The $0–100bn valuation band: segment structure of the primary market0—1000 亿美元估值带：一级市场的分段结构

1. The question: why "what number should we put on this round" is the wrong question

The primary market has no continuous quotation, so "how much" must first be replaced by "which segment".

The most expensive mistake in a financing negotiation is not answering wrongly. It is answering the wrong question. Glacier Institute has a fixed formulation for this: Answering the wrong question costs more than answering wrongly. (「答非所问，比答错更贵。」)

Why is the primary market so prone to answering the wrong question? Because its pricing structure differs from the public market's. The public market quotes a price every second, and the price itself carries information. The primary market completes one pricing only in the few weeks when a window is open. The rest of the time there is no price, only claims. So the two sides easily talk without coordinates: one side explains the technology at its own pace, the other asks for validation from its own checklist. Both feel they are having a serious conversation. In fact they are sitting two different examination papers.

Glacier Institute states this mismatch very directly:

The most common mistake is bringing the previous segment's answer sheet to the next segment's exam. However cleanly you explain the technical path, once you are in the crossing segment the exam is industrial validation. However good the growth data, one segment up the exam is shareholder structure and global carrying capacity. The hard part is that both sides think they are having a serious conversation. (「最常见的错，是拿上一段的答卷去考下一段的题。技术路线讲得再干净，进了跨越段，考的是产业验证；增长数据再漂亮，再上一段，考的是股东结构与全球承载力。为难之处在于，双方都以为在认真对话。」) — Glacier website, /institute.html, Glacier Institute essay No. 18, Classify First, Then Talk Price（《先分后价》）

So before talking price, a more basic question needs answering: which segment is this company standing in right now. Because price is a conclusion, not a starting point.

The finance literature supports this ordering from another direction. Gornall and Strebulaev [3] modelled 135 US unicorns and found that reported post-money valuations exceeded fair value by 48% on average; 14 of them were more than double. The gap did not come from the companies themselves. It came from the protective terms held by later-round investors: guaranteed IPO returns, vetoes over low-priced IPOs, liquidation seniority over other investors, and so on. In other words, an isolated valuation number was never a comparable quantity in the primary market to begin with. If the number itself is not comparable, then putting the company back into the segment of questions it faces is the steadier coordinate.

---

2. The framework: three segments, three examination papers

The three segments are not three thresholds but three rulers — each segment examines one thing only.

What follows is the original public formulation of the Glacier Institute framework. This paper quotes it as it stands, without rewording (see Table 1):

Table 1 / Exhibit 1　The three segments of the 0–100B USD valuation band and the proposition of each.

Source and notes: The text in the table is the original public formulation of the framework on the Glacier website, not reworded by this paper; the boundaries of the price bands are segment marks and do not express any valuation commitment.

Valuation band	Proposition	What this band answers
US$0–1B	Thesis	why this team, this route, this form, this timing
US$1–10B	Crossing	benchmark-customer validation, real growth data, a diversified shareholder structure, and milestone financing
US$10–100B	Succession	global governance architecture, platform-ecosystem logic, and the capacity to connect to global public capital markets

The expanded formulation also comes from the Glacier website:

US$0–1B, "thesis" stage: the core proposition to answer — why this team, this route, this form, this timing. (「0—10 亿美元（立论阶段）：回答核心命题——为什么是该团队、该技术路线、该产品形态以及当前的市场时机。」) US$1–10B, "crossing" stage: customer validation, growth evidence, shareholder-structure optimization and milestone financing, from a technology team to an industrial company. (「10—100 亿美元（跨越阶段）：以标杆客户验证、真实增长数据、多元股东结构与阶段性里程碑融资，完成从技术型团队向成熟产业型公司的跨越。」) US$10–100B, "succession" stage: global governance architecture, platform-ecosystem logic, and the capacity to connect to global public capital markets. (「100—1000 亿美元（承接阶段）：搭建全球化治理架构、平台化生态逻辑以及对接全球公开资本市场的综合承载力。」)

The relative positions of the three segments are shown in Figure 1. The purpose of the framework also has a fixed formulation:

The purpose is to match price, financing amount, dilution, milestones, investor type, and the company's next-stage capacity. (「框架的目的，是让价格、融资额、稀释度、里程碑、投资人类型与企业下一阶段承载力互相匹配。」)

It is followed immediately by a qualifying clause whose position and weight cannot be omitted: an analytical framework, not a valuation commitment. (「分析框架，非估值承诺。」)

Two things are worth pointing out first.

