On November 3, 2026, CrowdStrike argued that network-only security tools have a blind spot. Tools watching traffic, the company said, cannot stop an artificial intelligence (AI) agent acting on the endpoint, and "go blind" once the agent runs a local model. Seventeen days later, two people familiar with the matter told Reuters that CrowdStrike is in talks to buy Zscaler. Zscaler processes more than 750 billion transactions a day for about 11,000 customers and serves 14 of the 15 US Cabinet-level agencies.
The same day it made that argument, CrowdStrike launched an AI Kill Switch, which stops a rogue agent "first at the Gateway, and soon on the endpoint." The company describing network-only tools as blind shipped its newest control at the gateway first. Buying Zscaler would put that gateway, and the traffic of every customer behind it, under the same owner as the Falcon sensor on their laptops and servers.
The price would run past the US$33.5 billion Reuters said the market put on Zscaler before the report, more than six times the US$5.01 billion in cash CrowdStrike held on July 31, 2026. In return, CrowdStrike would add Zscaler's US$3.77 billion in annual recurring revenue to its own US$5.84 billion, a combined US$9.61 billion before overlap or attrition. That would land US$389 million short of the US$10 billion goal CrowdStrike moved into fiscal 2030 on September 2.
Neither company has confirmed the discussions, and the two people who spoke to Reuters say the talks could still collapse. Investors priced the trade within hours anyway. Zscaler shares rose 17 percent after the close on November 20, and CrowdStrike shares fell 8 percent.
If a deal closes, one company would watch a process launch on an endpoint, follow its connection across the wire, inspect its activity on the internet, and hold the power to stop the execution and the web-based traffic. Customers running Falcon and Zscaler today would hand the alarm, the gate, and the logs recording every step to a single owner.
The terms those customers should demand depend on whether CrowdStrike would be buying a missing control plane, a faster road to US$10 billion, or a scarce asset kept from a rival.
Disclosure. Scott Jarkoff ran CrowdStrike threat intelligence strategy across Asia-Pacific and Japan (APJ), and the Middle East, Turkey, and Africa (META) for five and a half years, leading the Counter Adversary Operations portfolio in those regions.
Operation Timeline
Trace the events, turning points, and connections behind the analysis.
CrowdStrike and Zscaler Timeline
Key events from the founding of Zscaler in September 2007 through every CrowdStrike and Zscaler acquisition to the reported talks in November 2026.
View Timeline Hide Timeline
-
September 2007
-
August 2011
-
February 21, 2012
-
June 2013
-
June 9, 2014
-
June 14, 2016
-
October 2017
-
March 16, 2018
-
June 12, 2019
-
February 19, 2020
-
April 16, 2020
-
May 22, 2020
-
September 23, 2020
-
December 23, 2020
-
February 18, 2021
-
March 29, 2021
-
April 15, 2021
-
June 1, 2021
-
October 28, 2021
-
November 1, 2021
-
December 27, 2021
-
June 17, 2022
-
August 1, 2022
-
September 20, 2022
-
February 8, 2023
-
February 2023
Elia Zaitsev Becomes Chief Technology Officer
United States
CrowdStrike announces Elia Zaitsev as global CTO at its 2023 sales kickoff in Las Vegas, after Sentonas moves to president.
-
February 20, 2023
-
September 19, 2023
-
March 5, 2024
-
March 8, 2024
-
April 12, 2024
-
May 14, 2024
-
June 24, 2024
-
July 19, 2024
-
July 20, 2024
-
November 20, 2024
-
May 6, 2025
-
August 1, 2025
-
August 30, 2025
-
September 12, 2025
-
September 26, 2025
-
October 31, 2025
-
November 21, 2025
-
January 13, 2026
-
January 14, 2026
-
February 3, 2026
-
February 5, 2026
-
February 20, 2026
-
March 18, 2026
-
May 27, 2026
-
June 3, 2026
-
July 16, 2026
-
August 20, 2026
-
August 26, 2026
-
September 1, 2026
-
September 2, 2026
-
September 2, 2026
-
September 3, 2026
-
September 8, 2026
-
November 3, 2026
-
November 3, 2026
-
November 3, 2026
-
November 20, 2026
Technology Already Connected
CrowdStrike and Zscaler have worked together for years. Zscaler said in 2023 the integration supported more than 900 joint customers. After Zscaler bought Red Canary, CrowdStrike Chief Business Officer Daniel Bernard described thousands and thousands of happy customers using their integrations. The September 2026 expansion blog put the joint solutions at thousands of enterprise customers worldwide.
Falcon and Zscaler already exchange consequential data. Under the partnership expansion announced September 2, logs from Zscaler Internet Access and Zscaler Private Access stream into Falcon Next-Gen SIEM, the CrowdStrike security information and event management product. Zscaler AI event logs follow them. Falcon shares user risk with the Zscaler Adaptive Access Engine in real time through the OpenID Shared Signals Framework. The partnership page adds automatically updated blocklists and a cross-platform quarantine workflow.
Falcon and Zscaler together create a useful sequence. Falcon can identify the process opening a connection, the identity behind it, and activity on the host. Zscaler can see the destination, the application, the data crossing the connection, and the policy governing access. One platform can find the malicious process. The other can deny the application or network path before the same identity reaches another device.
