# Claim audit: “The Scaling and Profitability Trade off”

**Audit date:** 2026-09-05  
**Verdict key:** **Supported** = the stated claim follows from identified evidence; **Qualified** = directionally or partly supported but the wording, metric, sample, or causal interpretation must change; **Unsupported** = the cited evidence does not establish it or contradicts it; **Unverifiable** = the relevant underlying source, context, or microdata are unavailable.

The units below are deliberately separated: company, investment, fund, manager/GP, and LP are not interchangeable; neither are exit value, marked value, proceeds, IRR, TVPI, DPI, PME, or alpha.

## Khosla statement and framing

The article includes a screenshot in which Khosla favors investment in future assets and discusses data-center spending. This publication paraphrases it; the source screenshot is linked in the original article.

The image does not preserve the post being answered, thread, timestamp, edits, or replies. A secondary report dates the disagreement to August 2026, but that does not recover the missing context. [Economic Times report](https://economictimes.indiatimes.com/markets/stocks/news/aswath-damodaran-disagrees-with-vinod-khosla-on-the-scaling-vs-profitability-debate-heres-why/articleshow/133730584.cms)

The text contrasts current profit/cash retention with investment in future assets, particularly data centers. It does **not** literally say every firm should scale regardless of economics. Damodaran acknowledges uncertainty about intent but then generalizes the statement into a wider VC norm; that interpretive move is not established by the screenshot.

