Which AI education startup is growing the fastest?

Why is it so hard to name the fastest-growing AI education startup?

The fastest-growing AI education startup today depends heavily on whether we mean growth rate, new revenue added or adoption inside schools.

That distinction changes the winner. Pingo AI says it has reached roughly $12 million ARR and more than 8 million users after launching in January 2025. Speak has already crossed $100 million in annualized revenue, making it far larger. MagicSchool says it has reached around 7 million users and more than $10 million ARR while spreading through U.S. schools at extraordinary speed.

These companies are also selling into different markets. Speak and Pingo can acquire an individual learner, charge a subscription immediately and expand country by country. MagicSchool and SchoolAI have to navigate districts, teachers, student privacy rules and procurement, but a successful district deployment can put their products in front of thousands of people at once.

Even the published metrics are different. Speak talks about annualized revenue and downloads. MagicSchool reports users and district penetration. SchoolAI reports classrooms and teacher engagement. Synthesis talks about families. Comparing those numbers as though they represented the same thing would give us a meaningless ranking.

For this article, we treat “fastest-growing” literally first: which AI education startup is expanding its business fastest now? We then check whether that growth is large enough, durable enough and commercially real enough to matter.

That points us toward Pingo AI, with Speak and MagicSchool winning different versions of the same race.

What should count as real growth for an AI education startup?

Revenue growth is the cleanest starting point for comparing AI education startups, but we also need to know the size of the business underneath the percentage.

A company growing from $100,000 to $1 million has expanded tenfold. A company moving from $50 million to $100 million has only doubled, yet it created roughly $49 million more annualized revenue than the smaller company.

We therefore look at both the percentage and the absolute increase.

User growth matters too, especially when a startup has not disclosed revenue. But cumulative registrations can flatter a business badly. Someone who tried an AI tutor once eighteen months ago should carry less weight than a subscriber paying every month. “Classrooms reached” can also include very different levels of actual usage.

Paid adoption becomes particularly useful in education because free distribution is easy to generate when teachers are curious about AI. Getting families, schools or consumers to keep paying is harder.

That is why Pingo's current revenue trajectory deserves more weight than its 8 million users, why Speak's $100 million-plus annualized revenue matters more than its 15 million downloads, and why MagicSchool's $10 million-plus ARR makes its enormous teacher adoption much more convincing.

Metric	What it tells us	Main weakness
ARR or annualized revenue	Whether users are actually paying at scale	Run-rate figures can move quickly
Paid subscribers	How many users have crossed the payment barrier	Does not show spending per customer
Active usage	Whether the product has become a habit	Rarely disclosed consistently
Registered users or downloads	How widely the product has spread	Can include inactive and free users
Schools or classrooms reached	Institutional distribution	Does not necessarily mean active paid usage

Is Pingo AI the fastest-growing AI education startup right now?

Pingo AI currently has the strongest claim to being the fastest-growing AI education startup because its revenue has gone from almost nothing to roughly $12 million ARR in about a year and a half.

The trajectory is unusually steep even by consumer-AI standards.

Pingo launched publicly in January 2025. By March, the company had around 50,000 users and $25,000 in monthly recurring revenue. By July, it had passed 300,000 users and approximately $200,000 MRR. Around its Y Combinator Demo Day later that year, TechCrunch reported roughly $250,000 in monthly revenue and growth around 70% month over month.

The curve kept moving after the initial launch burst.

Earlier Pingo hiring material in 2026 described more than 3 million users, around $6 million ARR and roughly 50% month-over-month growth. Another version put ARR around $7 million. More recent company hiring copy now says Pingo serves more than 8 million users and has reached approximately $12 million ARR, still claiming around 50% monthly growth.

Even allowing for the fact that these are company-reported numbers, the sequence is hard to ignore. Pingo went from a $300,000 annualized run rate at its first meaningful milestone to roughly $12 million. That is about 40 times higher.

The current $12 million base also means Pingo can no longer be dismissed as a tiny app posting a spectacular percentage from almost zero.

If we interpret “fastest-growing” in the normal sense of the company whose business is expanding fastest now, Pingo is our answer.

If you want more recent data on this point, please see our latest AI in education market report.

How crazy is Pingo AI's growth compared with a normal startup?

