Datadog DDOG

> ## ⚠ EARNINGS IN TWO DAYS

> Datadog reports second-quarter 2026 results on 2026-08-06 — 48 hours after the date of this dive. Consensus is adjusted EPS $0.583 and revenue $1,078.7 million (+30.5% year on year). The stock closed on 2026-08-04 at $288.15, EXACTLY its 52-week high, up 5.32% on the day and 96.4% over three months. At 119 times FY2026 consensus earnings the reaction function into that print is asymmetric, and the verdict below is set with this timing as an explicit input.

Reference table

1. The business, and how good it actually is

Datadog is a Nevada corporation (redomiciled from Delaware in April 2026 — Section 4) headquartered in New York, NASDAQ-listed, with approximately 8,100 employees, led by co-founder and Chief Executive Olivier Pomel.

And the March 2026 quarter, which is the most recent data anywhere in this archive:

Three observations.

The acceleration is real and it is the most important fact here. Revenue growth went 27.0% → 26.1% → 27.7% across three years and then 32.1% in the March quarter. A 27%-grower re-accelerating to 32% at $4 billion of scale is rare, and it is the change that justifies a large part of the 112.5% twelve-month return. What the archive does NOT contain is any explanation: no net-retention figure, no customer count, no product-attach disclosure, and no artificial-intelligence revenue line. This dive does not assert a cause.

The cash generation is genuine but heavily assisted. Free cash flow of $1,000.6M on $3,427.2M of revenue is a 29.2% margin, achieved with capital expenditure of only 1.4% of revenue. But $750.7M of stock-based compensation is added back to reach it — 75% of the free-cash-flow figure. Cash flow after treating share compensation as the real cost it is falls to roughly $250M, a 7.3% margin. Both numbers are true and the second is the one an owner experiences.

GAAP operating income was NEGATIVE in FY2025 — a loss of $44,373 thousand, worse than FY2023's loss and a reversal of FY2024's $54,284 thousand profit — because research and development rose from 42.9% to 45.2% of revenue. That is a company spending to defend and extend a position, and it is a defensible choice; it is also the reason the trailing GAAP price-to-earnings ratio is 748.8x.

2. Valuation — priced in or room?

At $288.15 (market cap $102.57B; corrected enterprise value ~$99.63B after $2,939.6M of net cash — see Section 5):

Note the row that decides this dive: adjusted EPS grows 17.5% in FY2026 against revenue growth of 27.0%, and 17.8% in FY2027 against 21.3%. Adjusted margins are compressing, not expanding. At 119 times current-year earnings, the price requires the opposite.

The two valuation lenses disagree, and the disagreement IS stock-based compensation. At 18.9 times FY2027 consensus revenue on a corrected enterprise value, Datadog is expensive but within the band best-in-class software has commanded for a 21-27% grower with an 80% gross margin. At 101.1 times FY2027 adjusted EPS — where "adjusted" adds back 21.9% of revenue — it is not. The gap between those two statements is $750.7 million a year of real economic cost paid in shares, and a reader who values Datadog on sales is implicitly assuming it disappears.

Estimate coverage is deep on the near years and collapses hard. 33 and 34 analysts on FY2026 and FY2027 revenue, 24 and 25 on EPS — the deepest coverage in this batch — then 19 on FY2028 revenue but only 6 on FY2028 EPS, and 4 and 2 on FY2029 and FY2030 EPS. The FY2027 EPS range is $2.684 to $3.522, a 29.4% spread. FY2028's range is $2.267 to $4.927 — a 118% spread on six analysts — and FY2029 and FY2030 are excluded outright for the reason in Section 5: the estimate block shows revenue DECLINING between them.

est.ebitdaAvg and est.ebitAvg are REJECTED on two grounds — fixed ratios and arithmetic impossibility. ebitdaAvg is exactly 13.02% of revenueAvg and ebitAvg exactly 11.93% in every forward year FY2026 through FY2030. And netIncomeAvg EXCEEDS ebitAvg in every single forward year — FY2026 $900.4M against $519.1M, FY2027 $1,107.5M against $629.9M, FY2030 $1,704.7M against $988.5M. Net income cannot exceed earnings before interest and tax by 70%. All forward valuation uses epsAvg.

Peer context is unusable and the set is genuinely strange. peers returns ADSK, CRWV, FICO, MSTR, NXPI, STX, TEAM, WDAY, WDC and ZS — a design-software company, a GPU cloud, an analytics scorer, a bitcoin treasury vehicle, a semiconductor maker, two hard-disk-drive manufacturers, a collaboration-software company, a human-resources platform and a security vendor. Not one is an observability company: Dynatrace, Elastic, Splunk and Grafana are all absent. No peer-multiple comparison is drawn.