## Detailed audit

| # | Claim as made or implied | Source/date | Definition/sample | Verdict | Explanation |
|---:|---|---|---|---|---|
| 1 | Khosla said profitability admits the firm has no higher-return investment and favored future assets over cash. | Article image; original X thread not recovered; likely Aug 2026. | One screenshot; surrounding post/reply context absent. | **Qualified** | The statement is paraphrased above; the screenshot appears in the original article. Treating them as “every business should put scaling first” is an interpretation, not the literal statement. The missing context stays missing. |
| 2 | Prioritizing scale over profit is the norm among many VCs and has intensified over two decades. | No direct empirical citation. | Undefined population of VCs, geographies, stages, and years. | **Unsupported** | Survey and contracting evidence show heterogeneous selection, monitoring, and value-add practices. A time trend in stated or revealed preferences is not estimated. [Gompers et al.](https://www.nber.org/papers/w22587) |
| 3 | Scale creates value only conditionally on market size/growth, unit economics, fixed-cost leverage, competition, and capital intensity. | Article’s analytical framework. | Firm/project-level incremental cash flows. | **Supported** | This follows from standard incremental NPV logic. It is the strongest part of the article, though “unit economics” must include all causally variable costs and capital needs. |
| 4 | The 2026 “NCVA” Yearbook shows continued growth in active VCs, cumulative capital, and AUM through 2026. | [2026 NVCA Yearbook](https://nvca.org/2026-nvca-yearbook/), released 2026-04-13; data primarily through 2025. | PitchBook/NVCA U.S. VC data; supplemental XLS is members-only. | **Qualified** | The organization is **NVCA**, not NCVA. The report confirms a large industry but also says firm count fell to 2,984 in 2025, the first decline. A 2026 endpoint in the article’s chart is not supported by a yearbook whose annual observations end in 2025. Exact chart reconstruction is blocked by the nonpublic data pack. |
| 5 | 2025 VC deployment remained very large. | [NVCA release](https://nvca.org/press_releases/nvca-releases-2026-yearbook-charts-a-venture-industry-in-transition/), 2026-04-13. | $320B across 15,352 U.S. deals in 2025; AI 65.4% of deal value. | **Supported** | The level is documented, but it coexists with slower fundraising, constrained liquidity, and high concentration. It does not itself show excess or value destruction. |
| 6 | “The median venture capitalist” has not harvested much alpha, as shown by Cambridge Associates. | [Cambridge Associates CY2025](https://www.cambridgeassociates.com/insight/us-pe-vc-benchmark-commentary-calendar-year-2025/), 2026-07-31. | Capitalization-weighted pooled horizon IRR, net of fees/carry, across 2,816 funds; public benchmarks are mPMEs. | **Unsupported** | A pooled, capitalization-weighted **fund index** is not the median VC, GP, fund, investment, or LP. Alpha also requires risk adjustment; raw IRR minus an index number is not automatically alpha. |
| 7 | VC returns in the table underperform Nasdaq in most periods. | Same Cambridge source; horizons ended 2025-12-31. | Pooled net VC IRR versus cash-flow-matched public mPME. | **Qualified** | The displayed horizons do show mixed relative performance and several Nasdaq shortfalls; 2025 was strong for VC at 21.1%. Results are horizon-, vintage-, weighting-, and valuation-dependent. Cambridge itself calls performance versus Nasdaq inconsistent. |
| 8 | Top 1% of VC investments generated 80% of all VC investor returns in 2023–2026. | [CF Private Equity](https://www.commonfund.org/cf-private-equity/ai-is-steepening-the-power-law-in-venture-capital?hs_amp=true), 2026-07-13; PitchBook data to 2026-06-18. | U.S. VC-backed **exit events** ≥$10K; share of total **exit value**. | **Unsupported as phrased; qualified underlying fact** | The source says top 1% of **companies represented 80% of exit value**, not investor returns. Excluding SpaceX lowers it to 45%. Entry basis, ownership, dilution, preferences, fees, and unrealized positions are absent. The window is incomplete and the microdata are proprietary. |
| 9 | The top 10% deliver the bulk of VC returns across periods. | Same CF/PitchBook chart. | Company exit-value concentration in four exit-date bins. | **Qualified** | The chart supports concentration of gross exit value. It cannot identify concentration of investment MOIC, fund IRR, DPI, or net LP wealth. Exit-date cohorts also mix many investment vintages. |
| 10 | Power-law skew means only about one-quarter of VCs beat the average each year. | No distributional source. | Undefined “VC,” return metric, weighting, year, and arithmetic/geometric mean. | **Unsupported** | Skew does not mathematically determine the fraction above the mean. Simple skewed samples can have 20%, 40%, or 80% above their means. The saved script demonstrates this directly. |
| 11 | VC success is more persistent than mutual-fund or buyout success. | [Kaplan & Schoar (2005)](https://doi.org/10.1111/j.1540-6261.2005.00780.x); [Harris et al. (2023)](https://doi.org/10.1016/j.jcorpfin.2022.102326). | Successive fund performance; older Venture Economics and later Burgiss cash-flow samples. | **Qualified** | Persistence is documented, especially for VC, and later work finds it using information available at fundraising. It is not immutable: access, reputation, vintage, selection, and measurement contribute. [Nanda, Samila & Sorenson](https://doi.org/10.1016/j.jfineco.2020.01.004) |
