Level Up Awards — Pitch Analysis

Pitch-by-pitch critique of the five Level Up Awards finalists, evaluated against the actual constraint they were given: a strict 60-second pitch, no slides allowed to run over, followed by unscripted jury Q&A. This is analysis and opinion, not transcript — see `06-awards-pitch-transcript.md` for exactly what was said.

1. Mo — Atamute (acoustic noise reduction)

What he did well: - Led with a concrete, felt problem (HVAC/data-center noise) that the jury immediately related to personally — the juror's first reaction was "this resonated because my neighbors complain about my HVAC," which is the best possible opening reaction a pitch can get. - Packed in real proof points fast: two mass-production customers, 15+ proof-of-concept projects, 100+ NDAs including large Asian and European companies, six patent applications (one granted), eight-person team including three professors. For 60 seconds, that's an unusually dense, credible traction summary. - In the Q&A, he answered the "how does a small player attack a big diversified market" question with real technical specificity (compressor jacket vs. fan silencer, exact thickness numbers — 1mm vs. 30cm foam) rather than a vague strategic answer. Specific numbers under pressure read as competence.

What could have been better: - He ran over the 60 seconds and was visibly cut off mid-sentence ("we have pilot production and now we are developing" — cut). For a competition that explicitly told finalists they'd be cut off, losing the closing line is a real cost — it means the pitch had no deliberate "landing" line, which is often the most memorable part of a pitch. - The pitch was almost entirely a list of proof points (customers, patents, team) with very little on the why now or the emotional hook — it reads more like a due-diligence data sheet than a story. Given he had the jury's attention from personal relevance in the very first Q&A answer, opening the pitch itself with that same personal/relatable framing (rather than saving it for the Q&A) would likely have landed harder in the 60-second window. - No explicit statement of market size or business model was given in the pitch itself — the jury had to ask about competitive landscape rather than hearing it upfront, meaning a full minute was spent without ever establishing "how big is this."

Jury Q&A quality: Strong. Both answers were specific, technical, and directly responsive — no deflection, no filler.

2. Jeroen — AIO (nanophotonic image sensors)

What he did well: - Best opening line of the five: "cameras that are 70% blind" is a genuinely arresting, counter-intuitive claim that immediately creates curiosity — a textbook example of leading with tension rather than a feature list. - Efficient use of the trillion-dollar-market list (physical AI, smartphones, smart cities, AR/VR, medical imaging, autonomous driving) — names the scale without dwelling on it, then moves straight back to the specific mechanism (light filtering to see color) that unifies all of them. That's a harder pitch craft skill than it looks: naming a big market without getting stuck justifying it. - In Q&A, the smartphone-camera-size explanation ("why the bump on your phone got bigger every year, and why this technology reverses that") was an excellent everyday analogy for a genuinely technical, hard-to-picture claim — this is the single best "make deep tech tangible" moment across all five pitches. - Directly and confidently answered the hardest question of his round (manufacturing dependency) with a specific, verifiable claim: technology already transferred into a commercial foundry, first product already running through the line. That's the right way to defuse a "you're dependent on others" objection — show it's already been operationalized, not just planned.

What could have been better: - This was, by his own account, effectively a repeat of the same pitch and company he'd already presented earlier in the day at the investor/deep-tech fireside session — for anyone in the room who'd seen both, there was very little new information here. In a competition explicitly about who pitches best live, repeating almost verbatim material is a missed opportunity to sharpen or specialize the pitch for this specific audience (a jury judging pitch craft, not an investor room judging investment fit). - The business-model answer ("we sell the chip, not the technology") was clear but slightly defensive in tone ("no, so we deliberately..." ) — phrasing it as a proactive strength from the start of the answer rather than a correction would have read more confidently.

Jury Q&A quality: Excellent — every answer was concrete, verifiable, and built the case rather than merely defending it.

3. Merrick — programmable haptic material (Heptal Tech) — the eventual winner

What he did well: - The single best 60 seconds of the five, and the most memorable pitch mechanic of the day: opening with "what if I told you, without technology, you could feel what I wanted to show you?" and then handing the jury a live, physical demo instead of slides. This directly converts the pitch's core claim (a new tactile interface) into something the jury experiences rather than just hears about — the strongest possible way to prove a genuinely sensory technology. - Explicitly named and neutralized his own apparent weakness up front ("some contestants might find it a pain today that we have no slides") before turning it into the pitch's strength — a good example of pre-empting an objection rather than waiting for the jury to raise it. - In Q&A, cited real academic perception-study grounding (University of Eindhoven spinout, formal studies with professors on what people can and can't feel) when asked about user research — this is exactly the kind of evidence a jury wants when the underlying claim (comfort/desirability of touch) is inherently subjective and hard to fake. - Let a juror physically try the device live during the Q&A ("can I try it?" / "of course you can") — turning Q&A time into more demo time rather than treating it as a defensive round. This is a distinctive, confident move few founders would risk with a limited prototype.

What could have been better: - When asked to pick a single focus market out of a wide range of potential applications (haptic screens, wearable braille coatings, consumer electronics), the answer named consumer electronics but didn't fully commit — it read as "we're talking to them" rather than "this is our beachhead and here's why." Given the jury's own closing feedback (see below) explicitly flagged this same breadth-vs-focus concern as the team's main open risk, tightening this answer in the room would have pre-empted the jury's single biggest reservation about the company. - The pitch itself (before Q&A) leaned entirely on the demo and the vision, with almost no traction, team, or business detail (no customer names, no revenue/funding figures) — a deliberate choice given the “no slides” framing, but it meant the win rested almost entirely on charisma and novelty rather than proof, which is a riskier bet than Mo's or Jeroen's proof-heavy approach, even though it clearly paid off here.

