We love reading the tea leaves and making predictions about the future. Here are our top five predictions (including which startups we think will fail this year, which will be acquired, and why the AI backlash is beginning).
There are downsides to every technology. Fire kept us warm, but also burned down our villages. - Ray Kurzweil
1. Learning Interrupted
The pandemic disrupted education around the world. 1.2 million students left the public educational system. The average grade 3–8 student lost half a year in math and a quarter of year in reading. This year is already a tragic testament to a new disruptor: Climate-related disasters.
In October, Storm Trami disrupted learning for 19 million students in the Philippines, my motherland. In the US, Hurricane Helene led to a full month of disrupted learning for students in the Southeast. Recent snow storms in the region mean that my co-worker’s two children, based in North Carolina, have had six weeks of near continuous learning disruption.
In Los Angeles, the devastating fires have led to school closures for at least half a million students. Over 20% of campuses (more than 200) in the district were closed. At least 9 schools have been destroyed.

We’re less than two weeks into 2025, and the LA fires are estimated to have caused more than $135B in damages, according to the BBC. That would be 74% of the total domestic climate-related damage from 2024–which was the 2nd worst year on record, behind 2023. 2025 might blow that away, just like we’ve blown past the temperature forecasts predicted by the Paris accords.
We predict that climate and weather disasters will disrupt education for many students across the country–underscoring the need for virtual emergency schooling.
2. The AI Backlash
I remember riding the NYC subway a few months ago and seeing the laziest f***ing ad that made me irrationally angry. It was for a furniture delivery company called Kaiyo.

AI has unleashed a torrent of horrible, tedious content. Yes, suddenly anyone can create an ad or video in two seconds, but that doesn’t mean that everyone should. I am tired of hearing the same emotionless AI voice on social media. I am tired of seeing clickbait ads everywhere.
I am tired of people saying that this thing will revolutionize everything. Yes, it has made writing a copy easier. It has made it easier for students to look up answers. But it is a very poor replacement for human beings, which is what the techno-futurists keep positing.
Dan Meyer wrote an excellent post last week slamming Unbound Academy (check out the Mathworlds post here). Unbound says they’re leveraging AI to replace teachers, in their stupidly grifty 2-hour learning model. The problem, as Dan points, is that they still hire teachers; they just call them “guides” (complete with the teacher certification requirements and worse student:teacher ratios).
I’m sick of technology promising to revolutionize XYZ, when all it’s actually doing is undermining paying family-supporting wages. We wrote about this in “Pizza Making Robots,” but deep-tech startups raised hundreds of millions of dollars promising…well, pizza making robots. AI startups are raising billions of dollars, and the best they can do is…rename "teachers” as “guides.”
3. Emergency Certifications
Teachers are quitting. There is no one to replace them. We predict that teachers are going to continue aging out or quitting, and states are going to continue lowering the bar for who can be a teacher. According to EdWeek, 260,000 educators are under-qualified (lacking the credentials or certifications normally required to teach in their state). NJ voted last year to remove the Praxis Basic Skills test (that same test disqualified 1,000 prospective teachers from pursuing that career in Louisiana in 2022).
While many states still refer to the Praxis Basic Skills test, Indiana, Alabama and Montana have made it central to their certification processes. We predict that their state legislatures will introduce bills removing that requirement, especially in Indiana, where the teacher shortage is more acute than the other states.
4. Die, Zombies!
If you were hoping for lower interest rates in 2025, you’ll be disappointed. A strong jobs report last week and an economy still humming along mean that rates will be elevated for the foreseeable future. We predict the reemergence of the zombie firm. Startups that raised capital or debt in a low interest environment who now have to put all of their revenue to servicing that debt.
Last year, Byju’s and 2U collapsed in the wake of their huge debt pileups. This year, it’ll be the next batch. Who’s next? Preply and GoStudent both closed on debt financing in the past two years (debt financing is only attractive in an interest rate environment like ours when valuations have collapsed). GoStudent raised $95M in debt and equity in 2023; Preply, $70M that same year. Instead of reinvesting, GoStudent and Preply will have to repay the debt portions of their financing. Brutal.
Multiverse closed a $220M funding round in 2022 at a $1.7B valuation: They burned through cash and closed their US operations. The UK economy is sputtering, meanwhile. No one will invest in Multiverse at their current valuation, and Multiverse won’t go public anytime soon (unless they want to demolish shareholder value). Paper is even worse shape. They raised $270M in a Series D in 2022, but let go of 45% of their staff last year…including their CEO.
We don’t think Paper survives the year.
5. Mergers and Acquisitions
Want to know who can come to the rescue? Private equity! Bain acquired PowerSchool last year. The chart below shows increasing PE deal flow relative to 2023 and 2022, both in deal value and deal count.
We predict that public market liquidity events will remain muted in 2025. A few EdTech firms might go public in 2025, and we will be astonished each time we hear about it.
Who do we think is primed for an acquisition? Upskilling firms which proliferated during the pandemic and Great Resignation. Maybe not Guild or Multiverse, but maybe Workera, Pathrise, Cybrery, etc.