CrowdStrike's Entire Guided EPS Range Fits Inside One Post-Split Cent, and Consensus Is Sitting On It. The Only Line Wednesday Can Surprise On Is $285 Million of Net New ARR.

CrowdStrike reports the fiscal second quarter of 2027 — the three months ended 31 July 2026 — after the US close on Wednesday 26 August 2026, with the call at 5:00 p.m. Eastern. The company announced the date by press release on 4 August; there is no 8-K behind it, which is how CrowdStrike has always done this.

Consensus is $0.29 of non-GAAP EPS on $1.44 billion of revenue, press-reported from third-party estimate feeds. Neither figure is news. CrowdStrike guided revenue to $1,436–1,442 million on 3 June, and the midpoint of that is $1,439 million, which is $1.44 billion. It guided non-GAAP EPS to $1.16–1.17, and consensus of $0.29 is $1.16 divided by four.

The division matters, because CrowdStrike split its shares four-for-one on 2 July 2026 and every per-share figure the company has guided was published a month before that. Run the guided band through the split and it becomes $0.2900 to $0.2925 — a range a quarter of a cent wide. Both ends round to $0.29. After the split, CrowdStrike's entire guided earnings range is a single number, and that number is the consensus.

So the two lines everyone quotes on Wednesday night carry almost no information. The line that does is not guided at all, and has to be backed out of the ARR guide: net new annual recurring revenue of about $285 million.

The points

• Net new ARR is derivable from guidance, and it is $285M. CrowdStrike guided ending ARR of $5,792.6–5,794.6 million for this quarter and disclosed ARR of $5,508.596 million as of 30 April 2026 in the 10-Q. The difference is $284.0–286.0 million. The subtraction is ours; the two figures are the company's.
• The year-ago quarter added $221.1 million. So the guide asks for +28.9% growth on the metric, against ARR growth of 24.4% and revenue growth of 23.1% on the same guide.
• CrowdStrike cleared this bar by 2.3% last time. The March guidance implied $250.0 million of net new ARR for the April quarter; the company delivered $255.8 million. Same arithmetic, one quarter earlier.
• The split has compressed the EPS bar to near-invisibility. In the last three quarters CrowdStrike beat the top of its own guided EPS range by 1 cent, 1 cent and 3 cents pre-split — 1.1%, 0.9% and 2.8%. Apply any of those to $1.17 and the post-split figure is $0.30. One cent is now the whole distance between meeting and beating.
• ARR is 108.1% of trailing revenue, our arithmetic on the disclosed $5,508.596 million against $5.09 billion of revenue in the four quarters through April. That gap is the reason the recognised-revenue score and the bookings score can point in different directions in the same quarter.
• Our trailing Rule of 40 is 51.1 — revenue growth of 23.2% plus a free cash flow margin of 28.0% on $1.42 billion of free cash flow. It is a strong score and it will not move on Wednesday, because none of its inputs are the thing the print turns on.
• Our forward model already uses the guide. Its build rate for the subscription line is $285 million of net new ARR a quarter, which is management's number for exactly this quarter.
• The CTO left on 20 August. Elia Zaitsev's departure after thirteen years was reported by Axios and stated by Zaitsev himself; CrowdStrike has not commented, has named no successor, and filed nothing. Details, and the limits of them, below.

What the guide already tells you about net new ARR

CrowdStrike does not guide net new ARR. It guides ending ARR, and it discloses ending ARR every quarter, so the build rate falls straight out of the subtraction:

Every ARR and net new ARR figure above is CrowdStrike's own, from its quarterly releases; the guided row is the company's outlook of 3 June 2026. The net new ARR implied by that guide, and the year-over-year percentage on it, are our subtraction.

This is what makes a preview possible on the metric that actually moves the stock. You do not need a Street estimate for net new ARR, because the company has effectively published one. $285 million is the bar. It is not a soft number and it is not ours.

And there is a track record on precisely this bar. Three months ago the March guidance of $5,501.8–5,503.8 million of ending ARR, set against the $5,252.8 million CrowdStrike had just reported, implied $250.0 million of net new ARR. The company delivered $255.8 million — $5.8 million, or 2.3%, above its own implied guide. Repeat that percentage on Wednesday and net new ARR prints at about $292 million.

The full-year guide sets the other half of the frame. CrowdStrike guided fiscal 2027 ending ARR of $6,531.7–6,555.5 million, which on the same subtraction is $1,278.9–1,302.7 million of net new ARR for the year against $1,010.9 million in fiscal 2026 — the 27.7% growth at the midpoint management said it was raising guidance to, up 520 basis points. Take the reported first quarter and the guided second out of it and the second half has to produce about $750 million, against $596 million a year earlier: +25.8%. That is the number Wednesday's guidance revision either confirms or does not.

