Venture capital directed $20.8 billion into EdTech in 2021, dwarfing the research budgets of the world’s largest school systems (HolonIQ, 2022). By 2024, annual investment had contracted to $2.4 billion, the lowest level in a decade (HolonIQ, 2025). First-quarter 2025 brought $410 million globally, a 35% decline from the same period a year earlier (HolonIQ, 2025). Capital receded not because confidence in the underlying technology evaporated. Two decades of funding had failed to yield any convincing proof of returns.

The tension between EdTech’s investment appeal and its fulfillment record predates the most recent cycle by decades. The pandemic appeared to resolve it.

Between 2020 and 2021, the industry operated under historically unprecedented circumstances. Campuses went remote and emergency government funding flooded the market. No competing alternatives existed. Mandatory adoption pushed platforms to accommodate waves of new users. A definitive affirmation of value appeared to be within reach.

The retrospective findings proved unfavorable and the failure was not confined to a single country or metric.

The U.S. National Assessment of Educational Progress recorded a five-point decline in reading and a seven-point decline in mathematics among nine-year-olds in its 2022 Long-Term Trend assessment. The reading drop was the largest since 1990. The mathematics decline was the first on record (NCES, 2022). The lowest-performing students absorbed the steepest losses. Reading declined ten points and mathematics fell by twelve, while scores at the 90th percentile slid two or three (NCES, 2022). OECD’s 2022 PISA assessment, covering approximately 690,000 students across 81 countries, documented a fifteen-point decline in mathematics, equivalent to three-quarters of a school year’s instruction (OECD, 2023). A meta-analysis published in Nature Human Behaviour by Betthäuser, Bach-Mortensen, and Engzell (2023), drawing on 42 studies across 15 countries, found students lost approximately 35% of a normal year’s learning.

Recovery proved elusive. Betthäuser observed that students had, “largely been unable to recover these learning deficits.” Losses arose early in the pandemic and persisted through the study period.

The data carried a pointed implication. Jack, Halloran, Okun, and Oster, publishing in the American Economic Review: Insights (2023), compared student outcomes across U.S. school districts by instructional mode. Math pass rates declined 12.8 percentage points and ELA pass rates fell 6.8 from 2019 to 2021. Districts with full in-person instruction saw math scores decline 13.4 percentage points less than districts relying on hybrid or virtual delivery.

In São Paulo State, Lichand and colleagues found that students taught remotely retained only 27.5% of what students learn in person, while dropout risk increased by 365% (Nature Human Behaviour, 2022). A follow-up study found the losses did not diminish when in-person classes resumed. The disparity increased after in-person classes resumed. The authors attributed the pattern to a sustained shock in student motivation and the latent factors behind learning (Lichand et al., PNAS, 2024).

Catherine Cheney, writing in Devex in 2021, described the COVID-19 pandemic as the moment that was supposed to validate educational technology. The validation did not arrive. Emiliana Vegas, writing for the Brookings Institution in March 2022, documented the institutional response. Countries reopening schools were, “going back to how education was delivered before the pandemic, instead of seizing the opportunity of the disruption to transform education.” Investors who had written checks on the premise that the pandemic would accelerate permanent adoption watched their wager vanish.

UNESCO’s 2023 study An Ed-Tech Tragedy? examined the promises of educational technology and the realities of what it delivered during school closures. The report identified, “numerous unintended and undesirable consequences,” from the hard pivot to remote courses. Reviewers noted the study exposed the consequences of outsourcing educational provision on unsupported technological promises rather than investing in teachers and institutional capacity.

Ahn (2022), using the ECLS-K:2011 longitudinal dataset, identified a compounding pattern. Higher-achieving students benefited, while lower-achieving students suffered additional losses. The result was a form of what the study termed “unfair divergence”, a newly documented widening of the third-level digital divide, which measures the differential in tangible outcomes derived from technology use. The finding was paradoxical. A core justification for the massive investment in educational technology had been bridging the first-level digital divide, or the unequal access to devices and connectivity. Access had been extended, but the gap between outcomes grew.

HolonIQ characterized the resulting investment landscape as a shift toward, “a more measured approach” (2025). Investors held capital in reserve until they could evaluate whether pandemic-era bets had justified the expenditure. A question deferred by venture capital for a decade had become impossible to avoid. Does any of this actually improve learning?

The research does not offer a clean answer and market players have long benefited from the ambiguity. Educational technology can improve measurable results under specific, well-documented circumstances. The 2023 UNESCO Global Education Monitoring Report, the most comprehensive international assessment of tool effectiveness published to date, identified the central constraint. The report stated, “learning benefits disappear if technology is used in excess or in the absence of a qualified teacher.” Distributing computers to students without active direction does not improve performance (UNESCO, 2023).