2a. What today's price assumes (the inversion)

At $288.15 — 119.0x FY2026 consensus adjusted EPS, 101.1x FY2027, 18.9x FY2027 revenue — the price embeds:

• Revenue reaches $4,352.3M in 2026 and $5,280.8M in 2027. (Consensus; 33 and 34 analysts.) The March quarter's 32.1% growth is well ahead of the 27.0% implied for the full year, so on revenue the consensus is beatable — and it has been beaten five quarters running.
• Adjusted EPS reaches $2.421 and $2.851. (Consensus; 24 and 25 analysts.) Note these grow slower than revenue in both years. The multiple therefore has to be justified by what happens AFTER FY2027, on six analysts' FY2028 estimates whose range is 118% wide.
• Stock-based compensation of 21.9% of revenue is a legitimate exclusion in perpetuity. (Our derivation.) This is the most fragile assumption in the price and it is not usually stated as an assumption at all. At $750.7M and rising — up 31.6% in FY2025 against revenue up 27.7% — it is growing faster than the business. If a reader treats it as a real cost, the FY2026 "earnings" fall by roughly $2.10 a share and the multiple ceases to be computable.
• The market keeps paying roughly 100-120x forward adjusted earnings. (Our number.) At 80x FY2027E the stock is $228; at 120x it is $342. The stock traded at $102.62 within the last twelve months, which is 36x the same FY2027 estimate.
• Nothing goes wrong on 2026-08-06. (Our number.) A print in two days, at a 52-week high, at 119x current-year earnings, with the consensus target 8% below the price.

2b. The return bridge (why the multiple moves)

Expected return over the next twelve months decomposes as: adjusted EPS growth (+17.8%, from FY2026E $2.421 to FY2027E $2.851) + multiple drift (COMPRESSION of roughly 26%, from today's 119.0x on the current year toward about 88x) + shareholder yield (0.0% — no dividend, and NO share repurchase in any year of the payload's cash-flow history) ≈ −13%.

Our base assumes substantial multiple compression, and the reason is that we cannot construct a defensible path in which it does not happen. Adjusted earnings growing 17-18% do not support 101-119x indefinitely; the consensus itself has FY2027 revenue growth decelerating to 21.3%; and the price has already run 8.0% past the consensus target and 2.8% past the median. A multiple that has to hold at 100x while earnings grow 18% is not a valuation, it is a bet on sentiment.

There is also no shareholder yield to cushion it. Datadog has never repurchased a share in the four years of cash-flow data in the payload, pays no dividend, and dilutes. Every dollar of return must come from earnings and the multiple, and the multiple is the larger term.

2c. Variant perception (where we differ, what would surprise)

• We are BELOW the street and the street is below the price — which is the most unusual configuration in this batch. Our base of $250 is 5.7% below the consensus target of $265, and the consensus target is itself 8.0% below the market price. Forty analysts rate this Buy and their collective target says the stock is 8% too expensive. That is not a disagreement between us and the market; it is a disagreement between the market and its own analysts, and we are on the analysts' side.
• Our specific divergence is the treatment of stock-based compensation, and it is a definitional disagreement rather than a forecasting one. $750.7 million, 21.9% of revenue, 75% of reported free cash flow, growing faster than revenue. Watchable number: stock-based compensation as a percentage of revenue in the 2026-08-06 print, against 21.9%. A decline toward 17-18% would materially change the quality of the earnings and would be the single most bullish disclosure available.
• We think the revenue acceleration is under-explained rather than over-stated. 32.1% growth at $4 billion of scale is genuinely impressive and the archive contains no net-retention figure, no customer count and no artificial-intelligence revenue disclosure to corroborate or contextualise it. Watchable disclosure: any net-retention number in the August print. The knowledge base's 2023 analysis put it above 130%; if it is still there, the acceleration is durable and the bear case weakens considerably.
• We think the governance change is being ignored. A redomiciliation from Delaware to Nevada carried by 62.5% of votes on a poll where Class B shares carry ten votes each, with 199.6 million votes against, followed two months later by the defeat of a simple-majority-voting proposal, is a coherent pattern. It changes nothing about the software and something about the recourse.
• Positive surprise that would force us to reverse: revenue growth above 35% in the August print with adjusted operating margin expanding and stock-based compensation falling below 20% of revenue; or the initiation of a share repurchase against the $2.94 billion of net cash, which the company has never done.
• Negative surprise that would confirm the verdict: revenue growth below 28%; soft third-quarter guidance; stock-based compensation above 23% of revenue; a return to GAAP operating loss; or simply the multiple normalising, which costs 20-30% with no operational change whatever.

Synthos fair values

All three anchors are multiples of the FY2027 consensus adjusted EPS distribution (mean $2.851, low $2.684, high $3.522, 25 analysts), cross-checked against FY2027 consensus revenue of $5,280.8M on a corrected enterprise value.