What a beat has actually looked like here

All figures pre-split, which is the basis every one of those guides was set on. The pattern is a company that lands 0.8% to 1.8% above the midpoint of its revenue guide and one to three cents above the top of its EPS guide, quarter after quarter. That is not a company the Street can be badly wrong about on the guided lines.

Post-split, though, that same performance reads differently. A one-cent pre-split beat is a quarter of a cent after the split; a three-cent beat is three-quarters of a cent. Run the last three quarters' beat percentages through $1.17 and each of them lands at $0.30 — a one-cent beat on a $0.29 consensus, which is where the headline generators will stop. The arithmetic that produces a 1% beat and the arithmetic that produces "beats by a penny" are the same arithmetic, and only one of them sounds like news.

The measurement problem, in CrowdStrike's own terms

Our Rule of 40 for CrowdStrike is 51.1: revenue growth of 23.2% over the four quarters through 30 April 2026, plus a free cash flow margin of 28.0% on $1.42 billion of free cash flow against $5.09 billion of revenue. Our arithmetic, on reported figures.

It is a good score, on a quarter we wrote up in June as one of the company's strongest since the 2024 outage. It is also entirely backward-looking, because both of its inputs are recognised revenue. CrowdStrike's installed ARR at the same date was $5,508.596 million — 108.1% of that trailing revenue. A dollar of ARR signed in July is recognised over the following four quarters, so the score describes bookings that closed up to a year ago.

That is not a general complaint about the measure. It is specific to this company and this quarter, and here is how specific: at the guided build rate, CrowdStrike will add about $285 million of ARR in a quarter whose recognised revenue moves by roughly $53 million from the last one. Annualise the reported $255.8 million and it is $1.02 billion, 18.6% of the installed base. Almost none of it is in the trailing score yet.

Which means the Rule of 40 we publish for CrowdStrike will barely move on Wednesday no matter what happens, and it will keep barely moving for three quarters after that. A quarter that missed the guided build by $65 million would change recognised revenue in that quarter by a few million dollars, and the trailing score by a fraction of a point. The score is not wrong. It is late, by about the length of a revenue recognition schedule.

The CTO who left, and what can actually be said about it

Elia Zaitsev, CrowdStrike's Chief Technology Officer, left the company on 20 August 2026 after more than thirteen years, having held the CTO title since February 2023. The departure was reported by Axios that day and stated by Zaitsev himself on LinkedIn. He is co-founding a venture fund named Cognition with Gur Talpaz and Tayler Sipperly, both formerly of CrowdStrike's corporate development team, to lead seed and Series A rounds in early-stage AI-security companies.

Four things about that paragraph are worth separating out, because the version circulating is tidier than the record supports:

• The $170 million is a target, not a close. The reporting describes a fund being raised to that size, not one that has raised it.
• CrowdStrike has said nothing. No press release, no blog post, no successor named as of this morning.
• There is no 8-K, and there did not need to be. A departure triggers an Item 5.02 filing when the executive is a named executive officer or a Section 16 officer. Zaitsev's name does not appear in CrowdStrike's filings, which is consistent with a technology leader who sat outside that group. The absence of a filing is not evidence either way.
• "Funding his own competitors" is an inference, not a fact. The fund's stated thesis is that agentic AI resets the security category. That thesis is implicitly unkind to incumbents, and CrowdStrike is the incumbent. But it is an early-stage fund writing seed and Series A cheques, and nothing in the reporting says it targets endpoint or cloud workload protection specifically. We are not going to convert a venture mandate into a competitive claim about the company whose earnings we are previewing.

What is left is real and worth one line on the call: the platform's technical leader of the last three years is gone, six days before the print, with no named successor. That is a question for management, not a number for the model.

What the print does to our model

We published a forward model for CrowdStrike yesterday, built on the April quarter. It drives the subscription line — 95% of revenue — off installed ARR treated as capacity: a base of $5,508.6 million, a build rate of $285 million a quarter compounding 3%, revenue at $250,000 per $1 million of ARR per quarter, which is definitional, and a utilisation of 94.8% of a quarter of ending ARR. That construction is ours. It is used because CrowdStrike has published no subscription customer count since fiscal 2024, so a customers-times-price model would have to invent its own denominator.

Note what the build rate is: $285 million is management's guide for this exact quarter. The model has already assumed Wednesday. A print at $285 million changes nothing in it, and this is the useful part of the reconciliation — the quarter is not the event, the compounding rate is.

The four cases, all ours, against $190.34 a share, the price the model carried when it was published on 21 August:

Two findings sit in that table, and the bars above this article show where the money in each case comes from.