Rebecca Silverman and colleagues at Stanford’s Graduate School of Education conducted a meta-analysis of EdTech interventions in elementary literacy, covering 119 studies from 105 manuscripts published between 2010 and 2023. The results were not uniform. A Stanford press release accompanying the study noted that effectiveness, “varied considerably, depending on particular features of the interventions and the skills they targeted” (Silverman et al., 2024). Positive effects appeared across decoding, reading comprehension, and writing, but with a wide range. Design features mattered more than the category of product.

A separate line of evidence reached the opposite conclusion about technology’s aggregate impact. OECD’s analysis of PISA 2022 data found that students reported an average of two hours of daily digital device use for learning. In 21 OECD member systems, each additional hour beyond the first was associated with a decline in mathematics scores (OECD Education Policy Outlook, 2025). The World Bank’s 2023 East Asia Pacific EdTech report reinforced the point from a different angle. As the scale of interventions increased, their measured effect on learning decreased (Yarrow, Abbey, Shen, & Alyono, 2023). The pattern carried a structural implication for any industry organized around rapid, broad deployment.

The scenario reflects an important notion.

The most consistent predictor of success is a teacher who knows how to use the tool. A working paper from the National Bureau of Economic Research by Bulman and Fairlie reviewed the empirical pattern directly. Supplemental programs, or initiatives that supplied technology to complement teacher instruction, produced mixed results. Substitution programs, where technology replaced supervised study, indicated learning losses. Di Pietro and Castaño Muñoz (2025), analyzing 740 estimates from 72 studies for Computers & Education, found an overall positive, but small effect size of d = 0.202 and stressed that, “the simple provision of technology to schools does not show a significant effect.”

Effective implementation depends on a knowledgeable instructor, which makes the most successful applications the most labor-intensive. Products designed for direct student use without teacher involvement deploy broadly, but deliver the weakest results. The industry is centered upon what works least.

UNESCO’s 2023 report framed the problem at market level. The report concluded that, “sound, rigorous and impartial evidence of technology’s added value in learning is needed more than ever, but is lacking.” If the evidence came only from the companies marketing the products, bias was a structural risk (UNESCO, 2023). Isabelle Hau, Executive Director of the Stanford Accelerator for Learning, reported that only 7% of global educational technology had any form of rigorous evidence behind them. Only 11% of education decision makers were examining studies of any kind when purchasing a product.

LearnPlatform (by Instructure) found that of the 100 most-accessed K-12 solutions, only a quarter publicly disclosed research satisfying ESSA standards. Of the quarter that did, nine had studies that met ESSA’s “Promising” threshold or above. Kucirkova, Brod, and Gaab, publishing in npj Science of Learning (2023), drew on the same LearnPlatform data and confirmed the finding. The authors warned that when sales are prioritized over evidence, student achievement remains out of reach.

The U.S. Department of Education’s 2024 National Educational Technology Plan acknowledged the dynamic. Providers, “sometimes claim (without independent, research-based evidence) that student assessment results will soar if school systems adopt a given digital resource.” The plan went further by citing EdTech researcher Justin Reich. Reich observed that predictions of imminent transformation are, “among the most reliable refrains in the history of educational technology.” The federal government was not describing a fringe marketing practice. It was describing endemic behaviour.

Brookings Institution research by Carter-Rau and Olsen (2023) documented how unsubstantiated claims have driven momentum across low and middle-income countries. Government purchasing decisions in EdTech are not motivated by the extent of due diligence. Instead, influence remains with who recommends the product, how its deployment will appear politically, and the reputation of the company. Decision makers are, “bombarded by marketing from edtech companies,” whose pitches are, “sophisticated and include information that looks like solid evidence.” The dynamic is not confined to developing markets. The 74 Million reported in December 2025 that U.S. districts access an average of 2,739 distinct tools annually. The criteria for purchasing remain dominated by cost and vendor relationships.

The fall of BYJU’S remains the most notorious case of the industry’s shortcomings. The company entered insolvency proceedings in July 2024 after India’s National Company Law Tribunal accepted a petition from the Board of Control for Cricket in India. Three years earlier, the company had reached a $22 billion valuation. By October 2024, founder Byju Raveendran described the company as, “worth zero” (TechCrunch, 2024).