• Bear ~$150 — 52.6x the FY2027 consensus LOW of $2.684, and roughly 9.6x FY2027 revenue on enterprise value. Cross-check: 46.2% above the 52-week low of $102.62; 7.9% above the lowest street target of $139; 10.9% below the 200-day average of $168.46. The scenario: growth decelerates below 20%, the market stops excluding stock-based compensation, and a decelerating consumption-software business de-rates to a normal high-growth multiple. −47.9%.
• Base ~$250 — 87.7x the FY2027 consensus MEAN of $2.851, and roughly 16.3x FY2027 revenue on enterprise value. Cross-check: 5.7% below the street's $265; 10.7% below the median of $280; 1.0% above the 50-day average of $247.50. Sensitivity, stated openly: 80x gives $228 and 100x gives $285. The whole answer lives in an 80-100x band, which is itself an admission that no multiple here is anchored to anything. The scenario: revenue compounds at 21-27%, adjusted margins stay roughly flat, and the multiple compresses toward the high eighties as growth normalises. −13.2%.
• Bull ~$360 — 126.3x the FY2027 consensus HIGH of $3.522, and roughly 23.6x FY2027 revenue. Cross-check: 12.5% above the highest street target of $320; 24.9% above the 52-week high. The scenario: the agentic-software thesis in the knowledge base is right, consumption growth re-accelerates above 35%, stock-based compensation moderates, and the market pays a scarcity premium for the control plane of production software. +24.9%.

Base is 13.2% BELOW spot; asymmetry roughly 0.52:1 (47.9% down, 24.9% up), with zero shareholder yield. A fair value below the market price, a payoff ratio of roughly one-to-two against, a street target 8% below the price, a close at the 52-week high, and an earnings report in two days is an Avoid — and the section below names precisely what would reverse it and at what level.

3. Knowledge base — four bullish claims, two of them gradeable and pointing opposite ways

Raw hits: 9. Case-sensitive entity hits naming DDOG or Datadog: 4. Name-level claims: 4, ALL BULLISH. Discarded: 5.

Two from one 2023 episode — business_breakdowns, 2023-05-13, speaker Peter Offringa, speaker_role: independent, entity DDOG:

> bullish · conviction 68 · thesis

> "Datadog's land-and-expand model (30%+ annual customer adds, 130%+ net retention) plus rapid product cadence drives 50-60% revenue growth while profitable, far faster than Dynatrace, New Relic, Splunk."

> bullish · conviction 65 · thesis

> "Datadog was first to unify the three pillars of observability (infrastructure, APM, logs) into one data model and UI, and keeps out-developing point players and weak cloud-native tools."

GRADING the first, because it carries a number. It asserted 50-60% revenue growth. Actual growth since: 27.0% (FY2023), 26.1% (FY2024), 27.7% (FY2025), 32.1% (Q1 2026). The quantitative claim has been WRONG by roughly a factor of two, for three consecutive years. The structural half — land-and-expand with high net retention and rapid product cadence — is consistent with what the company has delivered and with the March-quarter re-acceleration, and it is the part of the claim that still stands. We report the miss rather than crediting the thesis, because a knowledge base that only surfaces the half that worked is not a knowledge base.

One from 2026 — jordi_visser, 2026-05-09, bullish 45, skill 2.0, speaker_role: independent, entities DDOG and FTNT:

> "Cyber/analytics names (Datadog, Fortinet) are separating from seat-based software (Adobe); software may outperform and semis underperform for a month or two."

GRADED: spectacularly right, and closed. Made 2026-05-09; Datadog has returned 96.4% over the following three months. But note what it actually was: a conviction-45 tactical observation with an explicit one-to-two-month horizon, made three months ago. The horizon has elapsed, the move has happened, and it is a closed trade rather than a live thesis. Its structural observation — that consumption-priced analytics separates from seat-priced software — is the most useful sentence in this lane and is the core of the exponential case in the front matter.

One naming Datadog inside a list — no_priors, 2025-07-10, bullish 75, entities Microsoft, Oracle, SAP, Salesforce, Workday, Datadog and Epic:

> "These companies stayed dominant precisely because of insular integrated-platform architecture (internal frameworks, platform teams, interoperable apps) that critics mock."

Corroborates the unified-platform argument from a governance-of-architecture direction rather than a product one.

The five discards are free-text "observability" hits about other companies — no_priors on Salesforce and agent benchmarking; dwarkesh twice on Microsoft and GitHub's Agent HQ and on agentic infrastructure; cognitive_revolution on Cisco's distributed-systems DNA; and no_priors on the enterprise software layer thickening. All are thematically supportive of the category and none names Datadog, so all are discarded from the conviction pool. Two are quoted in the front matter because they articulate the long-run case better than anything the company says: "The agentic world grows underlying infra — databases, storage, e-discovery, identity, observability remain the rails whether work is done by humans or agents" and "The more enterprise the application, the thicker the defensible software layer — observability, monitoring, insight extraction, testing/simulation QA — beyond just the model."

Attribution. Channel distribution: no_priors 3, dwarkesh 2, business_breakdowns 2, cognitive_revolution 1, jordi_visser 1. No channel reaches a third, so no concentration sensitivity test is triggered — though it is recorded that two of the four name-level claims come from one 2023 episode by one speaker, so the effective independent-source count is three. speaker_role: independent fired correctly on both named speakers. No claim carries a management speaker_role for this registrant.