First, the boundaries of the three segments are not cut at random. The range of 0 to 100 billion US dollars, in the industry's notation, is 0–100B USD; the two boundaries fall at 1B and 100B. Both numbers were defined in the industry's vocabulary long ago: in 2013 Aileen Lee proposed "unicorn" for a private company valued above US$1 billion, and in the same article called those above US$100 billion "super-unicorns" [4]. Glacier Institute's segmentation coincides with these two accepted marks. The framework uses the ruler the market already uses rather than building its own coordinates.

Second, "only one proposition" per segment is a strong constraint, not a summary. One segment, one examination paper: the segment should not have its time taken up by other questions. This point is tested in reverse by the counterexamples later in the paper.

---

3. The segments argued

The thesis segment cannot produce data; it can only produce a causal chain that survives reverse checking.

3.1 The thesis segment (US$0–1B): turning what cannot be proved into what can be evidenced

The four elements of the thesis segment — team, technical route, product form, market timing — share one feature: at this stage none of them has external data to vouch for it. There are no customers at scale, no growth curve to extrapolate, and often no comparable company at all.

Finance has a mature explanation for this position. Gompers [1] shows that venture capital adopts staged financing precisely because the more intangible the assets, the more growth options and the higher the asset specificity, the higher the agency costs and information asymmetry. The response is not to provide all the money at once but to cut it into tranches, each tranche matched with one round of monitoring and re-evaluation. Based on a random sample of 794 venture-backed firms, the paper finds that investment is indeed concentrated in early-stage and high-technology companies, where information asymmetry is highest.

Translated onto the valuation band: the essence of the thesis segment is not "how much is it worth" but "on what grounds can one begin to believe" (「凭什么可以开始相信」). What the investor buys at this point is a set of judgements not yet validated, not a set of validated data. So the work of this segment centres on evidence, not price.

Glacier Institute formulates this way of evidencing as the standard of the Constructing layer (「构建」层):

The standard is coherence, not completeness. Every number and every chart has to survive being checked backwards. The work at this layer is, plainly, producing evidence. (「标准是自洽，不是齐全。每一个数字、每一张图，都要经得起反过来查。这一层的活，说白了就是举证。」) — Glacier Institute essay No. 20, From Outside In, Tightened Four Times（《由外而内》）

Coherence before completeness is an ordering specific to the thesis segment. Completeness is a requirement from the crossing segment onwards, when there is real data to lay out. The thesis segment cannot produce that much data. What it can produce is a causal chain that does not contradict itself. Complete the chain, and let the other side reach the conclusion themselves.

3.2 The crossing segment (US$1–10B): from a technology team to an industrial company

The four elements of the crossing segment — benchmark customer validation, real growth data, a diversified shareholder structure, milestone financing — point to one thing: the company must prove it no longer depends on the personal capabilities of the founding team.

What happens here is a change of identity. Glacier Institute's formulation is "from a technology team to an industrial company" (「完成从技术型团队向成熟产业型公司的跨越」). Hellmann and Puri [2], using hand-collected data on Silicon Valley start-ups, give this transition an observable form: venture-backed companies build human-resource policies earlier, introduce stock option plans earlier, set up a marketing function earlier, and are more likely, and quicker, to replace the founder with an outside CEO. They call this set of changes "professionalization" (「职业化」).