| 12 | Concentrated outcomes cause VCs to push firms toward reckless scale and gambling. | CF concentration description plus article interpretation. | No treatment/control or contract-level incentive test. | **Unsupported causal claim** | Portfolio convexity makes upside-seeking plausible, but ownership, follow-on reserves, reputation, downside, staged financing, and control rights can pull in different directions. Concentration is the outcome to explain, not an instrument identifying behavior. |
| 13 | Mutual funds increasingly invested in private companies. | [Kwon, Lowry & Qian (2020)](https://doi.org/10.1016/j.jfineco.2019.10.003), sample to 2016. | 16 mutual-fund families and VC-backed private firms; the third author is **Yiming Qian**. | **Supported** | The number of participating funds and dollar value rose sharply. The chart is historical and cannot by itself quantify the 2026 market. |
| 14 | Mutual-fund participation lets companies stay private longer. | [Authors’ summary](https://corpgov.law.harvard.edu/2017/04/18/mutual-fund-investments-in-private-firms/), 2017; final paper 2020. | 2SLS using VC–mutual-fund connections; estimates about 2.8 years longer and ~6% lower quarterly exit hazard. | **Supported with design caveat** | This is substantially stronger than a raw correlation. Instrument validity remains an identifying assumption, and the result is about financing and listing timing—not profitability or value destruction. |
| 15 | Crossover investors participate primarily because of FOMO and validate inflated VC marks. | Article discussion; no direct test. | Motives of mutual funds, sovereign funds, and other nontraditional investors. | **Unsupported** | Kwon et al. find delayed listing and return/IPO-allocation motives matter. Contract evidence also emphasizes liquidity and rights. “FOMO” is possible but not identified. [Chernenko et al.](https://doi.org/10.1093/rfs/hhaa100) |
| 16 | More private capital helped companies grow larger and remain private longer. | [Ewens & Farre-Mensa (2020)](https://doi.org/10.1093/rfs/hhaa053). | NSMIA-based quasi-experimental evidence for late-stage U.S. startups. | **Supported** | The study supports a financing channel and stronger founder bargaining power, while explicitly cautioning that NSMIA was not the sole cause and that the new equilibrium was not simply IPO-market failure. |
| 17 | Passive investing is above 50% and weakens fundamental price anchoring. | No consistent denominator or direct citation in article. | Could mean U.S. mutual-fund assets, all fund assets, equity ownership, trading, or market cap. | **Unverifiable as stated / unsupported mechanism** | The share depends on the denominator. More importantly, the article does not link passive flows causally to the plotted factor, VC pricing, or firm investment. Passive funds can influence governance. [Appel, Gormley & Keim](https://doi.org/10.1016/j.jfineco.2016.03.003) |
| 18 | Reversal has faded, leaving fundamental value investors without an anchor. | [Kenneth French data library](https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_Library.html), data downloaded 2026-09-05. | Article uses short-term reversal formed on prior-month returns; value HML and long-term reversal are separate factors. | **Unsupported causal interpretation** | Recent short-term reversal is weak/negative under several aggregations, but a one-month contrarian factor is not a measure of price-to-fundamental convergence. The article’s exact transformation code is absent and the chart is not exactly reproducible. |
| 19 | Momentum and value/reversal are natural opposites. | Article narrative. | Momentum uses prior 2–12 months; ST reversal uses prior month; LT reversal uses 13–60 months; HML uses book-to-market. | **Unsupported simplification** | These are distinct signals with different horizons and constructions. A fall in one does not imply a rise in another or identify a causal market regime. [French ST reversal definition](https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/Data_Library/det_st_rev_factor.html) |
| 20 | Median IPO age rose by about 11 years in the last 15 years. | [Ritter IPO data](https://site.warrington.ufl.edu/ritter/ipo-data/), public series through 2025. | U.S. operating IPOs after exclusions; age = IPO year minus founding year under documented conventions. | **Unsupported** | The annual all-IPO median was 11 in 2011 and 12 in 2025: roughly +1 year, not +11. The defensible wording is that median age rose **to around 11–12 years**. |
| 21 | Inflation-adjusted median IPO revenue tripled or quadrupled versus the 1980s. | Ritter sales table, updated 2026-02-25. | Real trailing sales; annual median; article averages annual medians by period. | **Qualified** | The broad level shift is supported, but results are composition- and cycle-sensitive. A mean of annual medians is not the pooled firm median; recent annual medians range from $1M in 2022 to $180M in 2025. |
| 22 | Over 80% of 1980s IPOs were profitable, versus less than one-quarter in the last decade. | [Ritter review](https://site.warrington.ufl.edu/ritter/files/2016/01/A-Review-of-IPO-Activity-Pricing-and-Allocations-2002-08.pdf) and updated sales tables. | Positive trailing earnings; U.S. operating IPO sample. | **Qualified** | Older research reports 19% negative earnings in the 1980s, supporting ~81% profitable. Recent profitability is much lower, but exact “last decade <25%” depends on firm-weighting versus averaging annual percentages, exact years, and industry composition. |
