"We sold 25,000 devices and generated $4 million in revenue — with zero dollars raised. Let me show you what we're building and why now is the inflection point."
"GPT was trained on the internet — text, images, web pages. But the next wave of AI needs to understand the physical world — health, environment, movement. That requires real-time sensor data from real devices in real places. Nobody is building this infrastructure."
2:00 — 4:00 · 解决方案 + 产品
硬件 + PINNs 验证
不要深入技术细节。一句话说清楚 PINNs，然后转到 traction。
"We build consumer devices — phones, wearables, glasses — with a physics-based AI verification layer called PINNs. It cross-validates sensor data against physical laws, so the data is provably real. Think of it as a lie detector for sensor data — except it uses the laws of physics instead of a polygraph."
4:00 — 7:00 · Traction (最重要的 3 分钟)
让数字说话
这是 VC 做 mental note 的时刻。每说一个数字，停顿 1 秒让它沉淀。
"$4 million gross revenue. [pause] 25,000 devices deployed across 40+ countries. [pause] 60% gross margin. [pause] near-zero customer acquisition cost — $0 paid ads. [pause] 100,000+ community members. [pause] And we did all of this with zero funding. [longer pause] The unit economics work. Now we need to scale."
7:00 — 9:00 · 商业模式
今天赚钱的 + 明天赚更多的
不要让 VC 觉得硬件是全部。展示进化路径。
"Revenue today is 100% hardware — and that's by design. You can't sell data from a network that doesn't exist. We built the network first. Phase 2, launching this year: data activation. Users opt-in to share verified, anonymized data. We take a platform fee. Phase 3: enterprise data licensing to pharma and insurance. The hardware is the profitable engine that funds network growth. The data is the exponential upside."
9:00 — 10:30 · 竞争 + 护城河
一句话定位 + World Labs 上下游
不要逐个比较竞争对手。给一个框架，然后用 World Labs 对比锚定你的价值。
"Helium proved DePIN distribution works but failed on economics — token before revenue. World proved biometric data has massive value but is centralized and privacy-invasive. We combine DePIN distribution with physics-verified data quality. And we're the only one that's already profitable."
"Let me give you a specific comparison. World Labs — Fei-Fei Li's company — raised $1.23 billion to generate synthetic 3D worlds. But synthetic models need real-world ground truth data to validate and calibrate. That's exactly what we provide. They're the downstream consumer. We're the upstream data supplier. They have $1.23B and zero deployed devices. We have $4M in revenue and 25,000 sensors in the real world. We're not competitors — we're supply chain partners."
World Labs 对比速查 (背下来)
World Labs
Oysterworld
角色
下游消费者
上游数据供应商
融资
$1.23B
$0 (bootstrapped)
设备
0
25,000
数据源
合成生成
真实世界 ground truth
收入
Pre-revenue
$4M
验证
无
PINNs 物理验证
核心需求
需要真实数据
我们提供这个
杀手句: "Not competitors. Supply chain partners. They spent $1.23 billion building the consumer of data. We spent $0 building the supplier — and we're already profitable."
"We're raising $1 million on a SAFE at $30 million post-money. The capital goes to three things: scale distribution to 50K nodes, launch data monetization, and hire key technical talent for the data layer. This sets up a Series A at $50M+ in 12-18 months."
11:30 — 12:00 · 过渡到 Q&A
把控制权递出去（但有方向）
不要说 "any questions?"。引导他们问你准备好的领域。
"I know that's a lot — I'd love to go deeper on any area. Most investors have questions about the PINNs verification technology, the data monetization roadmap, or how we built this without funding. Where would you like to start?"
不要让会议在 "we'll get back to you" 结束。要一个具体的 next step。
"Before we wrap — I want to leave you with one number. $4 million, zero funding. That's the floor. The question is whether you want to be part of the ceiling. Can I send you access to our data room, and would it make sense to schedule a follow-up with [partner/team] next week?"
