一句话： 好公司和好交易，靠的是同一个习惯：把少数几个问题想清楚，把答案说简单，再用对的人和体系，让答案一直成立。

1. 战略只回答两个问题：客户为什么选你？你怎么赚到钱？

低于成本卖也能拿到客户，但公司要走得远，两个答案必须同时成立。对合作伙伴也一样：先算对方能不能赚到钱，再算自己的，生意才做得长。

本周可以做： 补全这句话：我们向____卖____，客户付钱是因为____。如果每次说法不同，就从这里开始补。

2. 做不做一件事，看愿景和边界。今天有没有人，只决定快慢。

先画两条线：技术上从哪里做到哪里，客户是谁。线内的变化，要么是挑战，要么是机会；看不清怎么做时，先派两三个人去试，定好复盘的日子。线外的事，可以早一点、平静地说「不」。

本周可以做： 把两条线写在一页纸上，用它过一遍手上的每个项目。

3. 从一到十，先设计信息和业务怎么流动，再加管理者。

多设副总、多派亲信，管不住一家变复杂的公司。好的流程会让真实情况自己浮上来：合同签了，下一步却没有发生，系统就会显示。业务流是你和客户、供应商之间完整的交易过程，只回答速度、成本、质量；内部审批为它服务，组织跟着它走，才能真正授权。

本周可以做： 跟一张订单，从签约走到回款。标出它在哪里等待，由谁拍板。

4. 公司长大，不像一棵树慢慢长高，而是一次次变形。每个阶段要配不同的人。

把产品从零做到一的人，常常不是把公司做大的人；目标不是一直掌控公司，而是做对公司最有利的选择。找 CEO 看两点：从零做起过一件事，心里本来就想创业；再给他一个能降低创业风险的平台。

给投资人： 见创始人时问一句：下一个阶段，由谁来带队？

5. 技术是入场券。战略、运营和团队文化，决定公司能走多远。

文化说到底，是拿不准怎么选的时候，大家默认的那条规则；最怕的是每次选择，用的标准各不相同。招人也一样：最好的面试，是一起看两三个真实的项目，看能不能合作，也弄清对方想要什么。

本周可以做： 把团队最常遇到的两难，比如速度和质量，各写一句「默认怎么选」。

6. 你的优势，要是一个台阶，而不是一个折扣。

比大厂便宜一点，缺货时能拿到订单，却很难成为主流方案；能留下来的，是别人短期内做不到的事。产业客户最相信稳定的量产，几次送样只是起点；第一个标杆客户，最好选最挑剔的那一家。

给投资人： 问两个问题：哪件事别人短期内做不到？每个月稳定出货多少？

7. 你的故事，要经得起转述。

你见到的投资人，回去还要向老板和投委会汇报，所以故事要简单到小学生也能懂，一分钟内就让人想听下去。给最后拍板的人，再备好两句话的定位和一个他能亲自核实的事实；讲给大众听，就讲它在一天里怎么用、标价多少、谁愿意买。

本周可以做： 把路演压到五分钟，再请一位行业外的朋友复述。他漏掉的地方，就是要改的地方。

8. 故事要跟着阶段换。

早期先讲已经成立的事：利润、现金、真实的客户；等结果替你说话，再讲更大的技术故事。新概念先用市场已经听懂的词说清你是谁，再说明你是它的下一代。讲给谁听也要算清楚：做企业客户的公司，让行业里关键的几千人知道就够了。

本周可以做： 看一眼路演的第一页。它讲的，是你今天已经做到的事吗？

9. 报价只报一次，报真数。

信任来自前后一致：先在团队内部想清楚，再把价格、额度和时间一起讲明白，让对方能给出明确的答复。时间表按投资人真实的日历来排（很多人在长假前几天就已经休息了），也不要让一张特别大的支票，决定你整轮的节奏。

本周可以做： 每次见投资人之前，先把价格、额度和截止时间写成一句话。

10. 一个新的硬件品类，一般要三代才成熟。

第一代给爱尝鲜的早期用户，第二代修好问题、降低价格和门槛，第三代吸收市场反馈、靠口碑传开；一代大约一年到一年半。第一代要窄，样子和价格向人们熟悉的东西看齐，每一次演示都要证明一项底层能力，而不是表演。

给投资人： 问清产品在第几代、下一代要解决什么。带人工智能的产品，再问一句：换一个新任务要多少准备？看说明书、看两三次示范就能上手，是能规模化的好信号。

我们怎么用

庚辛康鹏资本（KANGPENG CAPITAL）在自己的工作里这样用：

• 做减法：我们把 80% 的注意力，放在 20% 的最头部项目上。
• 只接少数项目，长期陪伴，跟着公司的阶段，一起找下一阶段需要的人。
• 帮创始人准备一个五分钟的故事，让投资人能原样讲给自己的老板听。

马斯克和黄仁勋时代研究所的名字只为致敬一个时代，与马斯克、黄仁勋本人及其公司暂无合作关系，欢迎合作。

Answer the two questions first: ten notes for founders and investors

In one line: Good companies and good deals share one habit: get a few questions right, say the answers simply, and build the people and systems that keep them true.