Jury Q&A quality: Very strong, and the live-demo moment elevated the whole round — but the market-focus answer was the one soft spot, later echoed directly in the jury's own award rationale.

4. Paul — Cairo / Chiron (light-based PCB soldering)

What he did well: - Strong, punchy framing with a memorable, quotable line ("chips don't float, they need to be interconnected") that turns an unglamorous manufacturing detail into a vivid, unavoidable-sounding problem. - Good use of specificity to make an old, boring-sounding process (stencil printing) sound urgent: naming that the underlying method is 50 years old and was borrowed from the textile industry is a sharp, slightly absurd detail that makes the "nobody has touched this in decades" claim land. - The yield statistic in Q&A (30% yield with stencil printing on an advanced board, versus 80-90% with their machine) is an excellent answer — a single, concrete, dramatic number that does more work than a paragraph of explanation would. - Answered every question (confidence in market timing, addressable niche and its size, team, business model, market size) with a real number attached, every time — no answer was purely qualitative.

What could have been better: - The pitch script itself was noticeably rougher in delivery than the others — several sentence fragments read as if he was translating or recalling a memorized script rather than speaking naturally ("all investments were gone to semi-drawn," "hardly since 50 years there were no fundamental changes"). This likely cost some clarity in the room even if the underlying content was strong; a founder who has one shot at 60 seconds benefits disproportionately from over-rehearsing the opening lines specifically, since that's where jury impressions are formed fastest. - The confidence question ("why are people ready to adopt this now, after decades of no innovation?") was answered with proof of demand (customers placing orders) rather than actually explaining why now — i.e., what changed that makes today different from the last 50 years of stasis. It's a subtly different question than the one he answered, and the actual "why now" (AI chip interconnect density crossing a threshold current methods can't handle) was implied elsewhere in the pitch but never explicitly tied back to this specific question.

Jury Q&A quality: Strong on substance (every answer had a real, specific number), weaker on precisely answering the question as asked in one case.

5. Marwitz Kapteijn — Keebu (AI agent security)

What he did well: - Timely, high-relevance problem framing — explicitly ties the pitch to something the audience has almost certainly already encountered (AI agents taking real actions) and sharpens the stakes precisely ("it's more than just a hallucination — now they can actually delete a database"). - Handled the toughest, most adversarial question of the entire round — "how do you survive being crushed by fully-funded Silicon Valley competitors" — without flinching, and importantly self-corrected mid-answer: opened with a throwaway joke ("get more funding and grow aggressively") but then explicitly flagged it and gave a substantive answer (positioning as an external, independent governance/audit layer rather than competing with model vendors directly, plus a genuine European-regulation tailwind argument). Recognizing a weak first answer and visibly upgrading it in real time is a good recovery skill. - The follow-up on where the product's responsibility ends and institutional governance begins was the most conceptually sophisticated exchange of the whole session, and he handled it well: clearly separated "deciding the rule" (an institutional/governance question) from "enforcing the rule" (his product's job), which is exactly the right way to answer a question designed to probe whether he understood the limits of his own product.

What could have been better: - The pitch itself was the shortest and least specific of the five on quantifiable proof — no customer names, no revenue, no team size were mentioned in the 60 seconds (contrast with Mo and Paul, who front-loaded hard numbers). For a security/governance product where trust is the entire sale, the absence of any concrete deployment or pilot evidence was a noticeable gap the jury didn't happen to probe, but easily could have. - The company-name question ("what does Keebu stand for?") was answered with a joke about domain-name availability. Charming and honest, but a missed opportunity — even a slightly better-prepared, half-serious answer here would have used a "soft" question to reinforce the brand or mission rather than spend it on a shrug.

Jury Q&A quality: The best single answer of the day (the governance/enforcement distinction), let down slightly by an otherwise proof-light pitch that left easy follow-up questions on the table unasked.

Cross-cutting patterns across all five pitches

• The proof-heavy pitches (Mo, Paul) and the vision/demo-heavy pitch (Merrick) represent two genuinely different, both-viable strategies for a 60-second format — but only the demo-led approach won, which is a useful data point: in a room-based, jury-judged pitch competition (as opposed to a written investor memo), a felt experience appears to beat a denser list of facts, at least when the underlying technology is inherently demonstrable.
• Every founder who was asked "why now" or "why hasn't this been solved already" answered with evidence of current demand rather than a direct explanation of what changed — this was the single most common soft spot across the round. A sharper version of nearly every pitch would pre-load one sentence explicitly naming the shift (new regulation, a technology threshold crossed, a new class of customer demand) that makes now, specifically, the right moment.
• The strongest single technique across the whole session was translating a technical claim into something already familiar to the audience — Jeroen's smartphone-camera-bump story and Paul's yield percentage both did this well; the pitches that stayed at the level of specifications and jargon (parts of Mo's and Kapteijn's pitches) were harder to hold onto immediately after hearing them once.
• The jury's questioning style throughout was constructive, not adversarial (with the partial exception of the Silicon Valley competition question to Kapteijn) — most questions were genuinely curious "tell me more" prompts rather than attempts to find a hole, which likely reflects the event's framing as a showcase/award rather than a funding gate.