The first: the bull case is the price. $190.10 against $190.34 is a gap of 24 cents, a tenth of a percent. Turn the model around and solve for the exit multiple that makes fair value equal the current price, holding every other assumption fixed, and it is 23.7 times terminal-year revenue — against the 12x the base case uses and the 16x the bull case uses. Whatever else Wednesday does, it is being priced by a market that already assumes the guided pace of net new ARR holds without decaying for five years.

The second is stranger, and it is the reason to read the management-target row twice. CrowdStrike's own stated goal — $10 billion of ending ARR by fiscal 2031, from the September 2025 investor briefing — prices at $84.14 in our model, which is below the base case of $103.72. That is not a criticism of the target. It is arithmetic: the base case simply carries the current build rate forward with a modest deceleration, and doing that reaches about $12.7 billion of ARR by fiscal 2031. Hitting management's target does not require the run rate to improve. It requires it to slow down. The $20 billion by fiscal 2036 figure from the same briefing sits beyond the model's horizon and is not in any of these numbers.

So what would actually change the model on Wednesday? Not the quarter. The full-year ARR guide. The model's base case decelerates net new ARR growth to about 12.6% a year against the 27.7% guided for fiscal 2027; a raised full-year ARR range would say that deceleration starts later than assumed, and the fair value moves with the compounding rate, not with one quarter's build. A cut would do the reverse, and the bear case — which assumes net new ARR stops growing and starts shrinking, with the multiple rerating to 8x — is worth $48.26, a quarter of the current price. That is the size of the premium being carried.

One thing the model does not carry, and should be said out loud: share count is held flat at 1.03 billion, so roughly 2% a year of net dilution from stock compensation is charged nowhere in it. That flatters every case in the table.

What to watch

1. Net new ARR against $285 million. The guide implies $284.0–286.0 million. The year-ago quarter was $221.1 million. Last quarter came in 2.3% above the same implied bar, which on this one would be about $292 million. This is the number, and it is in the first bullet of the release.
2. The fiscal 2027 ending ARR range, currently $6,531.7–6,555.5 million. Raise it and the second half has to deliver more than the $750 million the current guide implies; leave it and the 27.7% full-year growth raised in June stands unchanged. This is the only line on Wednesday that speaks to the compounding rate our model turns on.
3. Non-GAAP EPS against $0.29 post-split, and whether the company restates the year-ago $0.93 to $0.23. Both figures will be quoted somewhere; only one basis is comparable. A print of $0.30 is a one-cent beat and a 1% beat at the same time.
4. Free cash flow against a 28.0% trailing margin. The April quarter ran 33.8%, an all-time high. The long-term target model management has stated is 34–38%, and a quarter inside that range would be the first evidence the target is a floor rather than an ambition.
5. Whether anyone asks about the CTO. No successor has been named. The question that matters is not the departure but the roadmap: who owns the Falcon platform's technical direction from here.

CrowdStrike reports the quarter ended 31 July 2026 after the US close on Wednesday 26 August, with the call at 5:00 p.m. Eastern; the company announced the date by press release on 4 August and filed no 8-K for it. Consensus of $0.29 and $1.44 billion is press-reported from third-party estimate feeds and is not a series this site stores or verifies. It is post-split, while every company guide quoted here predates the four-for-one split of 2 July 2026 and is therefore pre-split — each conversion between the two bases is ours. All guidance, for this quarter, for fiscal 2027 and in the earlier quarterly guides used to build the beat record, is CrowdStrike's own, from its releases of 27 August 2025, 2 December 2025, 3 March 2026 and 3 June 2026. Reported ARR, net new ARR, revenue and cash flow are the company's, from those releases and the 10-Q for the quarter ended 30 April 2026. Ours rather than the company's: every net new ARR figure implied by a guide, each of which is a disclosed ARR balance subtracted from a guided one and not a number CrowdStrike publishes; the comparisons built on them; the ratio of ARR to trailing revenue; and the trailing Rule of 40. The fair values, the build rate, utilisation, exit multiples, discount rate and flat share count are assumptions in our CrowdStrike model of 21 August 2026, not company forecasts, and the multiple of terminal-year revenue implied by the price is our solve on that model. The $10 billion and $20 billion ARR targets and the long-term margin model are management's own, stated at the September 2025 investor briefing and with the March 2026 results. The price of $190.34 is what the model page carried on 21 August 2026; a live quote will differ. Elia Zaitsev's departure and the Cognition fund are press-reported, originating with Axios on 20 August 2026 and Zaitsev's own statement; CrowdStrike has not commented, has named no successor and has filed nothing on the matter, and no figure in this article depends on any of it.