What an AI funding round actually means: Lovable's $6.6B and the edtech collapse

A funding round is a sale of new shares. A startup hands investors a slice of the company in exchange for cash, and the price of that slice sets a valuation. That is the whole mechanism. Everything else, the headlines, the leaderboards, the "Europe's most valuable startup" labels, is downstream of a private negotiation that neither side has to disclose. That is why the sourcing matters more than the number.

CNBC reported on 16 December that Lovable's latest round values the company at $6.6 billion. The figure comes from two people with knowledge of the deal who asked to remain anonymous, because the round is private information. Both said U.S. venture firm Accel participated, and one said Khosla Ventures is also in. CNBC approached all three companies for comment and none had responded when the article went live. Forbes had reported in November that the round would value Lovable at "around" $6 billion, so the number moved before it was confirmed.

The comparison CNBC draws is with Lovable's July round, which valued it at $1.8 billion. That is a multiple of more than three in roughly five months, and it would be the company's third round of 2025. Founded in 2023, Lovable reported $200 million in annual recurring revenue in November, just under a year after it first hit $1 million in ARR. The July raise was $200 million, from Accel plus Creandum, Klarna founder Sebastian Siemiatkowski, ElevenLabs founder Mati Staniszewski and Synthesia founder Victor Riparbelli. Lovable is based in Stockholm and is opening offices in Boston and San Francisco.

Why vibe coding attracts the money

Lovable's product takes text prompts and produces apps and websites, using models from providers including OpenAI and Anthropic. The company said 100,000 projects were being built on its platform every day when it announced the November ARR figures. Its investors are betting that a large population of people who cannot code will pay to ship software anyway.

The category is not a Lovable story alone. Anysphere, which makes the coding tool Cursor, raised $2.3 billion at a $29.3 billion valuation in November, according to CNBC. Replit hit a $3 billion price tag in September after picking up $250 million, and Vercel closed a $300 million round at a $9.3 billion valuation. Accel, the source says, is in the Lovable round and has also backed Cursor and Mira Murati's Thinking Machines. That tells you more about Accel's thesis than about any single product.

A valuation is not cash in the bank. It is not revenue, and it is not a promise that the next round clears the same bar. It is the price of one transaction, agreed by buyers and sellers with an interest in the number being high. The same caution applies to every figure in this article: they are reported valuations, from named outlets and anonymous sources, not audited accounts.

The other side of the same market

Money that goes into one sector comes out of another, or simply does not go in. Rest of World reported in April 2026 that global edtech investment peaked at $16.7 billion in 2021 and fell to less than $3 billion by 2025, according to Tracxn, a Bengaluru platform that tracks startup funding. The same piece notes that most global edtech funding comes from U.S. venture firms.

The founder pool shrank alongside the cheques. Rest of World cites 645 companies launched in 2025, against almost 10,500 in 2020. The failures are instructive rather than random: Byju's, once the world's most valuable education startup at $22 billion, collapsed under a financial crisis and aggressive sales tactics for expensive courses, while Nigerian startup Edukoya shut down in 2025 on weak profitability. Loot Drop, a database of more than 1,700 startup closures cited by Rest of World, points to high customer acquisition costs, long institutional sales cycles and low retention as structural problems in the sector. In China, the government's "double reduction" policy in July 2021 ended the K-12 online tutoring business more or less overnight: Yuanfudao, once valued at $15.5 billion, pivoted to AI learning hardware and now competes with Zuoyebang in that market.

HolonIQ, a research firm that advises governments and investors on education, wrote in a February 6 post cited by Rest of World that "venture capital flows reflected a shift from volume to intention" and that investors "concentrated capital in AI-enabled products, workforce-aligned platforms, and K-12 operations solutions that address cost or operational pressures, staffing challenges, and learning support at scale." That is a polite way of saying the money followed cost-cutting software, not classrooms.

Where the pitch goes wrong

Valuations can also be undone by things that have nothing to do with the cap table. The Register reported on 2 July 2026 that MeetingTV, a video conferencing and webinar startup, sued Palo Alto Networks, its Koi Security unit and Koi researchers over a December 30 blog post that linked MeetingTV to a Chinese corporate espionage operation. The complaint alleges Koi used an LLM to generate the threat report, that the model hallucinated findings, and that Koi published them as fact. It describes "reckless publication of an AI-driven cybersecurity report" and says the false attributions came from Koi's "Wings" analytical platform, which the suit says generated erroneous correlations with a group Koi called DarkSpectre.

A Palo Alto Networks spokesperson told The Register the company "is aware of the lawsuit brought by MeetingTV Inc. regarding a threat research report published by Koi Security prior to the acquisition" and declined to answer specific questions. The same spokesperson said: "We believe Koi's cybersecurity research reflects its commitment to identifying and exposing threats to users and enterprises, and we expect that this dispute will be resolved through the appropriate legal process."

MeetingTV founder and CEO Michael Robertson told The Register that Koi did not contact the company before publishing, and that he learned of the report when security vendors began blocking his domains. "If people on the internet are blocked from reaching your company, then that's a death sentence," he said, adding that providers including Verizon and Palo Alto Networks were still blocking the startup. The Register notes the blog was silently edited to remove references to Zoomcorder, MeetingTV's meeting recording service.

"Will these be made without human oversight? Will people have due process - see the accusations against them, present their own evidence, have a neutral arbiter? None of that happened in our case," Robertson told The Register. "They just declared us criminals and published it to the world."

The lesson is not that AI reports are worthless. It is that a claim published without the accused getting a look at it can cost a company its domains, its customers and its reputation faster than any funding round can repair them. Investors price that risk, or they should.

What to read in the next round

• Who is confirming the number: anonymous sources close to the deal, a term sheet, or a signed filing. The three are not equivalent.
• Whether the round is primary money into the company or secondary sales of existing shares. Only the first funds growth.
• Revenue quality: ARR is a run rate, not collected cash, and it can include discounts and annual contracts billed later.
• What comparable rounds did in the same month, because a valuation is a relative claim as much as an absolute one.

Lovable's round, if the sources are right, is a real event with real money behind it. It is also a data point in a market that has spent 2025 and 2026 concentrating capital into a narrow band of AI products while adjacent sectors emptied out. Both things are true at once, and the second one is the part founders tend to read last.