The Problem: Teacher Time Crisis in K8 Math

Four million American fourth graders (40% of the cohort) lack basic math proficiency. Without automatic recall of basic facts, these students will struggle with algebra, stall in STEM courses, and face limited prospects in an AI-driven economy. Cognitive science is clear: when children spend working memory on finger-counting, they cannot simultaneously hold multi-step problems in mind. Learning slows, confidence drops, and math becomes something to survive rather than understand.

By late 2026, the educational AI market will likely consolidate around a few commercial platforms. Districts will face multi-year contracts, premium pricing of up to $240 per student annually, and proprietary systems cannot easily audit or exit. The window to build a public alternative that serves all students, not just affluent ones, is closing fast.

The problem is urgent. The timing is critical. And the commercial solutions flooding the market are structurally misaligned with what K–8 students and teachers actually need.

XtraMath offers a different answer: AI that amplifies classroom teachers instead of replacing them, operates at public infrastructure costs instead of premium prices, and extends proven impact to millions more students before commercial lock-in eliminates alternatives.

This is the case for funding a public alternative while we still can.

What Teachers Actually Need

Mr. Marker teaches sixth grade in Mount Vernon, Washington. He is thoughtful, skilled, and deeply committed. Yet math time is still chaos.

Some students are working on subtraction. Others finished division weeks ago and sat waiting for more work. He spends hours every week analyzing data, creating groups, planning interventions, and writing progress notes.

He once said, “I know how to teach. I just can’t clone myself.”

That’s the real problem. Not that teachers are unqualified, but that there aren’t enough hours in the day to be as effective as they already know how to be.

Teachers need AI to handle the cognitive overhead that steals time from instruction and relationships in the classroom.

Why Commercial AI Tutors Won’t Solve This

The current AI gold rush in education follows a predictable pattern driven by venture capital economics that are fundamentally misaligned with public school needs.

1. Pedagogically Wrong for Early Grades
   Commercial AI tutors imagine a child alone with a screen, interacting with a chatbot. For K–8 learners, this sidelines the essential role of peer collaboration, teacher guidance, and social learning. Teachers become tech support for opaque systems they cannot customize. Community needs such as multilingual support, cultural relevance, and offline access become afterthoughts.
   Before students develop internal confidence through social learning, they are not ready to critically engage with AI. A child still finger-counting will simply obey what the bot tells them, trying to mimic the artificial teacher to earn rewards. This stifles curiosity and agency. ‍
   ‍
2. Misaligned Economics
   Venture-backed companies must deliver tenfold returns within five to seven years. That economic reality drives pricing like $240 per student per year—rational for firms burning millions on large models but excluding Title I schools. Market pressure pushes toward wealthy districts, not those with the greatest need.
   ‍
3. ‍ Vendor Lock-In by Design
   Proprietary Systems trained on public data to create switching costs and data moats. As computer costs rise and network effects compound, districts face escalating fees with few alternatives. We have seen this pattern in every prior EdTech cycle.

This is not a critique of AI itself. It is recognition that commercial incentives and K–8 pedagogy point in opposite directions.

The XtraMath Alternative: Shared AI Infrastructure for K–8 Math Educators

For fifteen years, XtraMath has done steady, essential work on behalf of teachers: daily math practice that builds automaticity. Nothing flashy. No games or gimmicks. Just adaptive practice that meets each student where they are.

We now reach 6.2 million students and over 500,000 teachers—about one in six U.S. elementary classrooms.

We have remained nonprofit throughout, operating at roughly $0.25 per student per year. We rely on revenue earned, offer a free tier for everyone, and have never taken venture capital. Our work is sustainable public infrastructure, not a commercial product.

And it works. Ninety-four percent of teachers report higher test scores, and 87 percent report reduced math anxiety.

Now we are extending that infrastructure with AI that works behind the scenes: analyzing the 1.3 billion minutes of practice students already complete each year, grouping students for interventions, documenting progress, and recommending next steps. Teachers reclaim time for what matters most.

This is not an AI tutor. There is no chatbot, no new login, no extra screen time.

Instead, teachers see AI-drafted summaries, groupings, and recommendations directly within their existing XtraMath dashboard. They can review, edit, and approve in seconds instead of spending hours.