Conclusion. Breadth 3, claim count 4, net conviction positive-low. The knowledge base likes the business, has been wrong about its growth rate by a factor of two, and was right about its stock for a two-month window that has closed. Nothing in it addresses the multiple.

4. Governance — Delaware to Nevada, and a ten-to-one vote

Two 8-Ks in this archive contain the entire governance story and neither is in the payload.

From the 8-K filed 2026-04-22 (Items 3.03, 5.03 and 5.07):

> "At the Effective Time, each outstanding share of Class A common stock, par value $0.00001 per share, of the Delaware Corporation... became one outstanding share of Class A common stock... of the Nevada Corporation, and each outstanding share of Class B common stock... became one outstanding share of Class B common stock... of the Nevada Corporation."

> "As described in the Proxy Statement, certain rights of the Company's stockholders were changed as a result of the Redomiciliation."

> "The Company's Class A stockholders and Class B stockholders, voting together as a single class, approved the proposal to redomicile the Company from the State of Delaware to the State of Nevada by conversion. The votes... were as follows: For 332,195,858 | Against 199,605,477 | Abstain 351,779."

And the crucial procedural detail, in the company's own words: "the Company's Class A stockholders were entitled to one vote per share and the Company's Class B stockholders were entitled to ten votes per share."

Three things follow.

The proposal passed with 37.5% of votes cast against it — 199.6 million against 332.2 million. That is an unusually large minority for a management proposal.

Those percentages are on a ten-to-one voting basis. Because Class B shares carry ten votes each and are held by insiders, the economic ownership opposing the move was proportionally far larger than 37.5%. The archive does not disclose the Class A/Class B split, so the economic percentage cannot be computed here and is not asserted.

The company itself states that "certain rights of the Company's stockholders were changed." It does not enumerate them in this filing, referring instead to the proxy statement, which is not in this archive. We record the fact and decline to characterise the change, because we have not read the document that describes it.

From the 8-K filed 2026-06-16 (Item 5.07), the 2026 annual meeting held 2026-06-15:

The pattern is coherent: a dual-class structure with ten-to-one voting, a redomiciliation to a state widely regarded as more management-friendly carried over substantial opposition, a defeated shareholder proposal to move to simple-majority voting, and 16.2% withheld votes against one director. None of this affects the software. All of it affects what a minority shareholder can do if something goes wrong, and it is unpriced at 101 times FY2027 earnings.

5. Data integrity — what we rejected and why

Six findings.

1. netDebt OMITS $4,073.5M of short-term investments — the company is NET CASH and the vendor says net debt. bal_a at 2025-12-31 reports cashAndCashEquivalents of $401,305 thousand and shortTermInvestments of $4,073,531 thousand, against totalDebt of $1,535,192 thousand — so the true position is $2,939.6M of NET CASH. The vendor reports netDebt of $1,133,887 thousand — net DEBT — by netting only the cash line. The error is $4.07 billion and it inverts the sign. It flows into enterpriseValueTTM of $103,428.6M, which should be approximately $99,630M — a 3.8% overstatement — and into evToSalesTTM of 28.17x (correct: ~27.1x) and evToEBITDATTM of 452.5x. This is the known short-term-investments omission defect, and here it is the largest instance in this batch.

2. The payload is a FULL QUARTER STALE and earnings arrive in two days. The most recent inc_q row is 2026-03-31; the most recent 10-Q in the archive is the one filed 2026-05-07 covering the same quarter; and second-quarter results are released 2026-08-06. Every operating figure in this dive is at least four months old, and this is stated at every use. The archive does contain a 10-Q filed 2025-11-07 covering the September 2025 quarter and a 10-K filed 2026-02-18, so the coverage is complete up to Q1 2026 and simply stops there.

3. est shows revenue DECLINING between FY2029 and FY2030 — incoherent forward rows, EXCLUDED. revenueAvg reads $8,304,128,274 for FY2029 and $8,287,299,999 for FY2030 — a decline — and ebitdaAvg likewise falls from $1,081.4M to $1,079.2M while epsAvg RISES from $3.98 to $4.69. A company cannot grow earnings per share 17.8% on falling revenue and falling EBITDA absent a buyback it has never performed. Both rows rest on 4 and 2 EPS analysts and are excluded entirely.

4. est.netIncomeAvg EXCEEDS est.ebitAvg in every forward year, and both EBITDA and EBIT carry the fixed-ratio fabrication signature — REJECTED. ebitdaAvg is exactly 13.02% of revenueAvg and ebitAvg exactly 11.93% in every year FY2026 through FY2030. netIncomeAvg is $900.4M against ebitAvg of $519.1M in FY2026 and $1,704.7M against $988.5M in FY2030. Net income cannot exceed EBIT by 70%. All forward valuation uses epsAvg.

