EdTech Funding Fell 89%. What Still Gets Funded Looks Very Different

Reach Capital, which EdWeek Market Brief calls one of the most active venture capital firms in K-12, said on August 18, 2026, that it had closed a $265 million fifth fund, about $50 million larger than its fourth and oversubscribed^[1][2]. That sounds like a comeback for education technology. The details say otherwise. The fund will back AI applications in learning, health and work, and EdWeek Market Brief reports that the firm is extending its reach beyond K-12 as district spending contracts^[1][2].

When a specialist education investor widens its brief, it is worth asking what happened to its home market.

From $20.8 billion to $2.4 billion

Research firm HolonIQ counted $20.8 billion of EdTech venture capital worldwide in 2021, across more than 1,500 funding rounds, 61 of them worth $100 million or more^[3]. By 2024 the annual total was $2.4 billion, an 89% fall from the peak and the lowest level in a decade^[4].

The recovery since has been faint. Funding rose roughly 11% to $2.6 billion in 2025, then slipped again: the first half of 2026 brought $1 billion, 26% less than the same period of 2025^[5][6]. For scale, one 2021 round, the $1.5 billion Series A raised by US company Articulate, was by our arithmetic half as large again as everything HolonIQ counted for EdTech startups worldwide in the first six months of 2026^[3][6].

Valuations followed. HolonIQ listed 32 EdTech unicorns, private companies valued above $1 billion, worth a combined $97 billion at the end of 2021. In May 2026 it listed 14, worth $34.2 billion^[3][7]. Not every departure was a failure: PhysicsWallah left through an IPO, while GoStudent, once valued at $3.2 billion, was marked below $1 billion^[7].

Other trackers count differently but tell the same story: Tracxn data cited by Rest of World show a 2021 peak of $16.7 billion and less than $3 billion in 2025. The same report says 645 EdTech companies were launched in 2025, against almost 10,500 in 2020^[8].

Why the money left

Three things broke at once. The first was China. Rules introduced in 2021 barred companies teaching the school curriculum from making profits, listing or taking foreign capital, and HolonIQ put that year's drop in Chinese EdTech investment at $8 billion^[3]. Zuoyebang, which raised $750 million in a single round in 2020, now competes in China's market for AI learning hardware^[16][8].

The second was the end of emergency demand. Reach partner Steve Kupfer told EdWeek Market Brief that US K-12 faces a perfect storm: enrollment falling with the birth rate, the federal stimulus era over, and health care, pension and energy costs rising^[2].

The third was the fate of some of the sector's prominent names. Byju's, once valued at $22 billion, was revalued by major investors at under $1 billion and came apart in a financial crisis^[7][8]. 2U, which paid $800 million for edX in 2021, filed for Chapter 11 bankruptcy in July 2024 under debt it could not carry^[9]. Chegg said in October 2025 that it would cut 388 roles, about 45% of its workforce, citing AI and reduced traffic from Google, and recast itself as a business-to-business skills company^[10]. HolonIQ says venture investors are still waiting on returns from their pandemic-era bets^[6].

What still gets funded

The companies now raising money or being acquired share four traits.

• Employers and institutions pay. Workforce training was the largest category in 2025 at 38% of deals and drew the majority of venture funding in the first half of 2026, with Preply's $150 million round, at a $1.2 billion valuation, among the notable raises^[5][6][7]. Workday bought Sana, an AI learning platform used inside companies, for $1.1 billion^[5]. Coursera and Udemy completed their merger in May 2026, promising a skills platform for the AI era^[15].
• AI is the base, not the pitch. MagicSchool, which raised $45 million in 2025, sells mainly to school districts that want a governed environment for generative AI^[5][13]. Founder Adeel Khan told Crunchbase News that "the model is no longer the differentiator"; what counts is applying it to real problems with specific expertise^[13].
• A person stays in the loop. Reach has backed services businesses built around people, on the view that technology should strengthen the student-teacher relationship rather than stand in for it^[2].
• Rounds are smaller and earlier. Reach plans checks of $1 million to $10 million into roughly 50 companies, mostly at pre-seed and seed^[1][2]. Across the market, deal volume held stable in the first half of 2026 while average deal size shrank^[6].

Evidence becomes the price of entry

HolonIQ expects capital in 2026 to favor products that deliver measurable, well-defined outcomes^[5]. Buyers are pushing the same way. Instructure's 2026 Edtech Top 40, drawing on usage by more than 12 million people in US K-12 through its Canvas platform, found that districts have access to an average of 3,001 digital tools, while students and educators interact with only four on average^[11].

The yardstick is the federal Every Student Succeeds Act (ESSA), which requires certain federal funds to support evidence-based interventions and grades research from Level I, the strongest, to Level IV. Of the top 40 tools, 52.5% had at least Level IV evidence, which by our arithmetic leaves 47.5% below even the lowest tier^[11][12]. A separate Instructure review of 150 widely used classroom technologies, produced with the nonprofit InnovateEDU, found ESSA-aligned evidence for 40% of purpose-built education tools and 2% of consumer technologies used in classrooms^[12]. Instructure sells Canvas, and the Top 40 draws on usage inside its own ecosystem, so these are a vendor's figures, but they point the same way as the investors.

