. Discovery/CNN … Or Another Iconic Media Firm
The Wall Street Journal reported that Elon is addicted to Ketamine. I believe that’s the delivery mechanism, but the nicotine (where his real addiction resides) is attention. For 10% of his net worth ($44b for Twitter) he can impose himself on all of us, nearly all the fucking time. Q: If he’s going to come undone, can’t he do it like the rest of us, in private?
Anyway, WBD has a market cap of $26b (plus debt). If the idea sounds outrageous, it isn’t. John Stankey (CEO, AT&T) put a condition on the sale of WBD that it had to be a single class of stock, to get the greatest price and net the company a takeover premium; in the words of Gordon Gecko, WBD is breakable, i.e., it can be acquired. After his fallout w/Trump, and the public’s increasing fatigue (Jesus, make it/him just go away) threatens to push him out of the spotlight, Elon will force himself back into the news cycle by (again) becoming the news. He could also buy MSNBC, as (unlike MSNBC) he does have a sense of humor.
Investment Opportunity: Emerging Markets
The S&P 500 outperformed Vanguard’s All-World ex-U.S. index ETF +56% to +23%, respectively, from 2023 through 2024. Historically, when U.S. equities fall, emerging markets rise. These cycles typically last about a decade. I believe we’re (over)due for a course correction. The U.S. stock market now makes up 50% of the total market cap globally; when stocks get this expensive, returns go down, and capital looks for greater returns elsewhere. Since 1989, emerging markets have typically outperformed developed markets by 27% after a Fed rate cut. Demographics are destiny; the growth in working-age populations favors India, Indonesia, and other developing nations. The share of institutional capital invested in the markets is at a cyclical low. A reversion to the mean would represent inflows of $910 billion to emerging markets.
The X factor is Trump. He’s called for a 10% to 20% tariff on all imports and a 60% to 100% tariff on goods from China. I don’t believe he’ll follow through, though, as tariff is Latin for tax. At the first hint of inflation, alarm bells will sound and the adults in the administration, looking at the bond market, will respond crisply and force the administration to slow their roll. And Republicans in Congress will find their backbones when they realize that 90% of the presents under the Christmas tree come from China, and their dear leader is, post-2026, a lame duck.
Platform: YouTube
Netflix didn’t win the streaming wars, YouTube did. Last year, YouTube, which spends zero dollars on content — it shares revenue with creators instead of paying them — became the first streaming platform to reach 10% of all television viewing. Eighty-one percent of Gen Alpha viewers said they watched YouTube recently, compared to 62% who said they watched a subscription streaming service, and 44% who said they watched TikTok. In the U.S. and U.K., one-third of kids aged 8 to 12 said YouTube was their No. 1 career choice; movie star didn’t make the list. Also, YouTube is the No. 1 podcast platform, adding a tailwind no other streamer has. If Alphabet were forced to spin off YouTube, the company would likely be worth half a trillion dollars, vs. Netflix’s market cap of $350 billion.
Media: Podcasts
I’m talking my own book here, but I’ve been in the podcasting business for almost a decade, and this is the first time I’ve called it the media of the year. The only ad-supported medium growing as fast as Meta, TikTok, Alphabet, and Reddit is podcasting. Of the estimated 3.2 million pods, 600k put out content each week, and I estimate only 600 are economically viable. This is a striking concentration of power, with the top 10 pods commanding 35% of the listenership. Kamala Harris would have needed to appear on CNN, Fox, and MSNBC three hours every night during prime time for two weeks to reach as many people as Donald Trump did going on Joe Rogan.
Podcasts’ share of attention is well ahead of their share of ad revenue. This delta will close. Since the election, our pods have seen a 30% increase in revenue. My prediction is that pods’ ad revenue will grow by 20+% in 2025. Listenership will continue to grow as well, and the ARPU, like those of Meta and Alphabet, will increase dramatically as advertisers discover this is where young, successful consumers have been hiding.
