Infrastructure is invisible until it fails. A bridge, a power grid, a water system—we notice them only when they break. The same is true for teacher support. The structures that help educators enter the profession, stay in it, and thrive are often taken for granted until they collapse under pressure.
In 2026, that pressure is immense. Political headwinds, funding instability, and rapid technological change continue to test the resilience of the STEM education ecosystem. But beneath the turbulence, something quieter is happening: organizations across our network are not simply reacting to disruption; they’re building systems designed to last. Not just programs, but infrastructure—financial models, mentorship networks, belonging practices, and AI frameworks that can hold up over time. This is work that rarely makes the news but determines whether the STEM teacher workforce strengthens or fractures in the years ahead.
As a network of over 200 organizations, Beyond100K has a unique vantage point. We see patterns invisible to any single organization—what’s emerging, what’s eroding, and where the field is placing its bets. This year’s Trends Report captures what we’re hearing: the field turning from pilot programs to permanent structures, from individual interventions to collective infrastructure. These four trends reflect that turn—and illuminate a tension running beneath them all.
The financial burden of becoming a teacher is not new—it has topped our list of challenges for years. What’s new in 2026 is the convergence of policy debates, funding models, executive orders, and legislative decisions that are quietly but powerfully reshaping who can enter and remain in the profession.
Across our network, partners continue to report that cost is the primary obstacle to a robust teacher pipeline—cited as the #1 challenge in more than ⅔ of discovery calls, conversations we had with over two dozen educator preparation program leaders, advocates, and funders of this work. Candidates struggle to afford preparation programs; clinical practice often means months without income; and once in the classroom, teachers find compensation tied to zipcode, embedding structural inequality into the profession itself. As one participant in network programming asked: “Can we reimagine how we pay educator salaries so increasing them isn’t solely dependent on property taxes?”
Yet 2026 also brings unexpected momentum. Going into our discovery calls, we wondered whether support for apprenticeships would be controversial: Traditional preparation programs and alternative pathways have not always seen eye to eye. We found the opposite. The UTeach Institute & Western Governors University are implementing apprenticeships to earn while you learn, and Prepared To Teach is advocating for them as a scalable pathway to the classroom. Austin Peay State University, which launched the first federally registered teacher apprenticeship, has now expanded to eight school districts with over 180 participants and received a $1M state grant in December 2025. The National Center for Grow Your Own is working in over 30 states, with more than 40 participating in the National Registered Apprenticeship in Teaching Network.
This approach is gaining traction faster than anyone predicted, in no small part due to broad bipartisan support at the local and national levels. But it comes with a risk we must name: the potential de-professionalization of teaching. Recent legislation reclassified teaching as a graduate rather than professional program, restricting borrowing limits for federal loans and raising questions about the profession’s standing. As public comment periods and policy shifts unfold, the field must advocate for financial access and professional standing.
Consider how Alabama A&M University, a UTeach site, frames the opportunity: “One degree, two career paths.” As one faculty member explained, “You are still going to be a chemist. You’ll just be a chemist who teaches students in the K-12 setting how to be a future chemist.” This reframe—teaching as a STEM career, not a departure from one—may not only be a way of drawing more STEM candidates into teaching, it may prove an essential hedge in a labor market that is shifting radically in light of automation.
Looking ahead: 2026 is a policy year. The choices made now—about apprenticeship funding, licensure pathways, and teaching’s professional status—will shape the pipeline for a decade or more. The door is open; the question is who we help walk through it.
We’re long past debating whether mentorship matters. Research shows it can cut teacher attrition by more than half, particularly when it goes beyond instructional coaching to include professional growth, leadership development, and community-building. The question now is different: How do we embed mentorship at scale without losing quality, trust, or relevance?
This is an infrastructure challenge, not an interpersonal one. And in 2026, we’re seeing the model evolve.
Traditional mentorship remains valuable—content-aligned support, culturally responsive matching, and external mentors with time and training. But the 1:1 structure that has defined mentorship for decades carries an inherent limitation: it is difficult to scale. One mentor can only support so many teachers well. Even the strongest programs eventually hit capacity.
In response, partners are increasingly experimenting with something more dynamic and communal: near-peer support groups, cohort-based learning communities, affinity spaces where educators with shared identities support one another, and peer coaching networks that distribute mentorship responsibility across many relationships rather than concentrating it in one. These models preserve relational depth while unlocking scale, expanding the circle of support without overburdening individual mentors.