5. seg_prod is EMPTY and seg_geo double-counts. seg_prod returns [] — no product, module or revenue-line breakdown of any kind for a company whose entire investment case is multi-product land-and-expand. seg_geo for FY2025 reports North America $2,433.1M and International $994.1M, which sum EXACTLY to the reported $3,427.2M of revenue — and then separately reports UNITED STATES $2,320.3M, which is a SUBSET of North America. Summing all three double-counts by 67.7%. On the correct basis the United States is 67.7% of revenue and North America 71.0%.

6. The trailing GAAP price-to-earnings ratio of 748.8x is arithmetically correct and analytically useless, and the reason should be stated rather than the number quoted. netIncomePerShareTTM of $0.384 against a price of $288.15 produces it. Datadog's GAAP earnings are suppressed by $750.7M of stock-based compensation and are therefore neither a good measure of economics nor a stable denominator: FY2024 GAAP EPS was $0.55, FY2025 was $0.31, and the direction has nothing to do with the business. This dive quotes both the GAAP and adjusted multiples at every point and treats neither as sufficient alone.

Where vendor and filing AGREED — recorded:

• inc_q Q1 2026 revenue of $1,006,426 thousand matches earn_cal's revenueActual of $1,006,426,000 exactly, as do all five reported quarters.
• cf_a reconciles internally in every year: FY2025 operating cash flow $1,050,135 thousand less capital expenditure $49,578 thousand equals free cash flow $1,000,557 thousand.
• seg_geo's two primary lines sum exactly to reported revenue.
• tech.max_dd_from_peak of 0.00% equals pct_from_hi of 0.00% — and it is CORRECT, because the stock closed AT its 52-week high. This is the one case in this batch where that field pairing is not a defect but a description. (quote.yearHigh of $289.90 is the intraday high; tech.hi52 of $288.15 is the closing high and equals the last price.)
• totalDebt of $1,535,192 thousand includes $256,187 thousand of lease obligations, which is a 20% inclusion of the T/TJX class — but it changes nothing here, because the company is net cash on either basis.

Vendor composite rating — accepted in direction. C+ / 2 overall, with 1/5 on both price-to-earnings and price-to-book, 2/5 on debt-to-equity and 3/5 on the returns and discounted cash flow. The two 1s are the correct reading and are the same two things this dive is most cautious about; the 2/5 debt-to-equity is wrong for the reason in finding 1, since the company is net cash.

Non-equity tripwire — checked and passed, with a note. DDOG is Class A common stock, par value $0.00001 per share, of a NEVADA corporation (redomiciled from Delaware in April 2026), NASDAQ Global Select Market. Note the separate Class B common stock carrying TEN votes per share, which is NOT this security and which is not disclosed anywhere in the payload. Price of $288.15 is not par-like; beta is 1.543; there is no dividend; volume was 4.44M shares (~$1.28B of turnover); the 52-week band of $102.62 to $288.15 is a 181% range. This is common equity, of a dual-class company.

6. Technicals and insiders

• Price $288.15. The stock closed AT its 52-week high — pct_from_hi is 0.00%. +180.8% above the 52-week low of $102.62. Position within the annual range: 100th percentile.
• +16.4% above a 50-day average of $247.50; +71.0% above a 200-day average of $168.46. RSI 65.6; MACD +8.20.
• Relative performance is the strongest in this batch by a very wide margin. 3-month +96.4% against SPY +7.6% and QQQ +7.7% — an 89-point outperformance in one quarter; 6-month +123.3% against SPY +11.1%; 12-month +112.5% against SPY +24.3% and QQQ +30.8%. The stock has roughly doubled twice over.
• Maximum drawdown from peak over the trailing year: 0.00% — correct, because the current price IS the peak.
• Sentiment: 40 buy, 7 hold, 1 sell across 48 analysts; consensus target $265 — 8.0% BELOW spot; median $280 — 2.8% below; high $320 (+11.1%); low $139 (−51.8%). A near-unanimous Buy rating attached to a target below the market price is the clearest possible statement that the price has outrun the thesis, and the 130% spread between the high and low targets says the analysts do not agree on what it is worth either.

Today's move. DDOG closed 2026-08-04 at $288.15, up 5.32% or $14.55 from $273.60. It opened at $269.06 — which was also the day's low — and traded up all day to close at $288.15, within 0.6% of the day high of $289.90, on 4.44 million shares. A 7.1% intraday advance from the open to the close, finishing at a 52-week high, two days before earnings. No company filing is dated 2026-08-04.

Insiders — one director, one day, $5.3 million sold

Every transaction in the file belongs to one director on one day. 40,000 shares were converted — Class B to Class A, the standard mechanic — and 19,900 were then sold across six tranches between $264.35 and $269.00, a single order working through the book, for approximately $5.29 million.