A — Acknowledge (承认): "That's a valid concern." 永远不要说 "不对" 或 "你理解错了"。
R — Reframe (重构): 把他的 concern 翻转成你的优势。
C — Concrete (具体): 用一个具体数字或事实钉死。
"$30M valuation is too high for a hardware company."
ARC Framework
A: "You're right — if we were just a hardware company, 7.7x on $4M revenue would be aggressive." R: "But you're not pricing hardware. You're pricing a 25,000-node sensor network that's already collecting data, with data monetization launching this year. That's infrastructure, not hardware." C: "Comparable DePIN networks — DIMO, Hivemapper — trade at 20-50x. World Labs raised at $5B+ with zero revenue and zero deployed devices. At $30M, we're the cheapest physical AI infrastructure deal on the market."
Why this works: Shifts the valuation anchor from hardware multiples to network/infrastructure multiples. The comps make $30M look like a bargain.
"Solo founder is a red flag."
ARC Framework
A: "I hear that from every investor and I respect it." R: "But I'm not really solo. I built a 32-agent AI development factory across 13 compute nodes. It runs 24/7 and shipped 5 products. That's not a metaphor — that's my engineering team. It cost me $100K/year. A human team doing the same output would cost $2M+." C: "That said — I'm actively hiring key technical talent for the data layer. One of the uses of this $1M is making that hire. So your capital directly de-risks this concern."
Why this works: Turns the weakness into a story about resourcefulness, then shows self-awareness about the gap, and ties the solution to the investment (making their money solve the problem).
"Revenue is 100% hardware. There's no data revenue yet."
ARC Framework
A: "Correct. Revenue today is entirely hardware." R: "And that was deliberate. You can't sell data from a network that doesn't exist. Every DePIN that launched tokens before revenue — Helium, IoTeX — crashed. We went the other way: revenue first, network second, data third. The hardware is the distribution engine that builds the network." C: "Even if data monetization takes longer than expected, the floor is a hardware company doing $4M at 60% margins. The downside is a profitable business. The upside is a data network."
Why this works: Reframes "no data revenue" as strategic discipline, uses Helium as a cautionary tale, and shows the downside protection.
"The Stripe numbers show $1M in reversals. That's suspicious."
ARC Framework
A: "I'm glad you caught that — I'd rather address it head-on." R: "The $1M in reversals was from a single distributor deal in 2024 where contract terms were restructured mid-fulfillment. Net on that deal was $750K. It's not chargebacks, not fraud, not customer dissatisfaction." C: "I can pull up the Stripe dashboard right now and show you every transaction. All other Stripe transactions are clean. Transparency is the policy."
Why this works: Being proactive about the messy number builds trust. Offering to show live data kills suspicion. Never be defensive about revenue details.
"Consumer phone sensors are too low quality. No pharma company will buy this data."
ARC Framework
A: "You're right — no single phone sensor matches medical-grade equipment." R: "But PINNs doesn't rely on a single sensor. It cross-validates 5+ sensor channels against physical laws. Like GPS — no single satellite is accurate, but four satellites triangulated together give you meter-level precision. Five noisy sensors cross-validated through physics produce a high-confidence result." C: "Our MAE is 0.35 against a clinical threshold of 0.61 — 1.74x more precise than required. And we benchmarked on MIMIC-III from MIT, the gold standard in biomedical AI."
Why this works: The GPS analogy is instantly intuitive. The MAE number is concrete and defensible. Citing MIMIC-III adds credibility.
"Why not just raise more? $1M seems small."
ARC Framework
A: "I could raise more." R: "But I've built a $4M business on $0. My cost structure is fundamentally different — AI factory at $100K/year vs. a $2M human team. $1M at this burn rate gives me 18+ months of runway. I'd rather give you a lower entry point and prove the data thesis, then raise a Series A at a much higher valuation." C: "At $30M post-money, your $1M buys 3.3%. If we hit $50M Series A — which requires 50K nodes and $1M ARR from data — that's a 1.67x in 12-18 months on a company with real revenue and real margins."
Why this works: Shows capital efficiency (a VC's favorite word), frames the small round as a feature not a bug, and gives them a clear return path.