Pingo AI's growth is extreme: the company appears to have compressed several years of normal consumer subscription growth into roughly eighteen months.

Consider the revenue progression. Around two months after launch, $25,000 MRR implied approximately $300,000 ARR. At $200,000 MRR four months later, the annualized run rate was already $2.4 million. More recent company figures put the business around $12 million ARR.

That means the annualized business expanded by roughly $11.7 million from its early $300,000 base.

The user curve was similarly sharp. Pingo went from 50,000 users to more than 8 million, an increase of roughly 160 times. User growth therefore ran even faster than revenue growth.

Its distribution helps explain how that happened. Pingo's marketing spread through short videos showing people having awkward, funny or deliberately blunt conversations with the AI. One widely circulated Pingo format turned language mistakes into entertainment rather than advertising the app like conventional education software.

Third-party analysis of its traffic also suggests that people frequently encounter Pingo elsewhere and then search directly for the brand. The company's website gets much more search traffic from variants of “Pingo” than from generic terms such as “Spanish AI,” which fits a creator-led acquisition model.

We should still keep one caveat attached to the $12 million figure: it comes from Pingo's own company material rather than audited accounts. But we are no longer looking at one isolated founder claim. Several snapshots of the company's hiring material show the revenue figure moving upward over time, from roughly $6 million to $7 million and then $12 million ARR.

That repeated progression gives us considerably more confidence in the underlying direction.

Is Speak actually growing more than Pingo AI in dollar terms?

Speak is still adding much more revenue than Pingo AI in absolute dollars, even though Pingo is currently growing faster.

This is the most important counterargument to giving Pingo the crown.

Speak was approaching roughly $50 million ARR around the end of 2024, when the company was still reporting growth above 100% year over year. Forbes later reported that Speak had crossed $100 million in annualized revenue.

So Speak appears to have added roughly $50 million of annualized revenue over about a year.

Pingo's entire current business is only around $12 million ARR. Speak therefore added several times Pingo's present revenue base during one recent growth period.

Speak also has a much broader commercial footprint. Forbes reported around 15 million cumulative downloads, consumer subscriptions generally priced from roughly $80 to $200, and about 500 companies using Speak for Business. The company first built a strong position in South Korea, then expanded in markets including Japan and Taiwan before pushing harder into the United States.

The comparison becomes straightforward once we separate the metrics.

Pingo is growing faster.

Speak is growing by more dollars.

For a question asking which startup has the highest current growth rate, we choose Pingo. If the question were “Which AI education startup is creating the most new revenue?”, Speak would probably win.

Company	Earlier revenue level	More recent level	What stands out
Pingo AI	About $300K annualized after two months	About $12M ARR	Roughly 40x from the early revenue milestone
Speak	Roughly $50M ARR	More than $100M annualized	About $50M of additional annualized revenue
SpeakX	About $3M ARR	About $7M ARR	More than doubled during 2025
MagicSchool	Started monetizing after viral teacher adoption	More than $10M ARR	Fast conversion of free K-12 adoption into paid contracts

Has Pingo AI become too big to dismiss as “small-base growth”?

Pingo AI's small starting base still explains part of its extraordinary percentage growth, but a roughly $12 million ARR business is now large enough for the growth to count seriously.

The denominator argument was much stronger when Pingo generated $25,000 a month. Going from $25,000 to $50,000 only requires another $25,000 of monthly revenue.

Going from several million dollars of ARR to $12 million is different.

Pingo's more recent company material says it has more than 8 million users. That puts its user footprint surprisingly close to Speak's reported 15 million downloads despite Speak having existed for many more years.

Revenue per user is clearly much lower at Pingo, and cumulative users cannot be treated as subscribers. Still, Pingo is now operating at a scale where another 50% increase would represent millions of dollars rather than a few thousand.

This is where our conclusion changed after checking the newest evidence.

At roughly $6 million ARR, we would have called Pingo the percentage-growth winner while still giving Speak the overall title. At roughly $12 million ARR, with the company still reporting rapid monthly expansion, that distinction has become harder to defend.

Pingo has moved far enough beyond the tiny-startup stage that its growth rate deserves to decide the title.

Is MagicSchool still the fastest-growing AI startup inside schools?

MagicSchool remains the strongest answer if we narrow the question to K-12 AI adoption.