Archish Mittal and Daniel Pianko, writing for the Cornell SC Johnson College of Business in July 2024, addressed the lesson plainly. As they noted, “There are no easy fixes to educational attainment. Byju’s device, with some software, does not a scholar make.” BYJU’S had sold investors on the premise that capital deployed at scale could substitute for legitimate learning practices. The business model prioritized user acquisition over instructional integrity. It collapsed when acquisition costs proved unsustainable and the promised academic gains failed to materialize. BYJU’S was not an outlier. The post-2021 contraction dismantled a generation of EdTech companies.

Chegg, the homework-answer platform, lost 99% of its value from a pandemic peak of $14.7 billion. Shares fell 48% in a single day in May 2023 after CEO Dan Rosensweig told investors that ChatGPT was affecting new customer growth (CNBC, 2023). The company’s core product, a database of over sixty million pre-written homework answers, was a content-delivery model. It was vulnerable to displacement the moment a free alternative could perform the same function. Chegg cut more than half its workforce across two rounds in 2025.

2U, once valued at $5.5 billion, filed Chapter 11 bankruptcy in July 2024 with a market capitalization below $12 million (Higher Ed Dive, 2024). The company had purchased edX from Harvard and MIT for $800 million in 2021, funded by debt that became unsustainable as interest rates rose. Accumulated losses approached $1.6 billion. The online program management model that 2U had scaled, treating university courses as revenue vehicles to be operated by an external vendor, did not survive its own cost structure.

VIPKid, once valued at $4.5 billion with $975 million in total fundraising, was destroyed in 2021 by China’s “Double Reduction” policy banning foreign online tutoring. At its peak, VIPKid reported 700,000 to 800,000 paid students and approximately 100,000 North American tutors. Attempted pivots to international and adult markets failed. In India, Unacademy (valued at $3.44 billion in August 2021) conducted five rounds of layoffs totaling 2,000 positions. Indian EdTech funding fell from $4.73 billion in 2021 to $283 million in 2023. An estimated 2,150 Indian EdTech startups shut down between 2020 and 2024.

The failures share categorical traits. Profit was prioritized over student success and companies expanded before demonstrating product efficacy. Acquisition-fueled expansion was financed by debt that could not survive rising interest rates.

Content delivery was mistaken for learning. Steven Mintz, writing for Inside Higher Ed, articulated the misconception. Educational technology has not lived up to its promise because, “innovators too often embrace an impoverished conception of teaching and learning.”

The broader contraction confirms what the research had already established. The gap between what an EdTech solution promises and what it delivers is not an aberration. It is the prevailing standard.

The data explaining the failure is settled. A deficit of appropriate design and institutional accountability sustains it, not a deficit of knowledge.

However, a response is emerging and its early results are worth examining.

Through 2023 to 2025, nine or more districts across seven U.S. states piloted outcomes-based contracts for EdTech procurement, coordinated by the Southern Education Foundation’s Center for Outcomes-Based Contracting with support from Harvard’s Center for Education Policy Research and Stanford (Government Technology, 2024). Under the model, 40% of contract value is tied to measurable student outcomes. Implementation rates ran ten times above the national average for typical EdTech adoption. In Ector County ISD in Texas, vendors that met performance benchmarks received full payment, while underperformers received a reduced share of contract value. Superintendent Scott Muri stated, “From this point forward, this is how we’ll do it.”

Digital Promise launched an ESSA Tier 3 Product Certification in April 2024, requiring a logic model supported by empirical citations and a quantitative study aligned with ESSA standards. Owl Ventures, a dedicated EdTech investment fund, now operates an internal outcomes-and-efficacy function across more than 100 portfolio companies. NewSchools Venture Fund commissions external research reviews at approximately $40,000 per portfolio investment, partnering with WestEd and Empirical Education (NewSchools, n.d.). Visible.vc, a fundraising platform that tracks the industry, described the shift as the “Efficacy Reckoning”, noting that founders who cannot present reliable logic models or third-party research face steep valuation discounts during due diligence. Securing procurement contracts increasingly requires documented gains in student proficiency of at least 10%.

The developing accountability infrastructure remains limited. Among the 40 most-accessed tools in the LearnPlatform annual report, 32% met any ESSA threshold in 2024. The percentage rose to 45% in 2025. Natalia Kucirkova, writing for the World Economic Forum in April 2024, proposed shared baseline quality standards for EdTech firms worldwide and a multidimensional index for measuring their impact. She noted that investors, “rarely ask for measures of learning effects when making their investments”, preferring solutions with large user bases and prioritizing profitability over effectiveness. Kucirkova cited the assessment of global agencies and summarized their position as, “industry’s lack of accountability, characterized by scaling without evidence of positive impact, has resulted in the widespread distribution of low-quality products in schools.”