How It Works: Three Layers Behind the Scenes

Layer 1: Specialized Teaching Assistants
Rather than one large model attempting every task, we build specialized agents optimized for specific functions, like:

• Report Analyst reads each student’s activity, flag patterns, and generates clear summaries.
• Standards Mapper connects performance to specific standards using the Learning Commons Knowledge Graph.
• Grouping Assistant clusters students by mastery profile and suggests differentiated instruction groups.
• Intervention Planner recommends evidence-based activities, timelines, and open resources.
• Documentation Writer drafts notes for parents and administrators in teacher-ready language.

Layer 2: Inference and Governance
Our architecture uses efficient small models to ensure both performance and affordability. District-level audit logs track all AI interactions. Human-in-the-loop workflows ensure educators review and approve outputs. Compliance and transparency are built into the foundation.

Layer 3: Knowledge Commons
A vectorized skill graph connects K–8 math standards, open educational resources, and evidence-based interventions. This shared layer allows interoperability with other trusted EdTech tools.

The Monday Morning Difference

Picture Mr. Marker on Monday morning. Instead of spending two hours reading reports:

• 30 seconds reviewing a summary highlighting five students who need attention
• 2 minutes accepting suggested small-group plans and making one quick edit
• 3 minutes reviewing three intervention drafts and adding brief personal notes
• 5 minutes approving weekly summaries for principals and parents

Ten minutes instead of two hours. The AI doesn’t replace his expertise; it amplifies it.

Why XtraMath Is Uniquely Positioned

1. Trust at Scale
We already serve millions of students through hundreds of thousands of teachers who choose our platform because it delivers results without manipulation. This network provides immediate distribution and rapid feedback loops. Schools do not need to adopt something new; we are enhancing infrastructure they already trust.

2. Proven Foundation
Our platform processes millions of students' practice exams each month with sub-second latency. Adding AI capabilities for teachers builds on that base. Advances in small language models now make it possible to deliver AI performance at a fraction of the cost of large commercial systems.

3. Affordable Sustainability
Fifteen years of operation have proven that educational infrastructure can scale through efficiency rather than extraction. We operate at ~$0.25 per student annually: evidence that public technology can be both effective and affordable.

What Success Looks Like

Efficiency Gains

Teachers reduce non-instructional workload by up to 90%, freeing 4–8 hours per week for direct instruction.

Learning Outcomes

Students receive better-targeted interventions and more teacher attention. Districts can track automaticity gains and dosage-to-outcome relationships with pre/post measures.

Equity at Scale
AI-enhanced support reaches under-resourced schools and Title I districts. Every teacher gets data-informed help previously available only in wealthy schools.

Our roadmap is outcome-based, not feature-based. We aim to double our student reach while maintaining flat operating costs. Guardrails are non-negotiable: open standards, auditability, and district data control.

The Choice Before Us

By 2026, network effects will favor a handful of commercial AI platforms. Districts will face rising costs and limited choice. We have roughly 18 months to establish a public alternative.

XtraMath is not competing with chatbot tutors. We are building the infrastructure public education actually needs: transparent, affordable, teacher-centered AI that operates at scale.

The question is not whether AI will reshape K–8 math education. It will.
The question is who it will serve.

Will AI widen achievement gaps with premium-priced tools for the privileged, or narrow them through shared public infrastructure that serves every child? Your support will decide.

Join Us

We are seeking:

• Philanthropic partners committed to public infrastructure over privatized solutions
• Research collaborators advancing transparent AI in education
• Policy leaders shaping standards for auditable AI in schools
• EdTech allies seeking open, affordable alternatives to commercial LLMs

Operating at pennies per student is not a gimmick—it is the point. It keeps public options truly public.

A third-grade teacher told us, “My students get through their other math work much faster now.” That small sentence captures a big truth: when fluency is in place, everything else opens up. That is what digital public infrastructure should do—make the important things easier, efficiently, and at scale.

The choice is simple but irreversible. Either we allow extractive providers to dominate educational AI, or we fund a public alternative that ensures universal access to essential math education.

We are building the latter. Will you join us?

Choose Your Impact

Your philanthropic dollar accelerates access and innovation while earned revenue from district subscriptions sustains operations.

1. ‍$500 supports AI-enhanced coaching for one school ‍
2. $50,000 funds a full-year implementation in a Title I district
3. ‍$1,000,000 enables a state-level deployment and evaluation framework that informs national policy

XtraMath is a 501(c)(3) public charity. For fifteen years, we have served millions of students with one mission: to deliver proven, trusted, and affordable tools for universal math fluency.

Roy King

CEO, Executive Director

‍