Its early growth was extraordinary. MagicSchool launched in 2023 and reached around 100,000 teachers within its first few months. Founder Adeel Khan later said it reached one million users in roughly five months.

By 2024, the company was reporting more than two million educators and thousands of school and district partners. More recent company and investor disclosures put MagicSchool at around seven million users, more than $10 million ARR and districts representing roughly one in five U.S. children.

The scale of that jump is worth sitting with. MagicSchool effectively built a multi-million-user education platform in the time many school software startups need to complete their first full procurement cycle.

The revenue is equally important. According to a recent interview with Khan, MagicSchool reached more than $10 million ARR after beginning monetization, using a sales organization made up overwhelmingly of former educators.

That means the teacher adoption did eventually turn into a real business.

Pingo is growing faster as a consumer company right now, but MagicSchool's rise inside K-12 is still the most impressive institutional adoption story we found.

If you want more recent data on this point, please see our latest AI in education market report.

Is MagicSchool still exploding, or has its growth started to normalize?

MagicSchool is still expanding today, although the breathtaking multiplication seen immediately after launch has clearly slowed.

Going from zero to one million users in around five months cannot continue for long. Nor can the jump from roughly 100,000 early teachers to several million educators be repeated indefinitely once a product has already spread across much of the U.S. school system.

At roughly seven million users, MagicSchool's next stage looks different.

The company now has to turn widespread teacher familiarity into deeper paid district adoption. Reaching $10 million-plus ARR suggests that process is already well underway.

This is a normal shift for a freemium education company. Early growth is dominated by individual teachers trying the product and recommending it to colleagues. Later growth depends more on district agreements, renewals, approved workflows, administration controls and budget allocation.

So MagicSchool is no longer the cleanest candidate for the highest percentage growth rate.

Its position has arguably become stronger anyway. Seven million users and meaningful recurring revenue give MagicSchool something more valuable than another viral quarter: a large installed base inside an industry where distribution is notoriously difficult.

Is SchoolAI catching MagicSchool?

SchoolAI is closing the credibility gap with MagicSchool, but we still do not have enough commercial evidence to say it is growing faster.

SchoolAI has reached more than one million classrooms across more than 80 countries and has been embedded through more than 500 education partnerships, according to figures published by OpenAI. Those numbers were achieved in roughly two years after SchoolAI launched in 2023.

The more interesting fresh evidence is about repeated use.

SchoolAI now points to Stanford research tracking 9,081 teachers on the platform. According to SchoolAI's summary of that work, 42% became regular or power users, 60% returned week after week, and half used the product on five or more days.

Those figures help answer one of the biggest weaknesses in education-AI adoption data. Reaching a classroom once tells us little. Returning every week looks much more like a real habit.

SchoolAI has also been pushing further into district controls for the new school year, including curriculum configuration, centralized guardrails, analytics and teacher-managed student AI experiences.

The gap versus MagicSchool is transparency around revenue. MagicSchool has disclosed more than $10 million ARR. SchoolAI has raised substantial capital and reported impressive institutional reach, but we still lack a comparable revenue trajectory.

Until that changes, MagicSchool remains easier to rank ahead.

Are SpeakX and Synthesis growing fast enough to challenge the leaders?

SpeakX and Synthesis are both growing quickly, but neither currently has the combination of recent speed and scale that Pingo, Speak or MagicSchool can show.

SpeakX is the stronger challenger on revenue.

The India-focused AI English-learning app was producing roughly $3 million ARR in early 2025. Later that year, investor Elevation Capital said SpeakX had passed 200,000 paying subscribers and approximately $7 million ARR.

That is roughly 133% growth from the earlier ARR figure.

The efficiency is impressive too. SpeakX reached that scale with a team reported at roughly 20 people, giving the company unusually high revenue per employee for an education startup.

Synthesis is attacking a different market: AI-supported elementary math tutoring. The company previously said subscribed students had grown about 4.5 times year over year and that revenue was tracking above a $10 million annual pace.

Its current website says more than 35,000 families use Synthesis Tutor. That is up from older pages showing 25,000 families, so the customer footprint is still moving.

Synthesis also has a clearer willingness-to-pay story than many classroom AI tools. Families currently pay meaningful subscription prices rather than simply registering for a free teacher product.