The educational technology emerging from the post-2021 contraction faces a defining question. Outcomes-based procurement, evidence certification, and investor-backed efficacy review represent the beginning of an accountability infrastructure. Whether the mechanisms reach the scale of the problem they were designed to address remains open. Capital will either organize around the requirements that research has confirmed are necessary or investment will coalesce around metrics that simulate progress without producing it. The research is not ambiguous about which path leads to learning. The industry’s willingness to follow it is.

References

Ahn, J. (2022). Exploring the negative and gap-widening effects of EdTech on young children’s learning achievement: Evidence from a longitudinal dataset of children in American K-3 classrooms. International Journal of Environmental Research and Public Health, 19(9), 5430. https://pmc.ncbi.nlm.nih.gov/articles/PMC9104322/ [Line 27]

Betthäuser, B. A., Bach-Mortensen, A. M., & Engzell, P. (2023). A systematic review and meta-analysis of the evidence on learning during the COVID-19 pandemic. Nature Human Behaviour, 7(3), 375-385. https://doi.org/10.1038/s41562-022-01506-4 [Lines 15, 17]

Bulman, G., & Fairlie, R. W. (2016). Technology and education: Computers, software, and the internet (NBER Working Paper No. 22237). National Bureau of Economic Research. https://www.nber.org/papers/w22237 [Line 39]

Carter-Rau, R., & Olsen, B. (2023). How to improve government decisionmaking around edtech innovations. Brookings Institution. https://www.brookings.edu/articles/how-to-improve-government-decisionmaking-around-edtech-innovations/ [Line 49]

Cheney, C. (2021, November 1). Ed tech, long overhyped, missed its moment amid COVID-19. What now? Devex. https://www.devex.com/news/ed-tech-long-overhyped-missed-its-moment-amid-covid-19-what-now-101835 [Line 23]

Di Pietro, G., & Castaño Muñoz, J. (2025). A meta-analysis on the effect of technology on the achievement of less advantaged students. Computers & Education, 226, Article 105197. https://doi.org/10.1016/j.compedu.2024.105197 [Line 39]

Government Technology. (2024, January 10). Leveraging ed-tech contracts for better student outcomes. Government Technology. https://www.govtech.com/education/k-12/leveraging-ed-tech-contracts-for-better-student-outcomes [Line 73]

Higher Ed Dive. (2024, July 26). A look at 2U’s path to bankruptcy. Higher Ed Dive. https://www.highereddive.com/news/2u-bankruptcy-restructuring-opms-education-department/722580/ [Line 59]

HolonIQ. (2022, January). Global EdTech venture capital report - Full year 2021. https://www.holoniq.com/notes/global-edtech-venture-capital-report-full-year-2021 [Line 7]

HolonIQ. (2025, January 15). EdTech VC reached ~$2.4B for 2024, representing the lowest level of investment in a decade. https://www.holoniq.com/notes/edtech-vc-reached-2-4b-for-2024-representing-the-lowest-level-of-investment-in-a-decade [Lines 7, 29]

HolonIQ. (2025, April 29). EdTech funding drops again in early 2025. Fewer deals, but bigger bets. https://www.holoniq.com/notes/edtech-funding-drops-again-in-early-2025-fewer-deals-but-bigger-bets [Line 7]

Jack, R., Halloran, C., Okun, J., & Oster, E. (2023). Pandemic schooling mode and student test scores: Evidence from US school districts. American Economic Review: Insights, 5(2), 173-190. https://www.aeaweb.org/articles?id=10.1257/aeri.20210748 [Line 19]

Kim, H. (2023, May 2). Chegg drops more than 40% after saying ChatGPT is killing its business. CNBC. https://www.cnbc.com/2023/05/02/chegg-drops-more-than-40percent-after-saying-chatgpt-is-killing-its-business.html [Line 57]

Kucirkova, N. (2024, April). Why measuring impact effectively is so important in EdTech. World Economic Forum. https://www.weforum.org/stories/2024/04/we-need-a-standardized-way-to-measure-the-impact-of-edtech-innovations/ [Line 77]

Kucirkova, N., Brod, G., & Gaab, N. (2023). Applying the science of learning to EdTech evidence evaluations using the EdTech Evidence Evaluation Routine (EVER). npj Science of Learning, 8(1), 35. https://doi.org/10.1038/s41539-023-00186-7 [Line 45]

LearnPlatform by Instructure. (2023). EdTech evidence report. Instructure. https://www.instructure.com/press-release/one-four-most-accessed-k-12-student-and-educator-focused-solutions-publicly-share [Line 45]

