Prisms.

A classroom in Pomona, and a tsunami

Somewhere in California right now, a 14-year-old is standing inside a tsunami. She is not on a beach. She is in a classroom, in a Meta Quest, watching a wave that is taller than her math teacher form, crest, and come down on a coastline she has never visited. Before the lesson ends, she will have estimated its volume, predicted its run-up, and - almost as an afterthought - written down an exponential function. The function is the easy part. The standing-inside-the-wave part is the lesson.

This is what Prisms looks like in 2026. Not a slide deck about virtual reality in education. Not a pitch. A wave, and a kid, and a piece of math that finally lands.

What Anurupa Ganguly saw before she built anything

The official story of American math education is that it is broken. The unofficial story is that it has been broken for so long the brokenness now looks like the floorplan. Kids memorize procedures. They forget them. They sit for a standardized test. They forget those. Somewhere in the middle, a generation decides that math is a thing that happens to them rather than a thing they do.

Anurupa Ganguly spent more than a decade inside that floorplan. She trained as an engineer at MIT, then taught math and physics, then ran the secondary math office for Boston Public Schools, then served as senior director of teaching and learning at the New York City Department of Education. That is roughly every seat at the table - student, teacher, district leader - and from each one she watched the same thing: students arriving at the symbol before they had the experience.

She decided the order was backwards.

The unfashionable case for a VR headset in school

In 2020, "VR in the classroom" was something between a cliché and a punchline. Most edtech pitches that used the phrase produced novelty - a butterfly garden, a planet, a heart you could rotate. Engagement up, learning flat. Ganguly's bet was different. She argued that VR was not a visualization trick. It was an embodied cognition tool: a way to give students a body inside a problem before asking them to abstract it.

If that sounds suspiciously like a learning-sciences paper, that is because it was one first. The National Science Foundation funded early prototype work. MIT's Venture Mentoring Service helped with the rest. By the time the rest of the world remembered what the Oculus Quest was, Prisms had a curriculum.

What you actually do inside a Prisms lesson

You put on a Meta Quest. A teacher sets the scenario. You walk - really walk - through a public health outbreak, a population curve, a coastal flood model, a chemistry reaction. You manipulate variables with your hands. Sometimes you fail the scenario and watch the curve go vertical. Sometimes you keep it flat. Either way, you are inside the function before you ever name it.

Then comes the symbolic part: the equation, the graph, the teacher pulling the class back to the whiteboard, the pencil work that, according to everything we used to think about math class, should have come first. Prisms inverts the order. The kid abstracts up from experience, not down from notation.

For teachers, there is a companion web dashboard with learning analytics, real-time progress, and curriculum alignment - the deeply unglamorous infrastructure that makes the deeply glamorous part work at scale.

The numbers, which are doing some heavy lifting

Edtech is a graveyard of charming demos. The graveyard is full because the demos rarely move scores. Prisms' early data is, to its credit, doing the unfashionable thing of moving scores. An independent study found an 11% lift in algebra performance. The Gates Foundation is funding the larger, multistate version of that question, which is the kind of question only takes a few years and several million dollars to answer well.

Built by teachers, which is more rare than it sounds

Most edtech is built by people who left the classroom. Prisms is built by people who, in many cases, never quite did. Ganguly has been explicit about this: the company is unusually full of former educators, and the product reflects it. Lessons start from instructional problems, not engineering ones.

There is also a quieter, more political claim underneath the company: that immersive technology should not be a luxury experience reserved for kids whose schools can afford it. Prisms has gone out of its way to ship into rural Kentucky and Ohio districts as well as urban systems. Educational equity is in the brand documents. It is also, less politely, in the customer list.

The end of the worksheet, or at least a serious dent in it

The honest tension at the heart of Prisms is not whether VR works in classrooms. The honest tension is whether American schools - underfunded, exhausted, and politically combustible - can absorb a new modality fast enough to matter. Hardware costs are falling. Teacher trust is, slowly, rising. The roadmap is widening from middle school into early college, and from math into the rest of the natural sciences.

If Prisms is right - if the order really is experience first, symbol second - then the worksheet is not going away because someone disliked it. It is going away because something simply teaches better. That is a quieter kind of revolution than edtech is used to. It is also harder to fake.

Back to Pomona

The student in Pomona takes off the headset. The wave is gone. The classroom is back. On the whiteboard, the teacher writes y = a·b^x, and for the first time in this kid's school career, those letters are not strangers. They are a souvenir from a place she just walked through.

This is what Prisms is doing. It is not making math fun, which is the lowest bar in edtech and arguably a slur. It is making math real first. The fun, when it shows up, is a side effect of finally getting to do the thing.

Links worth keeping