While often not labeled “mentorship“ in the traditional sense, programs like the Black Teacher Project’s Design Lab function as mentorship infrastructure: they create sustained affinity spaces where Black educators learn from one another, build leadership capacity, and develop practices collectively. Support is distributed across a community and an experience rather than concentrated in a single mentor relationship. At the Metropolitan State University of Denver, leaders are exploring near-peer mentorship as a lever to strengthen teacher candidates’ confidence and sense of belonging in STEM. Drawing on the research of Dr. Freeman Hrabowski, whose work on near-peer mentorship has transformed persistence in STEM professions, the university is asking a powerful question: Why don’t we treat teacher preparation the same way we treat STEM training? By normalizing uncertainty, modeling growth, and providing real-time affirmation, near peers help candidates internalize not just the skills of teaching, but the identity of being a STEM teacher.
This evolution reflects a recognition that no single mentor can meet all of a new teacher’s needs—and that peers at similar career stages often provide uniquely relevant support. As one high school computer science teacher told us: “I wish I had somebody there that truly understands what I’m going through. And doesn’t just tell me why—but says, ‘let’s figure out together.’” Near-peer and cohort-based models shift the equation: instead of one mentor carrying the full weight of support, multiple relationships create a web of guidance that is more responsive, sustainable, and scalable.
What doesn’t work is also becoming clearer: teachers “voluntold” to mentor without support or compensation; generic mentorship without STEM expertise; one-time matching without ongoing structure; relying solely on one mentor when teachers need a network of support.
Looking ahead: In 2026, expect more programs designing peer cohort support alongside traditional mentorship, increased attention to affinity-based support networks, and new compensation models that value peer leadership. The question “Who is your mentor?” may give way to “What is your mentorship ecosystem?”
In 2021, when Beyond100K’s unCommission pinpointed the centrality of belonging for student persistence and joy in STEM, it was a novel insight. In 2026, it’s become not just a rallying cry but something more concrete: an operational priority with structures, metrics, and institutional commitment.
Across our network, educator belonging and wellness has emerged as a non-negotiable retention driver. In 2025 Measurement CoLab launched our Belonging Question Bank as a resource for adoption and adaptation by our network community. Nashville Teacher Residency names it explicitly as their #1 priority, building affinity spaces into program design and extending community-building into graduates’ school environments—and they’ve doubled their cohort size for 2025-26. EnCorps pairs experienced educators with new CTE teachers, providing individualized guidance that teachers report leads to “greater confidence in instructional planning, increased feelings of support and belonging.” At our recent program-year launch event, when asked for one word that captures 2026, the majority wrote “belonging.”
What distinguishes 2026 is the shift from belonging as aspiration to belonging as infrastructure. Partners are asking harder questions: What are the early indicators of burnout or disengagement, and how do we proactively respond rather than react later? What conditions must be in place for belonging to be sustained, not just declared? A report from UpBeat cites belonging as critical to teacher retention, often facilitated and supported directly by school administration and leadership. As one Beyond100K program manager observed, “Belonging is a prerequisite to doing anything around diversity, equity, and inclusion. It is the glue that connects diversity, equity, and inclusion together.” At the same time, this shift is butting up against real infrastructure challenges. As one partner shared with us: “Administrators are charged with cultivating the professional ecosystems where belonging can thrive; however, they are rarely given the tools or mandates to make this their priority.”
This shift is happening alongside—and partly because of—a significant language pivot. In the current political climate, many organizations are finding that “belonging” can do work that “DEI” can no longer do publicly. Project Team (Beyond100K short-cycle innovation teams) members discussed this explicitly: “We shifted to really thinking about belonging and centering belonging… [The group is] focused on belonging and how belonging gets cultivated through instructional resources.” This is not abandonment of values; it’s strategic adaptation. But it raises a question partners are wrestling with openly: “Whose safety are we prioritizing with the use or non-use of certain language?”
There is also emerging interest in whether AI-powered tools might help—not to replace human connection, but to help leaders identify early signs of disengagement and respond before teachers leave. The key conditions: such tools must reduce cognitive load, not add to it, and must be built on trust and aligned with real needs.