One academic trial shows what clearing the bar can look like. In a randomized controlled trial with 900 tutors and 1,800 K-12 students, reported in a preprint, researchers Rose Wang, Susanna Loeb, Dora Demszky and colleagues found that students whose tutors had access to Tutor CoPilot, an AI assistant the team built, were 4 percentage points more likely to master topics, at a cost of $20 per tutor a year^[14]. The study covers live tutoring only, and tutors flagged that some suggestions were not grade-level appropriate. Still, it tests the model investors now favor: AI that supports a human.

The case against a comeback

None of this amounts to a recovery. Despite the interest in AI, EdTech venture funding fell again in the first half of 2026^[6]. Kupfer, whose firm backs AI founders, says many AI education tools are being commoditized because they offer little that a general-purpose AI account does not^[1][2].

“software exclusively is never going to move the needle in these systems”

Schools are also adopting AI cautiously. The only large language model among Instructure's 40 most-used K-12 tools was Google Gemini, in 38th place^[11]. And Chegg is a reminder that AI can shrink an education business as well as build one^[10].

The WAC view

The reset corrected a business model, not the case for technology in education. What capital now rewards, tools that fit institutional workflows, keep teachers and tutors at the center and can show independent proof of impact, is close to what good practice required all along. The risk is that a thin market funds too few experiments in schools and universities, where budgets are under the most pressure. Institutions that test tools rigorously and publish what they find will have unusual influence over what gets built next.

What to do with this

The events behind this article, in WAC News

• Reach Capital closes $265 million fifth fund for AI in learning, health and workAugust 18, 2026
• Workday agrees to acquire Swedish AI learning company Sana for $1.1 billionSeptember 16, 2025
• Coursera and Udemy agree to combine in all-stock mergerDecember 17, 2025
• MagicSchool raises $45 million Series B for AI tools for teachersFebruary 11, 2025
• Preply raises $150 million Series D at $1.2 billion valuationJanuary 21, 2026
• Zuoyebang raises $750 million in Series E round led by Tiger Global and FountainVestJune 29, 2020
• 2U agrees to acquire edX from Harvard and MIT for $800 millionJune 29, 2021

Sources

1. Reach Capital raises $265M Fund V to back AI founders building to 'expand human potential'. TechCrunch, August 18, 2026.
2. As K-12 Faces a 'Perfect Storm,' a Prolific Ed-Tech VC Changes Tack. EdWeek Market Brief, August 21, 2026.
3. Global EdTech Venture Capital Report - Full Year 2021. HolonIQ, January 2, 2022.
4. EdTech VC reached ~$2.4B for 2024, representing the lowest level of investment in a decade. HolonIQ, January 15, 2025.
5. EdTech hits $2.6B in investment as the market stabilizes. Bigger bets in AI and workforce training.. HolonIQ, February 6, 2026.
6. $1B in EdTech Venture Capital for 1H. Funding falls short of last year's midpoint. Asia & MENA buck the trend.. HolonIQ, July 14, 2026.
7. Global EdTech Unicorns: The Complete List of Global EdTech Unicorns. HolonIQ, May 18, 2026.
8. The global edtech boom is fading as investors look elsewhere (browser title: Edtech's pandemic boom is over as K-12 startup funding craters). Rest of World, April 23, 2026.
9. Long-embattled 2U declares bankruptcy. Inside Higher Ed, July 26, 2024.
10. Chegg to Remain a Standalone Public Company to Maximize Shareholder Value. Chegg (Business Wire press release, via Nasdaq), October 27, 2025.
11. New Instructure Data Shows K-12 Districts Are Demanding Evidence, Not Just Access to Edtech Tools (2026 Edtech Top 40). Instructure (PR Newswire), June 29, 2026.
12. As AI and Screen Time Scrutiny Rise, Instructure's 2026 Evidence Report Finds Most Classroom Consumer Technology Lacks Verified Proof of Impact. Instructure and InnovateEDU (PR Newswire), March 10, 2026.
13. 'Nobody Wanted to Give A Former Principal Money': How An Educator Built An Edtech AI Startup With $63M From VCs. Crunchbase News, August 5, 2026.
14. Tutor CoPilot: A Human-AI Approach for Scaling Real-Time Expertise (Wang, Ribeiro, Robinson, Loeb, Demszky). arXiv, October 3, 2024.
15. Coursera, Udemy Complete Their Merger. Inside Higher Ed, May 14, 2026.
16. Chinese online learning app Zuoyebang raises $750M. TechCrunch, June 29, 2020.

Written by the WAC editorial team. Facts are drawn from the numbered sources above and were checked against them before publication; opinions appear only under “The WAC view”. The organisations and people named are not affiliated with the World Assessment Council unless stated. Send a correction