Neither company should be ignored. SpeakX could become a serious language-learning competitor if it reproduces its Indian growth abroad, while Synthesis has one of the strongest paid tutoring models in the category.

Today, though, Pingo's curve is steeper and Speak's revenue base is much larger.

Are millions of AI education users actually worth anything?

Millions of AI education users only become convincing growth evidence when we can see that people return, pay or pull the product deeper into schools.

This is why the rankings look different once we move beyond headline user counts.

Pingo says more than 8 million people use its language-learning product, alongside approximately $12 million ARR. Speak has around 15 million downloads and more than $100 million in annualized revenue. Both companies therefore give us commercial data alongside their consumer reach.

MagicSchool also passes this test better than it did during its first viral year. Its roughly seven million users now sit beside more than $10 million ARR and large district penetration.

SchoolAI gives us less revenue transparency but better usage depth. Its recent teacher research suggests 60% of tracked users returned week after week, which is a much stronger data point than simply saying the platform has touched one million classrooms.

Synthesis takes another route entirely: it reports families using a paid tutoring product. Even though the absolute audience is much smaller, 35,000 families paying for math help can be commercially more valuable than hundreds of thousands of lightly engaged free accounts.

So we do use user numbers, but we rarely let them decide the ranking alone.

The stronger pattern across this market is that the leaders are beginning to show a second layer beneath adoption: Pingo and Speak have subscriptions, MagicSchool has district revenue, SchoolAI has repeat teacher behavior, and Synthesis has paying families.

If you want more recent data on this point, please see our latest AI in education market report.

Are new school AI restrictions going to slow MagicSchool and SchoolAI?

New restrictions on student-facing AI are becoming a real obstacle for K-12 startups, especially products that put generative AI directly in front of younger students.

The clearest recent example comes from New York City, the largest public-school system in the United States.

For the new school year, New York City Public Schools has imposed a one-year moratorium on student-facing generative AI from 2-K through eighth grade. The policy affects nearly 600,000 students. High-school AI use will be limited to vetted programs, supervised pilots and specific career-related contexts.

Teachers can still use approved AI for instructional planning and operational work.

That distinction matters enormously for the startups in this article.

MagicSchool initially spread through teacher tasks such as lesson planning, differentiation, assessments and communications. Those workflows fit much more comfortably inside the new rules.

SchoolAI also offers teacher tools, but a large part of its product vision involves students interacting directly with teacher-controlled AI “Spaces” and personalized tutors. Flint follows a similar approach.

New York is only one district, so we should not turn one policy into a national conclusion. But its scale makes the decision difficult for the rest of the market to ignore.

For K-12 AI companies today, teacher productivity has become the easier growth lane. Student-facing AI has to prove itself under much heavier scrutiny.

Could school AI restrictions actually help MagicSchool and SchoolAI?

Tighter school rules could eventually help the strongest education-specific AI platforms because districts now need controlled alternatives to unrestricted consumer chatbots.

The direction of New York City's policy shows what that future could look like.

High school students will still encounter AI, but through approved programs, limited pilots, teacher supervision and formal AI-literacy lessons. The city says its pilots were selected partly around safety, privacy, teacher-guided instruction and keeping students responsible for the thinking.

Those requirements look much closer to SchoolAI's product design than to giving every student an unrestricted ChatGPT account.

SchoolAI lets teachers create and monitor AI learning spaces, while districts can apply guardrails and centrally manage what gets used. MagicSchool has also spent years building education-specific controls around teacher workflows.

The emerging market may therefore reward products that can answer a district's compliance questions rather than whichever chatbot has the smartest general-purpose model.

There is a slightly counterintuitive possibility here: regulation slows raw student adoption in the short term while making approved institutional distribution more valuable. A school AI startup that survives procurement, privacy review and curriculum scrutiny may face fewer casual competitors on the other side.

MagicSchool and SchoolAI are both positioned for that world.

If you want more recent data on this point, please see our latest AI in education market report.

Can Pingo AI keep growing this fast once the viral videos stop working?

Pingo AI's biggest weakness today is that we still do not know how durable its extraordinary consumer growth will be.

Its acquisition machine has worked remarkably well. Pingo turned AI language conversations into entertaining short-form content, and that content appears to have generated huge awareness without requiring the advertising budgets normally associated with consumer apps.

That creates an obvious risk too.