Two readings and we give both. The generous one: a director converting and selling roughly half of a converted block is routine diversification, it was executed five days after the March-quarter reporting window and well inside a normal 10b5-1 arrangement, and 20,100 shares were retained. The unforgiving one: the sale was executed at an average of roughly $266.60, which is 7.5% BELOW today's close, five days before an earnings report, by the only insider who transacted at all.

What the file does not contain: any transaction by chief executive Olivier Pomel or by the chief financial officer, and not one purchase by anybody.

7. Verdict, kill-criteria and flip conditions

Avoid.

This is a verdict on price and timing and we want that stated first, because the business deserves better than the tier implies. Revenue accelerated to 32.1% in the March quarter — faster than the 27.7% of FY2025 — on a 79.9% gross margin, with $1,000.6 million of free cash flow, capital expenditure of 1.4% of revenue, $2,939.6 million of net cash, and five consecutive earnings beats averaging 12.7%. The knowledge base's structural claims about the category are the best-articulated exponential argument in this batch. If this dive were about the enterprise, it would be a positive one.

It is about the security, and four things about the security are wrong at once.

First, the price closed at exactly its 52-week high after a 96.4% three-month advance. pct_from_hi is 0.00%. There is no drawdown, no pullback and no valuation cushion.

Second, the multiple. 119.0 times FY2026 consensus adjusted EPS, 101.1 times FY2027, 748.8 times trailing GAAP earnings, 25.5 times book — and the adjusted figure adds back $750.7 million of stock-based compensation, 21.9% of revenue, 75% of reported free cash flow, growing faster than the business, against a GAAP operating LOSS of $44.4 million. Consensus adjusted EPS grows 17.5% in FY2026 against revenue growth of 27.0%: the earnings the multiple is struck on are growing more slowly than the top line.

Third, the street's own target is below the price. Forty of forty-eight analysts rate this Buy, and their consensus target of $265 is 8.0% below the market, with the median of $280 also below. We are not taking a contrarian position against the sell side; we are agreeing with it and noting that the price has left it behind.

Fourth, the timing. Second-quarter results are released 2026-08-06, two days after this dive, into a 52-week-high close at 119 times current-year earnings. The reaction function there is asymmetric in a way no analysis improves.

Pre-registered REVERSAL conditions — what would take this to Watch or better. We expect to revisit this name and these are the terms:

• A price near $200, which is 70.2x the FY2027 consensus of $2.851 and roughly 13x FY2027 revenue on a corrected enterprise value. At that level the base case is +25% and the asymmetry turns positive without anything changing at the company. That is a 31% decline — smaller than this stock's three-month ADVANCE.
• Revenue growth above 35% in the 2026-08-06 print with adjusted operating margin EXPANDING, which would mean the acceleration is durable and the earnings can grow into the multiple rather than requiring it to hold.
• Stock-based compensation below 20% of revenue and falling, from 21.9%. This is the single most valuable disclosure available and it converts a large part of the "adjusted" earnings into real ones.
• The initiation of a share repurchase against $2,939.6 million of net cash. The company has never repurchased a share in the four years of cash-flow data available, and at a 0.98% free-cash-flow yield the case is admittedly weak — which is itself informative.
• A disclosed net-retention figure above 115%, which the archive does not currently contain and which the knowledge base's 2023 analysis put above 130%.

Pre-registered CONFIRMATION criteria — what would make this a durable Avoid rather than a timing call:

• Revenue growth below 28% in the August print, ending the acceleration narrative.
• Stock-based compensation above 23% of revenue.
• A return to GAAP operating loss after the March quarter's marginal $7.3 million profit.
• Soft third-quarter or full-year guidance, which at 119x is worth more than the quarter itself.
• Further insider selling by executive officers rather than directors.

Where DDOG fits in the Synthos Framework Portfolio. No position, and none contemplated above $200. It stays on the watch list at $200 (70x FY2027 consensus) for a 1.5% initial sleeve in the software/AI-infrastructure bucket. Sizing note: this is a 1.543-beta security with a 181% twelve-month price range that has doubled in a quarter. The discipline is not to short it — the business is accelerating and the category argument is strong — but to refuse to buy it at the high, two days before a print, at a multiple its own analysts will not endorse. Logged as a tracked Synthos call (Avoid) as of 2026-08-04 at $288.15, with the fair-value anchors, the reversal conditions and the $200 trigger all gradeable. We expect this call to be graded harshly if the August print is strong, and we make it anyway: a 13% negative expected return with 0.52-to-1 asymmetry is not something to own regardless of what happens on Thursday.

Single biggest risk to this verdict — stated because Avoid is the tier most likely to be wrong. Datadog is accelerating at scale, the observability category may genuinely be the control plane for a world of autonomous software, and momentum in names like this has historically run far beyond what multiples justify. If the August print shows 35% growth and expanding margins, this stock can trade at $360 without anything in the analysis above being false — the bull case in Section 2 is exactly that, and it is 24.9% above the current price. What we are saying is narrower and we hold it firmly: at a 52-week-high close, at 119 times current-year earnings whose definition excludes a fifth of revenue, with the consensus target 8% below the price and a report in forty-eight hours, the distribution of outcomes is unattractive and the correct action is to wait.