In other words, the crossing segment does not examine whether the technology can get stronger. It examines whether the organisation can carry more money, more customers and more shareholders. Benchmark customer validation answers "has the outside world already recognised this route with real money" (「外部世界是否已经用真金白银承认这条路线」); real growth data answers "can that recognition be repeated" (「这种承认能否重复」); a diversified shareholder structure answers "does the company still depend on the will of a single investor" (「公司是否还依赖某一家资方的意志」); milestone financing answers "can the capital cadence be checked from outside" (「资本节奏是否可以被外部核对」).

The last item is especially easy to underestimate. Elsewhere, Glacier Institute gives milestones a very hard criterion:

A clear milestone is one an outsider can sign off. (「所谓里程碑（milestone）清楚，就是外行也能验收。」)

If an outsider can sign off, the company's progress no longer depends on insiders' explanations. This is the watershed between the crossing segment and the thesis segment: the thesis segment relies on explanation, the crossing segment on sign-off.

The crossing segment is also where answering the wrong question most often happens. The reason is that the thesis segment's answers still sound good here: the technical route is still clean, the team still excellent, the timing still valid. But this segment examines industrial validation. Material that keeps talking about the technical route is not wrong material. It is expired material.

Takeaway | The watershed of the crossing segment: the thesis segment relies on explanation, the crossing segment on sign-off.

3.3 The succession segment (US$10–100B): who takes the next leg

The three elements of the succession segment — global governance architecture, platform-ecosystem logic, the capacity to connect to global public capital markets — together answer one question: after the primary market, who takes over.

The size of capital in this segment limits the kinds of subsequent buyers: sovereign and quasi-sovereign capital, multinational industrial players, global long-horizon funds, and finally the public capital markets. Each has its own entry conditions, and most of them are "fail and you are out" (「不满足就直接出局」), not "can be compensated by price" (「可以用价格补偿」). Whether the governance architecture withstands review across jurisdictions, whether data and supply chains can be compliant across borders, whether the revenue structure can be read by public-market accounting standards — these are not valuation discounts. They are qualifying conditions.

Platform-ecosystem logic is another layer. At this size a company can hardly convince buyers with the growth of a single product; it needs to prove it is the foundation for others' growth. Glacier Institute writes the difficulty of this layer into Controlling (「控局」):

What makes a late-stage deal hard is not the valuation, it is the number of interests inside it. Existing shareholders wanting out, the option pool, leftover performance ratchets, disagreements between management and finance — each additional party is another veto. The workload does not grow linearly with the amount, it grows combinatorially with the number of parties. (「后期交易难的不是估值，是利益关系太多。老股东的退出诉求、期权池、对赌残余、管理层与财务口的分歧，每多一方，就多一道否决。工作量不随金额线性增长，而随相关方数量组合式增长。」)

This passage explains a counter-intuitive phenomenon: the main workload in the succession segment is not pricing but coordination. At this size, price is usually the interval left after the constraints of all parties have been intersected, not the product of negotiating skill.

Takeaway | In the succession segment, most conditions are qualifying conditions, not discount items.

---

4. Two horizontal mechanisms

What the buyer fears is not the price, it is having no coordinates.

The three segments are the vertical structure. The framework also has two horizontal mechanisms that run through all three.

4.1 One step earlier than classification: who does the classifying

Glacier Institute has a sentence that is easy to skip:

One step earlier than classification is who does the classifying. The same company filed under an application industry, or under a class of foundational capability, draws completely different comparables, investment-committee language and tolerance. Early commercialisation usually has only one use case, and at the level of fact it does look like that industry. Do not fight for this card, and someone else nails down the valuation anchor first. (「比分类更早的一步，是被谁分类。同一家公司，被归进某个应用行业，还是被归进一类基础能力，对应的是完全不同的对标（comparable）、投委会话术与容忍度。早期商业化往往只有一个场景，事实层面也确实像那个行业。这张牌不争，估值的锚就被别人先钉住了。」)

This is a judgement prior to the valuation band. The band answers "which segment"; "who does the classifying" answers "which set of comparables to measure by" (「按哪一套对标来量」). Confusing the two produces a particularly hidden error: a company clearly in the crossing segment is measured against mature companies in some application industry, so all its growth data look insufficient; or the reverse, it is measured against a class of foundational capability, so all industrial validation is deemed not yet necessary.