Lichand, G., Doria, C. A., Leal-Neto, O., & Cossi Fernandes, J. P. (2022). The impacts of remote learning in secondary education during the pandemic in Brazil. Nature Human Behaviour, 6(8), 1079–1086. https://doi.org/10.1038/s41562-022-01350-6 [Line 21]

Lichand, G., & Doria, C. A. (2024). The lasting impacts of remote learning in the absence of remedial policies: Evidence from Brazil. Proceedings of the National Academy of Sciences, 121(22), e2316300121. https://doi.org/10.1073/pnas.2316300121 [Line 21]

Mintz, S. (2021, March 22). Why most ed tech fails. Inside Higher Ed. https://www.insidehighered.com/blogs/higher-ed-gamma/why-most-ed-tech-fails [Line 67]

Mittal, A., & Pianko, D. (2024, July 1). What investors should learn from the fall of edtech unicorn Byju’s. Cornell SC Johnson College of Business Hub. https://business.cornell.edu/hub/2024/07/01/what-investors-should-learn-from-fall-edtech-unicorn-byjus/ [Line 53]

National Center for Education Statistics. (2022). NAEP long-term trend assessment results: Reading and mathematics. Nation’s Report Card. https://www.nationsreportcard.gov/highlights/ltt/2022/ [Line 15]

NewSchools Venture Fund. (n.d.). Is my ed tech tool making a difference? An entrepreneur’s guide to using research to improve products and measure impact. NewSchools. https://www.newschools.org/edtechresearch/ [Line 75]

Organisation for Economic Co-operation and Development. (2023). PISA 2022 results (Volume I): The state of learning and equity in education. OECD Publishing. https://www.oecd.org/en/publications/pisa-2022-results-volume-i_53f23881-en.html [Line 15]

Organisation for Economic Co-operation and Development. (2025). Education policy outlook 2025: Nurturing engaged and resilient lifelong learners in a world of digital transformation. OECD Publishing. https://www.oecd.org/en/publications/education-policy-outlook-2025_c3f402ba-en.html [Line 35]

Silverman, R. D., Keane, K., Darling-Hammond, E., & Khanna, S. (2024). The effects of educational technology interventions on literacy in elementary school: A meta-analysis. Review of Educational Research, 95(5), 972-1012. https://doi.org/10.3102/00346543241261073. Quote drawn from accompanying Stanford press release: https://acceleratelearning.stanford.edu/story/new-study-explores-what-makes-digital-learning-products-more-or-less-effective/ [Line 33]

TechCrunch. (2024, October 17). Byju’s founder says his edtech startup, once worth $22B, is now “worth zero.” TechCrunch. https://techcrunch.com/2024/10/17/byjus-founder-says-his-edtech-startup-once-worth-22b-is-now-worth-zero/ [Line 51]

The 74 Million. (2025, December). 2,739 ed tech tools later, where are the outcomes? The 74. https://www.the74million.org/article/2739-ed-tech-tools-later-where-are-the-outcomes/ [Line 49]

UNESCO. (2023a). An ed-tech tragedy? Educational technologies and school closures in the time of COVID-19 (M. West, Author). https://www.unesco.org/en/articles/ed-tech-tragedy-educational-technologies-and-school-closures-time-covid-19 [Line 25]

UNESCO. (2023b). Global education monitoring report, 2023: Technology in education: A tool on whose terms? https://www.unesco.org/en/articles/global-education-monitoring-report-2023-technology-education-tool-whose-terms [Lines 31, 43]

U.S. Department of Education, Office of Educational Technology. (2024). A call to action for closing the digital access, design, and use divides. 2024 National Educational Technology Plan. https://eric.ed.gov/?id=ED641164 [Line 47]

Vegas, E. (2022, March 11). Education technology post-COVID-19: A missed opportunity? Brookings Institution. https://www.brookings.edu/articles/education-technology-post-covid-19-a-missed-opportunity/ [Line 23]

Visible.vc. (2026). Top EdTech VCs: A founder’s guide to fundraising. https://visible.vc/blog/edtech-investors/ [Line 75]

Yarrow, N., Abbey, C., Shen, S., & Alyono, K. (2023). Using education technology to improve K-12 student learning in East Asia Pacific: Promises and limitations. World Bank. https://openknowledge.worldbank.org/entities/publication/2cb32acc-7e84-429e-b664-39575d0f5a34 [Line 35]

Disclaimer: The contents of this article have been edited and fact-checked using a large language model (LLM). Claims are made in good faith and are subject to correction.