Provenance & disclosures

• Traceability: 4 name-level knowledge-base claims name Datadog out of 51,928 distilled claims, all BULLISH (raw hits 9, case-sensitive entity hits naming DDOG or Datadog 4, name-level 4, discarded 5; breadth 3, claim count 4, net conviction positive-low). TWO come from a single business_breakdowns episode dated 2023-05-13, speaker Peter Offringa, speaker_role: independent, entity DDOG — on land-and-expand with '30%+ annual customer adds, 130%+ net retention' driving '50-60% revenue growth' (bullish 68), and on Datadog being 'first to unify the three pillars of observability (infrastructure, APM, logs) into one data model and UI' (bullish 65). GRADED: the quantitative half is WRONG — revenue growth has run 27.0%, 26.1%, 27.7% and 32.1%, roughly half the 50-60% asserted, for three consecutive years — while the structural half stands. ONE is jordi_visser, 2026-05-09, bullish 45, skill 2.0, speaker_role: independent, entities DDOG and FTNT — 'Cyber/analytics names (Datadog, Fortinet) are separating from seat-based software (Adobe); software may outperform and semis underperform for a month or two.' GRADED: spectacularly right — Datadog has returned 96.4% in the three months since — but it was a conviction-45 call with an explicit one-to-two-month horizon that has now elapsed, so it is a closed trade rather than a live thesis. ONE is no_priors, 2025-07-10, bullish 75, naming Datadog among seven integrated-platform companies. The five discards are free-text 'observability' hits about Salesforce, Microsoft/GitHub, Cisco and the enterprise software layer generally — thematically supportive, none naming this company; two are quoted in the front matter because they state the long-run case better than the company does. Channel distribution: no_priors 3, dwarkesh 2, business_breakdowns 2, cognitive_revolution 1, jordi_visser 1 — no channel reaches a third, so no concentration test is triggered, though two of four name-level claims come from ONE episode by ONE speaker, making the effective independent-source count three. speaker_role fired correctly on both named speakers; no claim carries a management speaker_role for this registrant. All quotes are verbatim from the stored claim text.
• Data as-of: fundamentals through 2026-03-31 ONLY — the payload and the filing archive are both a full quarter stale, and second-quarter results are released 2026-08-06, two days after this dive · governance from the 8-Ks filed 2026-04-22 and 2026-06-16 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873601 = 2026-08-04T20:00:01Z ($288.15, +5.32%; 50-DMA $247.50; 200-DMA $168.46; RSI 65.6; MACD +8.20; distance from 52-week high 0.00%) · knowledge-base claims 2026-08-04. Datadog's fiscal year is the calendar year. All figures come from the Synthos vendor data file for DDOG or from the SEC filings in the DDOG archive; no figure comes from memory, recall or external retrieval.
• Filing archive contents: 10-K filed 2026-02-18 (fiscal 2025); 10-Q filed 2025-11-07 (September 2025 quarter); 10-Q filed 2026-05-07 (March 2026 quarter — the most recent financial statements available anywhere); 8-K filed 2026-04-22 (Items 3.03, 5.03 and 5.07 — the REDOMICILIATION from Delaware to Nevada by conversion, approved at a special meeting on 2026-04-21 by 332,195,858 votes to 199,605,477 with 351,779 abstentions, on a poll in which Class A shares carried one vote and Class B shares TEN votes each, with the company stating that "certain rights of the Company's stockholders were changed as a result of the Redomiciliation"); 8-K filed 2026-05-07 (Item 2.02, first-quarter results); 8-K filed 2026-06-16 (Item 5.07 — 2026 annual meeting: four directors elected with 9.9M, 57.7M, 83.4M and 1.0M votes withheld respectively; executive compensation approved; Deloitte & Touche LLP ratified; and a SHAREHOLDER PROPOSAL FOR SIMPLE-MAJORITY VOTING DEFEATED by 296,078,084 against to 217,163,770 for). All carry preserved [TABLE] statement data. The 8-K corporate-action scan was run and returned no acquisition or financing event; the material findings are governance.
• Where the filings contradicted or corrected the vendor (detailed in Section 5): netDebt of $1,133,887 thousand OMITS $4,073,531 thousand of short-term investments, inverting the sign — the company holds $2,939.6M of NET CASH against cash of $401.3M plus short-term investments of $4,073.5M and total debt of $1,535.2M — which overstates enterpriseValueTTM by roughly $3.8 billion (3.8%) and invalidates evToSalesTTM (28.17x against a corrected ~27.1x) and evToEBITDATTM (452.5x); the payload is a FULL QUARTER STALE, with the most recent inc_q row dated 2026-03-31 and earnings due 2026-08-06; est shows revenue DECLINING between FY2029 ($8,304.1M) and FY2030 ($8,287.3M) while epsAvg RISES from $3.98 to $4.69, which is incoherent absent a buyback