1. Strategy answers two questions: why do customers choose you, and how do you make money?

Selling below cost can win customers, but to go far, both answers must hold at the same time. The same goes for partners: first work out whether they make money, then your own numbers. That is how a business lasts.

This week: finish one sentence: we sell ___ to ___, and they pay because ___. If the answer keeps changing, start there.

2. Decide what to do by your vision and boundaries. Today's team only sets the speed.

Draw two lines: from where to where you work in the technology chain, and whom you serve. Inside the lines, a change is either a challenge or an opportunity; when you can't yet see how to do something, send two or three people to try and set a review date. Outside the lines, you can say no early and calmly.

This week: put both lines on one page and check every project against it.

3. From one to ten, design how information and business flow before you add managers.

More deputies and trusted aides cannot hold together a company that has grown complex. A good flow lets the facts surface on their own: if a contract is signed but the next step never happens, the system shows it. The business flow is how you trade with customers and suppliers from start to end, and it answers only speed, cost and quality; internal approvals serve it, the org chart follows it, and only then can you truly delegate.

This week: follow one order from signing to payment. Mark where it waits and who decides.

4. A company does not grow like a tree. It changes form again and again, and each stage needs its own people.

The person who takes a product from zero to one is often not the one who builds the company; the goal is not to stay in control but to make the choice that is best for the company. When you look for a CEO, look for two things: they have built something from zero, and they already want to start a company. Then give them a platform that lowers that risk.

For investors: ask who will lead the company at the next stage.

5. Technology is the entry ticket. Strategy, operations and team culture decide how far a company goes.

Culture, in the end, is the default rule people follow when they are not sure how to choose; the real risk is a different standard each time. Hiring follows the same logic: the best interview is to look at two or three real deals together, to see whether you can work as one and what the other person wants.

This week: take the trade-offs your team meets most often, such as speed against quality, and write one line for each: our default choice.

6. Your edge should be a step, not a discount.

Being a little cheaper than a large player can win orders in a shortage, but it rarely makes you the standard; what lasts is something others cannot do soon. Industrial buyers trust stable volume, and a few good samples are only a start; for your first reference customer, pick the most demanding one.

For investors: ask two questions. What can this company do that others cannot do soon? How much does it ship, steadily, every month?

7. Your story must survive being retold.

The investor you meet must brief a boss and a committee, so make the story simple enough for a child and good enough that people want more after one minute. For whoever has the final say, add two sentences on where you stand among peers and one fact they can check; for a broad audience, say how it is used in a day, what it costs, and who would buy it.

This week: cut the pitch to five minutes and ask a friend outside your field to retell it. What they drop is what you fix.

8. Change the story as the company grows.

Early on, lead with what is already true: profit, cash, real customers; tell the bigger technology story when results speak for you. For a new idea, first use words the market already understands to say who you are, then show you are the next generation. Also be clear about who needs to hear it: a company selling to businesses only needs the key few thousand people in its industry to know it.

This week: look at the first page of your deck. Does it say what you have already done?

9. Quote your price once, and make it the real one.

Trust grows from consistency: decide the number inside the team first, then state the price, the amount and the timing together, so the other side can give a clear answer. Plan by the investors' real calendar (many are away several days before a long holiday), and do not let one unusually large check set your whole timeline.

This week: before each investor meeting, write the price, the amount and the deadline in one sentence.

10. A new hardware category usually needs three generations to mature.

The first generation serves early users who love new things, the second fixes problems and lowers the price and the barrier to use, and the third absorbs market feedback and spreads by word of mouth; one generation takes about a year to a year and a half. Keep the first generation narrow, let its look and price follow something people already know, and make every demo prove a core ability rather than put on a show.

For investors: ask which generation the product is on and what the next one will fix. If it runs on AI, also ask what it takes to learn a new task; learning from a manual or two or three demonstrations is a good sign that it can scale.

How we use this

KANGPENG CAPITAL uses these ideas in its own work:

• We subtract: 80% of our attention goes to the top 20% of projects.
• We take on few companies and stay with them for many rounds, and help each one find the people its next stage needs.
• We help founders build a five-minute story that an investor can retell to a boss.

The Musk–Huang Era Institute is named to honour an era. It has no current partnership with Elon Musk, Jensen Huang or their companies. We welcome one.