Looking ahead: We predict that measuring belonging will become an important leading indicator of retention rates. Organizations that fail to operationalize belonging—to move it from mission statement to daily practice—will continue to lose teachers.
The urgency around AI literacy is real. Teachers need to understand AI—not just how to use it, but how to teach it with discernment, ethics, and critical thinking in age-appropriate ways, so students are ready for a world where AI will be ubiquitous.
But the struggle to respond to AI exposes a pattern we’ve seen before: School systems racing to address urgent or trendy priorities while crowding out joyful learning, enduring skills, and non-tested subjects. The pattern disproportionately affects kids in higher-poverty schools, who are more likely to be taught to consume new tools — but not how they work or how to build them.
According to the 2018 National Survey of Science and Mathematics Education, only 17% of K-3 teachers teach science all or most days—compared to 99% for math. The average time on science? Just 18 minutes per day, versus 89 for reading and 57 for math. The 2024 NAEP results show the consequences: 8th-grade science scores fell, with 38% of students below basic proficiency.
Now AI is the new urgent priority. Districts providing AI training doubled between 2023 and 2024—from 23% to 48%—and a 2025 executive order mandates AI professional development “across all subjects.” This attention is warranted.
But history reminds us that when a new priority captures scarce leadership focus and professional learning time, other long-standing instructional goals can lose momentum (remember “the science of reading”), not because they lack importance, but because systems have limited bandwidth.
The question, then, isn’t whether AI deserves attention. It does. The question is how we use this moment. If districts are already revisiting curriculum, investing in professional development, and rethinking instructional practice in response to AI, this is an opportunity to strengthen, not sideline, deeper goals like STEM access, student belonging, and rigorous inquiry. Rather than treating AI as a standalone initiative layered onto already stretched systems, we can leverage its urgency to address persistent challenges in engagement, equity, and instructional coherence. If the attention is here, let’s use it to buoy all boats.
Across our network, partners are finding better ways forward. Lipscomb University and Arizona State University’s Mary Lou Fulton Teachers’ College are teaching candidates AI-powered strategies to enhance daily instruction. A network for K-12 systems, launching in early 2026 with The Lawrence Hall of Science, is developing practitioner-facing tools to help leaders and teachers decide “when AI genuinely improves outcomes and when restraint is the wiser choice”—prioritizing pilots in under-resourced and multilingual-focused districts, schools, and classrooms. And at the state level, 32 states have piloted statewide guidelines, programs, and evaluation frameworks, ensuring that guidance reaches classrooms before confusion does. California’s emerging framework, in particular, offers robust direction for local education agencies on teaching about AI, interacting with AI tools, and training educators to use AI responsibly and effectively.
Looking ahead: The programs that thrive will be those that integrate technology exposure during teacher preparation, treat AI literacy as part of broader STEM fluency (not a replacement for it), and teach AI literacy in service of students navigating an uncertain future with confidence, curiosity, and connection to what matters.
If there is a meta-trend running beneath the others, it is this: The organizations in our network have spent the past year adapting—to political headwinds, funding instability, and shifting expectations. They have found new language when old language became risky. They have pursued new funding models when traditional sources dried up. They have built new structures when existing ones proved inadequate.
The scale of disruption is significant, and it’s not slowing down. Government funding made up approximately 33% of nonprofit revenue before recent cuts. Philanthropic giving to education has continued its decade-long decline. And as one partner told us: “We have funders that have decided [educational opportunity] isn’t an area they want to invest in anymore… it’s not like they directly say, ‘Oh, it’s because we don’t care about diversity,’ but it’s sort of in there.”
Partners have demonstrated remarkable creativity in response. They are deepening local partnerships, exploring earned revenue models, pursuing bipartisan state funding, and leaning into collaboration—recognizing that more can be gained together than through competition for limited resources.
As we look across the field, there are some questions we should all be asking: What has been preserved? What is at risk of being erased? And is there a third way — adaptation without erasure? As organizations shift from “equity” to “belonging,” how will we preserve the substance without losing the specificity of who has been excluded and why? How will we lower financial barriers to teaching without de-professionalizing a field that has fought for recognition? In the midst of so much upheaval, we must find the space and time to collectively decide how we preserve hard-fought progress, what new realities demand adaptation, and where we need new approaches to meet the challenges — and opportunities — of this moment.