Note that early commercialisation usually has only one landing scenario, so "looks like that industry" (「看起来像那个行业」) is often true at the level of fact. That is why this card must be fought for actively: it is not an argument about facts but an argument about the frame of reference.

4.2 What the buyer wants is not a low price but coordinates

The framework's second horizontal effect takes place at the negotiating table:

From what we see, what investors resist is mostly not the absolute price but their inability to place it — they do not know who set the last round or who is meant to take the next. With a lower round underneath and a higher band moving in parallel above, every price has a reference and expensive stops being the problem. As we see it, the first duty in investing is safety, not odds. (「据我们了解，投资人抗拒的多数不是绝对价格，而是无法定位这个价格——不知道上一轮是谁给的，下一轮又该由谁接。前面有更低的轮次垫着，后面有更高的带子同步推进，每个价位都有参照系，贵就不再是问题。目前看，做投资的第一要务是安全，不是赔率。」)

This passage and the Gornall–Strebulaev [3] conclusion cited earlier corroborate each other. If an isolated valuation number is not comparable because terms differ, then what reassures the buyer cannot be the number itself. It can only be the structure around the number: who priced the segment below, who is moving in parallel in this segment, which kind of capital takes over in the segment above. That structure is exactly what the valuation band provides.

A practical criterion follows: forcing a valuation up is not worth much. "A price can be pulled up once. It cannot be held up a second time." (「价格能被拉高一次，但是拉不住第二次。」) The one time it is pulled up does not change the segment the company is in. It only brings the next segment's examination paper forward — and that paper does not get easier because the price went up.

---

5. Boundaries and counterexamples

The framework describes "which segment right now"; it does not predict "bound to move up".

A framework with only positive examples does not deserve trust. The following three situations are where this framework is known to fail or to need additional explanation.

Counterexample one: companies whose segment and question are out of step. There is indeed a class of companies whose segment sits far above the questions they can answer — typically independent entities incubated by platform companies, or second ventures by mature teams. They may stand in the crossing segment's price while holding only the thesis segment's answers. Glacier Institute's response to this is guarded: "Are there companies that belong in a higher band by nature? Yes. But the band is not chosen, it is decided by those three exam papers." (「会不会有公司天生就该在更高的带子里？会啊。但是带子不是自己挑的，是被这三张考卷判出来的。」) That is, the framework admits such companies exist, but does not admit they can skip the paper. They only answer it faster.

Counterexample two: companies whose valuation does not rise segment by segment. The three-segment structure is progressive in its formulation, but real paths often are not. Down rounds, structural reorganisations, spin-offs and mergers all move a company backwards between segments. The framework itself contains no assumption about direction in time. It describes "which segment right now" (「此刻在哪一段」), not "bound to move up" (「必然会往上走」). Reading progression as inevitability is the most common misreading of this table.

Counterexample three: companies that do not advance on equity financing. A considerable share of hard-tech companies advance mainly on orders, government special programmes, prepayments from industrial partners or debt instruments; equity financing is only a supplement. For such companies the valuation band can describe the capitalisation stage but not the operating stage, and the two may differ by a whole segment. Using the band to judge the maturity of such companies will systematically underestimate it.

There is also a boundary of formulation that must be stated: this paper discusses a classification framework, not a pricing method. The framework produces no price and endorses no interval. In its public formulation Glacier Institute places this sentence directly beneath the table rather than in a footnote, and this paper keeps that position — an analytical framework, not a valuation commitment.

---

6. Within what range the framework holds

The framework's range of validity: US-dollar denominated, hard tech, the capitalisation stage of equity financing; it classifies, it does not price.

A framework that cannot state its range of validity has lost a dimension. The following are the boundaries within which this paper holds the framework to be valid.

One, currency and measurement range. The framework is denominated in US dollars and covers 0 to 100 billion US dollars (that is, 0–100B USD). For companies above US$100 billion, the main pricing venue is no longer the primary market, and the framework no longer applies. Renminbi-denominated transactions can be converted and fitted, but exchange rates and cross-border structures change comparability.