"I don't understand the crypto angle. Is this a crypto company?"
ARC Framework
A: "Fair question — the crypto component confuses some investors." R: "We're not a crypto company. We're a hardware + AI company that uses blockchain for one specific thing: tamper-proof data attestation. 46% of revenue happens to come through crypto payments because our customer base in Southeast Asia prefers it. That's distribution optimization, not ideology." C: "We accept Stripe and crypto. The product works regardless of payment method. If TON disappeared tomorrow, the hardware and PINNs verification still work."
Why this works: Separates the technology choice (attestation) from the payment method (convenience). Shows you're pragmatic, not religious about crypto.
3. 不同类型 VC — 怎么讲
同一个 pitch，对不同 VC 要调不同频道。5 分钟 research 一下 VC 的 portfolio，就知道该强调什么。
在意: Unit economics, supply chain, defensibility, sensor technology
不要说: "DePIN" "token" "crypto" — 这些词让他们紧张
Lead with: "60% gross margins on consumer hardware, near-zero CAC — $0 ad spend, with a physics-based data verification layer that creates recurring revenue."
AI-Focused VC (Conviction, Air Street, Radical Ventures)
在意: PINNs innovation, data quality, AI agent ecosystem, model architecture
不要说: "We sell phones" — 听起来太 commoditized
Lead with: "We're building the perception layer for Physical AI. PINNs gives AI agents verified eyes in the physical world. We have 25K sensors already deployed."
Generalist / Seed Stage (Y Combinator, First Round, Initialized)
在意: Founder-market fit, traction velocity, market size, team velocity
不要说: 太多技术细节 — 他们在投你，不是投 PINNs
Lead with: "Solo founder. $4M revenue. Zero funding. 25K customers. 5 products shipped. I built an AI factory that replaces a 20-person engineering team."
Asia-Based (HashKey, Fenbushi, Spartan, DFG)
在意: SEA/LATAM distribution, Telegram/TON ecosystem, community size
不要说: "We're focused on the US market" — 你的优势就在亚洲
Lead with: "140K+ community (100K Telegram, 44K X), 25K devices, 70% of sales from SEA/LATAM through Telegram channels. We're already where the next billion users are."
4. 买/不买信号 — 实时判断
会议中不用猜 VC 感兴趣还是不感兴趣。这些信号会告诉你。
买入信号 (加码)
🟢 "Can you walk me through the financials in more detail?"
🟢 "Who else is in this round?"
🟢 "What's the timeline on the SAFE?"
🟢 "How much is allocated?"
🟢 "Can I introduce you to our operating partner?"
🟢 "Let me bring in [another partner] for a second call"
🟢 问关于 cap table、董事会席位、governance
🟢 问关于具体用钱计划和 milestone
🟢 会议超时但他们不着急结束
无兴趣信号 (不要追)
🔴 "Interesting, keep us posted."
🔴 "We don't typically invest in this stage."
🔴 频繁看手表或手机
🔴 没有任何 follow-up 问题
🔴 "Have you talked to [competitor]?"
🔴 "We need to see more data revenue traction."
🔴 会议准时结束，没有延长
🔴 "Send me your deck and I'll share internally."
🔴 不愿意 introduce 给其他 partner
收到买入信号时怎么做
加速: "We're targeting to close this round by [date]. I have 2-3 other conversations at similar stage. I'd love to move quickly with you if there's mutual interest."
创造紧迫感（诚实地）: 不要虚构 competing term sheet。但可以说 "我们有几个 active conversations" — 如果确实有的话。
收到无兴趣信号时怎么做
不要恋战。 问一个问题然后优雅离开：
"I appreciate your time. Before we wrap — what would need to be true for this to be interesting to you? I'd love to know what threshold you'd want to see."