the company has never performed — both rows excluded; est.netIncomeAvg EXCEEDS est.ebitAvg in every forward year ($900.4M against $519.1M in FY2026), alongside a fixed-ratio signature (exactly 13.02% and 11.93% of revenue in every year FY2026–FY2030); seg_prod is EMPTY for a company whose entire investment case is multi-product land-and-expand; and seg_geo reports UNITED STATES $2,320.3M as a subset of North America $2,433.1M, so a naive sum double-counts by 67.7%. Where vendor and filing AGREED — recorded: inc_q quarterly revenue matches earn_cal revenueActual exactly in all five reported quarters (Q1 2026: $1,006,426 thousand); cf_a reconciles internally in every year (FY2025: OCF $1,050,135K − capex $49,578K = FCF $1,000,557K); seg_geo's two primary lines sum exactly to reported revenue; and tech.max_dd_from_peak of 0.00%, which equals pct_from_hi, is CORRECT here rather than defective — the stock closed AT its 52-week high, the only instance in this batch where that field pairing describes reality rather than an error.
• Basis note: analyst estimates are on adjusted EPS, and the adjustment on this name is the largest in this batch. March-quarter GAAP EPS of $0.15 against an adjusted actual of $0.60 is a 4.0x adjustment; FY2025 GAAP EPS of $0.31 against a consensus adjusted $2.019 is 6.5x. The principal component is stock-based compensation of $750,671 thousand in FY2025 — 21.9% of revenue, 75% of the $1,000,557 thousand of reported free cash flow, and growing 31.6% year on year against revenue growth of 27.7%. GAAP operating income for FY2025 was a LOSS of $44,373 thousand and the trailing GAAP operating margin is MINUS 0.67%. Both bases are quoted at every point in this dive and neither is treated as sufficient alone: 119.0x and 101.1x on adjusted FY2026 and FY2027, 748.8x on trailing GAAP, and ~22.9x and ~18.9x on corrected forward EV/sales. Trailing adjusted EPS of $2.20 is the sum of the four most recent quarterly earn_cal actuals ($0.46 + $0.55 + $0.59 + $0.60). No GAAP-to-adjusted reconciliation is available in the extracted filings and the composition above is derived from the cash-flow statement's stock-based-compensation line.
• Estimate coverage: 33 and 34 analysts on FY2026 and FY2027 revenue and 24 and 25 on EPS — the deepest coverage in this batch — with an FY2027 EPS range of $2.684 to $3.522, a 29.4% spread; then 19 analysts on FY2028 revenue but only 6 on FY2028 EPS, whose range of $2.267 to $4.927 is 118% wide; and 4 and 2 on FY2029 and FY2030 EPS, both rows excluded as incoherent.
• Peer note: the vendor peer set — ADSK, CRWV, FICO, MSTR, NXPI, STX, TEAM, WDAY, WDC and ZS — contains a design-software company, a GPU cloud, an analytics scorer, a bitcoin treasury vehicle, a semiconductor maker, TWO hard-disk-drive manufacturers, a collaboration platform, a human-resources platform and a security vendor. NOT ONE is an observability company — Dynatrace, Elastic, Splunk and Grafana are all absent. No peer-multiple comparison is drawn.
• Fair-value caveat: the $150 / $250 / $360 anchors are multiples of the FY2027 consensus adjusted EPS distribution — 52.6x the low of $2.684, 87.7x the mean of $2.851, and 126.3x the high of $3.522 — each cross-checked against FY2027 consensus revenue of $5,280.8M on a corrected enterprise value (roughly 9.6x, 16.3x and 23.6x). Stated arithmetic, not a discounted cash flow. The base is sensitivity-disclosed: 80x gives $228 and 100x gives $285 — a band whose width is itself an admission that no multiple on this security is anchored to anything. The base case assumes substantial forward-multiple COMPRESSION, from 119.0x toward roughly 88x, because adjusted earnings growing 17-18% do not support 100-119x and consensus itself has FY2027 revenue growth decelerating to 21.3%. Our base of $250 sits 5.7% BELOW the street consensus of $265 — which is itself 8.0% BELOW the market price.
• Timing — the operative disclosure on this name: second-quarter 2026 results are released 2026-08-06, TWO DAYS after the date of this dive, against consensus adjusted EPS of $0.583 and revenue of $1,078,658,000 (+30.5% year on year). The stock closed on 2026-08-04 at $288.15, exactly its 52-week high, up 5.32% on the day and 96.4% over three months. The verdict is set with this timing as an explicit input and would not be the same tier at $200. The most recent insider filing covers 2026-07-30, when a director converted 40,000 shares and sold 19,900 between $264.35 and $269.00 — approximately $5.29 million, at an average roughly 7.5% below today's close. 2026-08-04 carried no company-specific filing.
• Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling. The earnings banner at the head of this document is marked by a blockquote and a heading, not by colour.
• Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
• Version: 2026-08-04-full.