| 23 | The market-cap/float chart describes tech IPOs through 2026. | Article figure attributed to Ritter. | Figure period counts versus Ritter all-IPO and tech samples; public cutoff 2025. | **Unverifiable / likely mislabeled** | The figure’s counts appear to match all operating IPOs rather than Ritter’s tech-only series, while its title says tech. The 2020–2026 bin extends beyond the cited public data. Exact custom aggregation is undisclosed. |
| 24 | Recent IPO median market cap exceeds $1B and proceeds are a shrinking fraction of market cap. | Article custom figure; Ritter public summaries. | Period averages/medians, unclear all-versus-tech population; incomplete 2026. | **Qualified direction; exact values unverifiable** | Larger private financing and later listing plausibly lower primary float, but the chart’s labeling/sample inconsistency prevents exact verification. Proceeds/market cap is not itself a measure of capital need or value creation. |
| 25 | SpaceX went public in June 2026 at about $1.8T. | [SpaceX 10-Q](https://www.sec.gov/Archives/edgar/data/1181412/000162828026052535/spcx-20260630.htm); [SEC offering disclosure](https://www.sec.gov/Archives/edgar/data/1181412/000162828026040874/spacexukfwp.htm). | 638.9M offered shares at $135; pro forma A+B shares 13.076B. | **Supported with definition** | The IPO occurred; $135 times pro forma shares is ~$1.765T, reasonably rounded to $1.8T. It is an offer-price implied equity value, not proceeds or the ~$2.1T first-day closing market cap. [Nasdaq](https://www.nasdaq.com/newsroom/spacex-makes-history-raising-85-billion-through-nasdaq-listing) |
| 26 | OpenAI and Anthropic have trillion-dollar-plus IPO “pricing promises.” | Article references anticipated listings. | Forecast or private-market expectation, not completed offers as of research date. | **Unverifiable as outcomes** | Pipeline expectations are not IPO pricing evidence. They must not be added to realized exit statistics as if complete. |
| 27 | Early Google and Facebook scaled while profitable. | [Google 2004 S-1](https://www.sec.gov/Archives/edgar/data/1288776/000119312504073639/ds1.htm); [Facebook 2012 filing](https://www.sec.gov/Archives/edgar/data/1326801/000132680113000003/fb-12312012x10k.htm). | Google 2003 and Facebook 2011 pre-IPO financials. | **Supported** | Google reported $961.9M revenue and $105.6M net income for 2003; Facebook reported $3.711B revenue and $1.756B operating income for 2011. These are selected winners, not prevalence estimates. |
| 28 | Amazon rationally put scale before profit. | [Amazon 1997 letter](https://www.sec.gov/Archives/edgar/data/1018724/000119312517120198/d373368dex991.htm). | One company; contemporaneous customer/revenue evidence and stated capital discipline. | **Qualified** | Amazon explicitly prioritized growth, but also tracked repeat purchases, tested returns, promised to drop failed programs, and balanced growth with long-run profitability/capital management. It supports evidence-gated scale, not blind scale. |
| 29 | WeWork’s long-lease/short-membership duration mismatch made scaling dangerous. | [WeWork filing](https://www.sec.gov/Archives/edgar/data/1813756/000181375623000016/we-20221231.htm), 2022. | Future undiscounted lease payments $27.9B; committed sales $2.5B; ~15-year leases versus 19-month memberships. | **Supported mechanism, not universal inference** | Fixed long liabilities financed cancellable/shorter revenue expose the firm to occupancy and price shocks. Governance and financing also mattered; one case cannot estimate the average effect of scale. |
| 30 | Ferrari is a high-value niche star because it limits volume. | [Ferrari FY2025 results](https://www.sec.gov/Archives/edgar/data/1648416/000164841626000018/fnvfy2025results.htm). | 13,640 shipments; €7.146B revenue; 29.5% EBIT margin; €1.538B industrial FCF. | **Qualified** | Constrained volume and premium economics are documented. The causal “because” and market-cap comparison require a counterfactual and same-date valuation; brand, mix, personalization, and scarcity jointly matter. |
| 31 | Remaining private longer weakens governance and causes delayed business-model building. | Article interpretation; related evidence from Lerner/Nanda and Ewens/Farre-Mensa. | Governance rules, founder control, private financing, and later operating outcomes. | **Qualified concern; causal outcome unsupported** | Large private firms avoid some public disclosure and oversight, and evidence shows stronger founder bargaining power plus concern over relaxed VC governance. No cited design shows that listing delay itself causes worse unit economics or economic profit. [Lerner & Nanda](https://doi.org/10.1257/jep.34.3.237) |

## Cross-cutting causal verdict

The article identifies plausible mediators:

$$
More\ private\ capital \rightarrow longer\ private\ life \rightarrow different\ control\ and\ financing\ constraints.
$$

The first arrow has credible empirical support. The next proposed chain—toward scale-first management, weak profitability, inflated exits, and eventual social loss—is mostly correlational. Company selection, sector composition, technology, falling early capital requirements, intangible accounting, macro rates, listing rules, and winner survivorship are alternative explanations. Establishing causation requires linked company-level financing, contract, cohort economics, governance, and outcome data with a credible source of exogenous variation in scale speed or investor incentives.