这句话的价值：(1) 获得真实 feedback, (2) 可能打开 "come back when you have X" 的门, (3) 展示成熟度。
会议结束时没有 next step。 "We'll be in touch" = 永远不会 touch。要一个具体的日期。
焦虑地填充沉默。 说完大数字后停 2 秒。如果 VC 在想，不要打断他。沉默 = 你的数字在沉淀。
6. 会后跟进 — 72 小时决定成败
24 小时内：感谢邮件 + Data Room
Subject: Following up — Oysterworld ($4M revenue, raising $1M SAFE)
Hi [Name],
Thank you for the conversation today. I appreciated your questions about [specific topic they asked about — shows you listened].
As promised, here's access to our data room: [dataroom link]
Password: [password]
Key documents inside:
- Gross revenue report ($4M across Stripe, TON, MixPay)
- 3-year financial model
- Product overview and unit economics
I'd love to continue the conversation. Would [specific day] work for a follow-up call with [specific goal: "deep dive on data monetization" / "intro to your partner"]?
Best,
Howard
如果没回复 — Day 3 Nudge
Subject: Re: Following up — Oysterworld
Hi [Name],
Wanted to share a quick update — since we spoke, [one new concrete development: "we onboarded 500 new nodes" / "got a warm intro to [pharma company] for a data pilot" / "closed another $X from an angel"].
Still keen to continue our conversation if there's interest on your end. Happy to hop on a quick call whenever works.
Howard
如果明确拒绝 — 保持关系
Subject: Re: Following up — Oysterworld
Hi [Name],
Totally understand — timing and fit matter. I appreciate you being direct about it.
If you don't mind me asking: what would need to be true for this to be interesting in the future? I'd love to come back when we hit that milestone.
Either way, I'll keep you updated on major milestones. Thank you for the time.
你说: "I appreciate the offer. At $4M revenue and 60% margins, $15M would be a 3.8x revenue multiple. That's below typical SaaS early-stage multiples, and we're a profitable hardware company with a network asset. I'm open to finding the right number, but $20M is the floor that reflects the business fundamentals. Would $25M work as a middle ground?"
Rule: Never accept the first counter. Always have a middle ground ready. $25M is your real floor. $30M is your ask.
9. Pitch 练习模块 — 自我训练
一个人练到条件反射。每个 pitch 都应该像呼吸一样自然。
60 秒 Elevator Pitch (必须能脱口而出)
"Oysterworld is building the perception layer for AI agents in the physical world.
We deploy consumer devices — phones, wearables, smart glasses — that act as sensor nodes. Each device collects health and environmental data, verified by physics equations through our PINNs engine. Fake data can't pass. Real data gets monetized.
We've sold 25,000 devices, generated $4 million in revenue with zero outside funding, at 60% gross margins.
Now we're launching the Verified Fitness Foundation Model — training AI on physics-verified real-world data from our sensor network.
We're raising $1 million on a SAFE at $30 million post to accelerate distribution and data monetization."
计时器: 对着镜子说，目标 55-60 秒。超过 65 秒就太长了。
2 分钟 Full Pitch (会议开场)
[问题 — 10秒]
"AI is getting smarter every day, but it still can't reliably sense the physical world. Health data gets spoofed. Environmental data gets fabricated. There's no trust layer for physical intelligence."
[方案 — 15秒]
"Oysterworld is an open sensor network. We put hardware in people's hands — phones, wearables, glasses — that collect real-world data verified by physics equations. Our PINNs verification engine uses cardiovascular models, metabolic equations, and signal propagation physics to mathematically prove data is real."
[Traction — 20秒]
"We've sold 25,000 devices, generated $4M in revenue with zero funding, at 60% margins. All bootstrapped. 100,000-person community. And we don't need this money to survive — hardware already covers all costs. EBITDA positive."
[Foundation Model — 20秒]
"Now we're launching the Verified Fitness Foundation Model — an AI model trained on physics-verified real-world data from our 25,000-node sensor network. No one else has this data. No one else can verify it at this scale."
[The Ask — 15秒]
"We're raising $1M on a post-money SAFE at $30M — that's 7.7x trailing revenue. Use of funds: 80% go-to-market, 20% technology. Path to Series A in 12-18 months at $75-100M."
[Hook — 10秒]
"Happy to dive into any area — PINNs technology, unit economics, or the data monetization roadmap. What's most interesting to you?"