Sudip Rokaya knew the comparison would raise eyebrows. When he posted it to LinkedIn in late April, the numbers told a lopsided story: his startup's tool had just generated a complete whiteboard explainer video in 11 seconds flat. Google's NotebookLM, handed the same prompt, took more than 16 minutes according to Rokaya's comparison—though that figure, like the 11-second claim, comes from the founder's own testing rather than independent verification.
For a company that barely exists yet—two founders, a waitlist, and a spot in Y Combinator's current batch—it was an audacious shot across the bow.
Lamina Labs, the San Francisco outfit behind that claim, is now pulling back the curtain on Simi, a product it describes as "the visualization layer for AI-native EdTech." The premise is deceptively simple: type in a request, wait a handful of seconds, and out comes an animated educational video in the style of those whiteboard explainers that have colonized YouTube tutorials and online courses. The target market? EdTech platforms that need to crank out visual explanations at industrial scale, where minutes of wait time per video start to hurt.
Whether that 90-times speed advantage holds up under scrutiny from anyone other than the founders themselves remains an open question. So does the broader bet: that raw speed is what EdTech buyers actually care most about.
A Waitlist, a Claim, and No Customers (Yet)
Lamina's pitch doesn't meander. The website (laminalabs.ai) lays out a streamlined process tailored for contexts where pacing and clarity trump cinematic flair. According to the company's Y Combinator directory page, the infrastructure produces "accurate visual explanations" from text prompts, with an emphasis on what Lamina calls "deterministic animation." That phrasing, which Rokaya has used on LinkedIn, hints at frame-level control designed for educational precision rather than the looser, sometimes unpredictable output of generative models.
What Lamina hasn't disclosed: pricing, a commercial model, or any public customers. As of mid-May 2026, the company is onboarding users exclusively through a waitlist, as seen on laminalabs.ai. No logos grace its homepage. No case studies. No partnership press releases.
The 11-second figure, for its part, comes from Rokaya's self-reported LinkedIn post, not from independent verification. Third-party benchmarks haven't surfaced. Google's NotebookLM did roll out "Cinematic Video Overviews" to paid users in March, and the company's own help documentation concedes generation "can take a while." But absent external tests, the comparison remains founder-reported—a data point that aligns with Lamina's narrative, if not yet with anyone else's.
Still, if the claim is accurate, it's the kind of infrastructure edge that can matter in markets where volume dictates viability.
Two Founders, One Bet
Lamina was founded in 2025 by Rokaya and Kartikesh Mishra, both of whom trace their roots to MIT. Rokaya left the university's computer science and math program to launch the company, announcing the leap in an April LinkedIn post with a headline that didn't bury the lede: "Leaving MIT to build the next generation infrastructure for animated explainer videos." Mishra, who holds a BS and MEng in electrical engineering and computer science from MIT (class of 2024 and 2025, respectively), rounds out the duo.
Y Combinator's directory lists Lamina's team size as two, though the company's LinkedIn profile suggests a range of 2–10 employees. That fuzziness is common for early-stage startups still figuring out contractor versus full-time headcount. Lamina is part of YC's Spring 2026 cohort, which runs April through June, with Demo Day slated for June 16.
Third-party funding trackers like CB Insights and Dealroom estimate the company's total raise at around $500,000, consistent with YC's standard deal: $125,000 for 7% equity, plus $375,000 on an uncapped most-favored-nation SAFE. Lamina hasn't confirmed those figures publicly, and for a pre-Demo Day company, that silence isn't unusual.
A Crowded Field, With Room for Speed
Lamina is hardly alone in chasing AI-generated explainer videos. Google's NotebookLM looms as the big-platform contender, while a thicket of startups crowd the same corner. Golpo AI, another YC alum according to Y Combinator's directory, bills itself as an "AI whiteboard explainer video generator." Tools like Sketchly and Skribcast offer similar capabilities, though generation times often clock in at minutes rather than seconds. Legacy platforms—VideoScribe, Vyond—still command market share, but they lean on manual workflows or template libraries that don't mesh well with the "one prompt, instant video" ethos.
Then there are the code-first approaches. Newer tools that convert text into Manim code (like Animo, updated as recently as May 11) target STEM explainers but still require separate rendering steps, adding latency that Lamina claims to bypass.
The academic side of the conversation has gotten noisier, too. A Nature Scientific Reports study published May 13 found students preferred personalized AI-generated videos over non-personalized human-recorded ones—a finding that suggests the tech has cleared some baseline threshold of pedagogical credibility. Separate reviews in Frontiers in Computer Science and SAGE journals have cataloged the taxonomy and limitations of AI video in higher ed and medical training. And a February paper from Stanford's SCALE initiative critiqued end-to-end video models for instruction, arguing that pedagogy-aware systems—not just fast generation—should be the goal.
That framing might be where Lamina's "deterministic animation" language becomes more than marketing speak. If the company can demonstrate not just speed but also control that aligns with learning outcomes, the wedge gets sharper.
What Happens Next
For now, Lamina's website offers a waitlist and a scheduling link to "Talk to Us." What it doesn't offer: the names of EdTech platforms that have signed on, pilot results, or even a pricing model. The timing of Simi's launch—weeks before YC Demo Day—looks like a momentum play, a way to gather data and interest before the main event on June 16.
The big unanswered question isn't technical. It's market fit. Do EdTech buyers prioritize speed over pedagogy? Over content quality? Over integration depth with existing learning management systems? Lamina's bet is that latency is a real enough pain point that a 90x improvement opens doors. Perhaps it does. Or perhaps, as the Stanford researchers might suggest, speed is table stakes and the harder problem is building systems that actually teach well.
For a two-person team with a waitlist and a bold LinkedIn post, that's the test ahead. The infrastructure claim is on the table. The market will decide whether it's the right wedge—and whether 11 seconds is fast enough to matter.