IPO: Shein (Disclosure: Investor)
One-third of Gen-Z consumers say they’re “addicted” to fast fashion. Traditional retailers release 100 new styles a week. Fast fashion retailers put out 100 styles per day. Shein pushes out 7,000 styles per day. Its operations are remarkably asset-light, as Shein is an IP business that doesn’t own any factories, trucks, or stores. Instead, its software tracks activity on the site, sends orders to factories based on their ability to calibrate demand, and then puts in motion the transportation. Also, there are effectively no returns (the Achilles’ heel of any retail business), as the products are so cheap people don’t go through the hassle of sending them back. Similar to other asset-light winners (e.g., Airbnb, Nvidia, Uber), Shein’s revenue per employee dwarfs that of the incumbents.
Business Trend: M&A
A historic amount of cash is on the sidelines. Since 2003, private equity’s dry powder, i.e., the committed capital not yet allocated, increased 8x to $4 trillion. Corporate cash holdings total $4.1 trillion. Context: U.S. GDP is around $27 trillion.
The average closing time for U.S. dealmakers in 2022 was 161 days, a 14% increase since 2018. For deals exceeding $10 billion in value, closing times have surged by 66%, to an average of 323 days. Over the past four years, Lina Khan has been an aggressive antitrust enforcer, and the Biden administration has published 209 “economically significant” regulations — more than any president since Reagan. The lesson? Elections have consequences. Setting aside whatever grievances Trump may hold against specific tech and media companies, the perception is that his administration will likely be more friendly to M&A. Some predictions re who will be on top of some big transactions: Comcast, Uber, and (see above) Musk. Also, I believe someone will take Intel and/or Boeing private.
Tech Movement: Banning Phones
When we look back on this age, the thing we’ll regret most is letting our kids become addicts. The substance is social media, the delivery mechanism is the phone. On a typical day, a teen receives 237 notifications. One study found that 97% of kids use their phone during school hours for a median of about 43 minutes per day. Think about that: Basically every teen in America misses 10% of school every day.
Giving students unrestricted access to phones has been a great move, said no teacher ever. Banning them in school is a return to sanity. The good news: 18 states have passed laws restricting the use of phones in school, and roughly three-quarters of schools have policies restricting their use in the classroom. Better news: Our response, while slow, is bipartisan. Best news: Test scores have improved by 6% in schools that have banned phones.
Chemical: Testosterone
Women are ascendant (something to celebrate), while young men are struggling. There has never been a cohort that’s fallen further, faster than young men living in Western democracies. The percentage of young men aged 20 to 24 who are neither in school nor working has tripled since 1980. Workforce participation among men has fallen below 90%, while median hourly wages are $3 less per hour, adjusted for inflation, than they were in 1970. This is deadly; over the past 20 years, America’s incremental deaths of despair totaled 414,000, exceeding the 407,000 Americans killed in World War II. It’s also a mating crisis, as women traditionally mate horizontally and up socioeconomically, whereas men mate horizontally and down. When the pool of horizontal-and-up young men shrinks, there are fewer mating opportunities. And without the guardrails of a relationship, young men behave as if they have … no guardrails.
Families feel this. I believe the 2024 election was about struggling young people, especially struggling young men. If your son is in the basement vaping and playing video games, you don’t really care about trans rights or Ukraine, you just want change, i.e., chaos and disruption. The Trump campaign saw this and flew into the manosphere with coarse language, crypto, Rogan, UFC, and Hulk Hogan. Trump gained 15% with young men — the biggest pivot from Democrats to Republicans of any age group. Another big shift was among women aged 45 to 64. I believe those are the mothers of struggling young men. America elected President T; the “T” stands for testosterone. The election was supposed to be a referendum on women’s rights. It was instead a referendum on failing young men.
2025 Will Be a Great Year for You
How do I know this? A: I don’t. However, I do (really) hope your year is full of prosperity and time with loved ones. I’ve read that if you write a goal down, it’s 40% more likely to happen. And I’ve done this — see two sentences ago. So we have that going for us.
Life is so rich,
P.S. All our podcasts are available on YouTube, including Office Hours, where I answer your questions.
P.P.S. Section CEO Greg Shove is talking to Reid Hoffman (the guy who founded LinkedIn) about AI’s ability to unlock our superpowers on Jan 29. Don’t miss this one.