Two, industry range. The framework comes from primary-market practice in hard tech and frontier technology. For industries with stable cash flows that can be priced directly by traditional valuation methods (such as mature consumer, traditional manufacturing, real estate), the explanatory power of these "segment propositions" (「分段命题」) falls markedly: the questions in those industries can be answered directly by a financial model, without classifying first.

Three, stage range. The framework describes the capitalisation stage of equity financing, not the company's operating maturity, nor its technical maturity. The three often move together, but not necessarily; see counterexample three above.

Four, what it cannot do. The framework does not predict price, does not give a valuation interval, does not judge whether a deal will close, and does not constitute investment advice of any kind. It does one thing only: it assigns the questions a company currently faces to a segment, so that both sides talk over the same examination paper.

Five, strength of evidence. The external studies cited in this paper [1][2][3] are all based on US market samples, spanning the 1990s to the 2010s. What they support is the general mechanism behind the framework — information asymmetry varying with stage, organisational professionalization accompanying the entry of capital, private valuations made non-comparable by terms — not the specific values of the three segment boundaries. The reasonableness of the boundaries comes from their coincidence with the industry's common marks (US$1 billion / US$100 billion) [4] and from Glacier Institute's own practice formulations; this paper does not claim they have been independently tested empirically.

Six, an honest gap. What this paper has not done is use public data to test "whether companies in a given segment do in fact have their financing questions concentrated on that segment's proposition" (「处于某一段的公司，其融资问题是否确实集中在该段命题上」). That needs a cross-segment sample with term-level detail, which public data cannot yet support. Until that work is done, the status of this framework is a falsifiable classification hypothesis, not a validated empirical regularity.

---

7. Conclusion

Pricing in the primary market is discontinuous, so an isolated valuation number cannot carry much information. The 0–100B USD valuation band divides the range into three segments — thesis, crossing, succession — and each asks one question only. Its value lies not in telling a company what it is worth but in telling both sides what to talk about now.

Three sentences summarise the paper:

1. Price is a conclusion, not a starting point — accept the segment first, then talk price.
2. Answering the wrong question costs more than answering wrongly — bringing the previous segment's answers to the next segment's exam is the most common and the most expensive mistake.
3. What the buyer fears is not the price, it is having no coordinates — the valuation band provides exactly those coordinates.

And the status of the framework itself, from beginning to end, is one sentence: an analytical framework, not a valuation commitment.

---

Eleven papers are not eleven topics. They are eleven links on one chain: segment first, then sequence, then take apart the one move in each link that most often goes wrong.

Read nextGI-WP-2026-P10　One Chain, Five Buckets: Why the Way a Value Chain Is Classified Decides the Way It Is Priced先被分类，再被分段。P10 是 P1 的上游。

FOR FOUNDERS

For founders这一轮怎么打、要多久、我们接不接，一次说清。进去看 →

转人工 · ASK A HUMAN把这个问题交给赵羚茜（Gogo Zhao，投资人关系负责人）安排——她会把它交到对的人手里。 联系庚辛

庚辛人生（Certainty of Money） · 庚辛研究院通讯论文 · GitHub · PDF 全文 · DOI · 可引用，CC BY 4.0。

References

Format: Author (Year). Title. Journal/Publisher, Volume(Issue), Pages. DOI or accessible link. The five entries below were verified as reachable, item by item, on 2026-09-20; this list contains no unverified entry.