Viral formats eventually tire. Competitors can imitate them. Social algorithms change. Users acquired because a funny AI clip appeared on TikTok may behave differently from people who deliberately searched for a serious language-learning program.

Pingo's own hiring priorities show that the team understands the next problem. Current roles focus heavily on memory, personalization, real-time speech, dialogue quality, latency and D1/D7/D30 retention. Those are exactly the systems a viral app needs once acquisition stops being the main constraint.

Speak has already travelled further down that road. It has spent years building language curricula, expanding internationally and getting users to speak repeatedly rather than simply test the AI once.

That gives Speak more evidence of durability.

Pingo therefore wins the growth-rate question today, but we have less confidence that Pingo will still be growing fastest a year from now than we have in Speak remaining a very large AI education business.

Which AI education startup has the hardest growth to fake?

Speak still has the strongest growth evidence overall because more than $100 million in annualized revenue is much harder to explain away than registrations or viral traffic.

This is where Speak remains ahead of Pingo despite losing the title on speed.

Forbes reported Speak's $100 million-plus annualized revenue alongside roughly 15 million downloads and around 500 enterprise customers. The company also has years of expansion across South Korea, Japan, Taiwan and other markets behind the figure.

Pingo's approximately $12 million ARR is impressive, and the progression across several company snapshots makes the trajectory increasingly credible. But the number remains self-reported, and we have much less public information about retention, subscriber cohorts and geographic mix.

MagicSchool has strong evidence from another direction. Its multi-million-user reach now sits beside more than $10 million ARR and deep U.S. district penetration.

SchoolAI is becoming more convincing because we finally have repeated-use data rather than just a giant classroom number.

Our confidence therefore differs by company.

We are very confident Speak has built the largest AI-native education business among the startups examined here. We are fairly confident Pingo currently has the fastest revenue trajectory. We are also confident MagicSchool leads specialist K-12 AI adoption. We have less confidence when ranking SchoolAI, SpeakX and Synthesis against those three because their public growth data are less complete.

Question	Best answer today	Confidence
Who is growing fastest by current rate?	Pingo AI	Fairly high
Who is adding the most revenue dollars?	Speak	High
Who leads AI-native K-12 adoption?	MagicSchool	High
Who has the strongest classroom challenger position?	SchoolAI	Medium
Who is the strongest smaller language-learning challenger?	SpeakX	Medium
Who has the clearest paid AI math-tutoring traction?	Synthesis	Medium

If you want more recent data on this point, please see our latest AI in education market report.

Which AI education startup is growing the fastest right now?

Pingo AI is currently the fastest-growing AI education startup we can identify, while Speak remains by far the larger business and MagicSchool remains the clearest K-12 leader.

The newest Pingo numbers are what push us over the line.

Pingo launched in January 2025, reached roughly $25,000 MRR and 50,000 users within about two months, moved to around $200,000 MRR by July, and later crossed several million dollars of ARR. Current company hiring material says the business has now reached approximately $12 million ARR, more than 8 million users and roughly 50% month-over-month growth.

That trajectory is faster than the other credible contenders we examined.

Speak still wins on absolute scale. Its annualized revenue has crossed $100 million, and it appears to have added around $50 million of annualized business during one recent year. If we were asking which AI education startup has created the most new revenue, we would pick Speak without much hesitation.

MagicSchool owns another version of the title. Going from launch to roughly seven million users, more than $10 million ARR and district penetration reaching about one in five U.S. children makes it the standout generative-AI startup inside K-12 education.

But the title asks who is growing fastest.

As of now, Pingo AI is the sharpest answer. A roughly 40x increase from its early annualized revenue level, followed by continued expansion into an eight-million-user, eight-figure-ARR business, is too large to write off as a small-denominator trick.

The more interesting question from here is whether Pingo can hold that pace once the viral acquisition engine matures. Speak has already shown that an AI language tutor can become a $100 million-plus business. Pingo is currently trying to cover the distance much faster.

Who is the author of this content?

NEW MARKET PITCH TEAM

We track new markets so founders and investors can move faster

We build living "market pitch" documents for emerging markets: AI, synthetic biology, new proteins, and more. Instead of outdated PDFs or hallucinated LLM answers, our clients get a clean, visual, always-updated view of what's really happening: key players, deals, regulations, and signals that matter. Learn more about us.