[1] Gompers, P. A.（1995）. Optimal Investment, Monitoring, and the Staging of Venture Capital. The Journal of Finance, 50(5), 1461–1489. https://doi.org/10.1111/j.1540-6261.1995.tb05185.x

[2] Hellmann, T., & Puri, M.（2002）. Venture Capital and the Professionalization of Start-Up Firms: Empirical Evidence. The Journal of Finance, 57(1), 169–197. https://doi.org/10.1111/1540-6261.00419

[3] Gornall, W., & Strebulaev, I. A.（2020）. Squaring Venture Capital Valuations with Reality. Journal of Financial Economics, 135(1), 120–143. https://doi.org/10.1016/j.jfineco.2018.04.015（开放获取工作论文版：NBER Working Paper No. 23895, https://www.nber.org/papers/w23895）

[4] Lee, A.（2013）. Welcome to the Unicorn Club: Learning from Billion-Dollar Startups. TechCrunch, 2013-11-02. https://techcrunch.com/2013/11/02/welcome-to-the-unicorn-club/（访问日期 2026-09-20）

[5] Glacier Institute (2026). 0—1000 亿美元估值带分析框架 / 60 天标杆执行节奏 / 交易方法论 [The 0–100B USD valuation band framework / 60-day benchmark execution cadence / transaction methodology] [in Chinese]. Glacier Capital official website. https://glacier.mba/institute.html (accessed 2026-09-20)

Sources of quoted Glacier website passages

All website quotations are taken from [5] https://glacier.mba/institute.html (retrieved 2026-09-20), from the following passages:

• The "0–100B USD Valuation Band analytical framework" (「0—1000 亿美元估值带分析框架」) subsection (the three-segment table; the expanded formulation; "an analytical framework, not a valuation commitment" (「分析框架，非估值承诺」); "The purpose is to match price, financing amount, dilution, milestones, investor type, and the company's next-stage capacity" (「框架的目的，是让价格、融资额、稀释度、里程碑、投资人类型与企业下一阶段承载力互相匹配」))
• Glacier Institute essay No. 18, Classify First, Then Talk Price（《先分后价》） ("Answering the wrong question costs more than answering wrongly" (「答非所问，比答错更贵」); "price is a conclusion, not a starting point" (「价格是结论，不是起点」); "Three segments, three exam papers" (「三段，三张考卷」); "It sounds like three thresholds. It is closer to three rulers" (「听上去像三道门槛，其实更像三把尺子」); "One step earlier than classification is who does the classifying" (「比分类更早的一步，是被谁分类」); "What the buyer fears is not the price, it is having no coordinates" (「买方怕的不是贵，是没有坐标」); "forcing a valuation up is not worth much" (「硬拉估值意义不大」); "Are there companies that belong in a higher band by nature" (「会不会有公司天生就该在更高的带子里」))
• Glacier Institute essay No. 20, From Outside In, Tightened Four Times（《由外而内》） ("The standard is coherence, not completeness" (「标准是自洽，不是齐全」); "What makes a late-stage deal hard is not the valuation, it is the number of interests inside it" (「后期交易难的不是估值，是利益关系太多」); "The workload does not grow linearly with the amount, it grows combinatorially with the number of parties" (「工作量不随金额线性增长，而随相关方数量组合式增长」))
• The "60-Day Benchmark Execution Cadence" (「60 天标杆执行节奏」) subsection (context of the milestone formulations)
• Website homepage FAQ "What does the 0—100B USD price band mean?" (「0—100B USD 价格带是什么意思？」) (the US-dollar scale formulation)
• Website homepage FAQ "What does the 0—100B USD price band mean?" (「0—100B USD 价格带是什么意思？」) (the US-dollar scale formulation)

---

Glacier Institute 庚辛研究院. The $0–100bn valuation band: segment structure of the primary market[R]. Glacier Institute Working Paper GI-WP-2026-P1-EN. 2026. DOI:10.5281/zenodo.22869076.

Glacier Institute. (2026). The $0–100bn valuation band: segment structure of the primary market (Glacier Institute Working Paper No. GI-WP-2026-P1-EN). Zenodo. https://doi.org/10.5281/zenodo.22869076

@techreport{glacier2026p1en,
  author      = {{Glacier Institute}},
  title       = {{The $0–100bn valuation band: segment structure of the primary market}},
  institution = {Glacier Institute},
  type        = {Working Paper},
  number      = {GI-WP-2026-P1-EN},
  year        = {2026},
  doi         = {10.5281/zenodo.22869076},
  url         = {https://doi.org/10.5281/zenodo.22869076}
}