Ownership would make the sequence easier to package and sell. It could also let one engineering organization design detections around both data sets, align product roadmaps, and remove commercial negotiations from every integration change. CrowdStrike would gain the network enforcement plane the company does not own. Zscaler would gain endpoint and identity telemetry without depending on a partner whose priorities can change.
Technical integration does not require corporate ownership. A contract can already connect the products. An acquisition buys control of pricing and product sequencing. It also buys customer data rights, distribution, and the decision over which integrations receive the deepest support. Another owner could weaken the CrowdStrike connection or favor a competing endpoint platform.
Ownership would not erase the boundaries around customer data. Government contracts, privacy law, residency requirements, and the terms under which each customer supplied telemetry would still control what the combined company could join. Buying Zscaler does not automatically give CrowdStrike permission to pour every packet record into Falcon. A deal would make the architecture possible, not make every use lawful.
The same automation producing faster containment can amplify a mistake. A false Falcon signal could deny access through Zscaler. A bad Zscaler alert could trigger endpoint isolation. The AI Kill Switch already acts at the gateway before the endpoint. Joining both directions creates a control loop. It needs rate limits, staged enforcement, independent approval, and a rollback outside the system taking the action.
CrowdStrike Can Buy Years of Growth
CrowdStrike cannot pay for Zscaler from the cash it reported on July 31, 2026. Any transaction near the Zscaler market value would require CrowdStrike shares, new debt, or both, and a takeover premium would push the cost higher still.
Zscaler reported revenue of US$3.35 billion for fiscal 2026, which ended July 31, 2026, up 25 percent, and free cash flow of US$779.1 million. Red Canary, the managed detection and response company Zscaler bought for US$675 million, CRN reported, supplied part of that growth. Excluding Red Canary, revenue grew 20 percent. Annual recurring revenue also grew 25 percent, matching CrowdStrike, but the acquisition contributed US$141 million of it and organic growth was 20 percent.
The Zscaler slowdown shows in its fiscal 2027 guidance. Zscaler expects fiscal 2027 annual recurring revenue growth of about 16.6 to 17.4 percent, while CrowdStrike reported record net new annual recurring revenue of US$333 million in its quarter ended July 31, 2026, up 51 percent.
Zscaler free cash flow fell to US$60.8 million in its fourth quarter, from US$171.9 million a year earlier, as capital spending and internal-use software rose to US$218.5 million from US$78.7 million. Capital expenditure alone was US$200 million, 22 percent of quarterly revenue, Zscaler said on its earnings call. Zscaler still calls itself the leader in cloud security and is expanding into data security, artificial intelligence security, and security operations.
CrowdStrike entered the talks with the more valuable currency. Stocktwits put its market value near US$235 billion at the close of regular trading on November 20, roughly seven times Zscaler's US$31.77 billion. CrowdStrike shares had gained 96 percent during 2026, while Zscaler had fallen 13 percent. A stock-heavy offer could exchange the higher valuation for Zscaler's revenue and cash flow. The buyer can create financial value before one engineer combines a product.
Buying Zscaler's recurring revenue would also purchase time for CrowdStrike. At the September 2, 2026 briefing, CrowdStrike said it now expects US$20 billion in ending annual recurring revenue within fiscal 2035. The annual cash incentive plan in its 2026 proxy statement uses net new annual recurring revenue as one performance measure, and the board created a new long-term retention mechanism for George Kurtz. None of those incentives proves management wants a deal for compensation. They make scale an explicit corporate objective.
Both companies have built flexible consumption programs to sell more products into existing accounts. CrowdStrike said accounts using Falcon Flex exceeded US$2.29 billion in ending annual recurring revenue as of July 31, 2026, up 101 percent. Zscaler said Z-Flex generated more than US$1.7 billion in total contract value during fiscal 2026. A combined contract could let customers shift spending across endpoint and identity. The same agreement could cover cloud, secure access, and data protection without another procurement cycle.
The amount of new revenue available inside those accounts remains unknown. Neither company publishes the percentage of Zscaler customers also buying CrowdStrike, their shared annual recurring revenue, or the products already deployed in common accounts. The 900 joint customers Zscaler disclosed in 2023 and later references to thousands establish material overlap. They do not establish how much room remains to cross-sell.
Raw customer counts do not settle the question. The Zscaler annual report counts approximately 11,000 customers as of July 31, 2026. CrowdStrike counted more than 88,000 organizations as of January 31, 2026, including customers of managed security service providers. Those measures use different definitions and different dates. Dividing one into the other would produce a percentage without producing intelligence.
Zscaler would also change CrowdStrike's geographic mix. The Zscaler annual report says the company earned 47 percent of revenue from international customers in the fiscal year ended July 31, 2026. CrowdStrike earned 35 percent from international customers in the six months ended the same day. The target brings relationships, channels, and government access in markets where the buyer still has more room to grow.
Four Motives Fit the Talks
No public account identifies who opened the discussion or what each board wants. Four hypotheses fit the technical position, the financial record, and the competitive timing. They can operate together, but each predicts a different deal structure and a different explanation if the companies announce terms.
Hypothesis 1, Buying the Missing Control Plane
CrowdStrike built its position from the Falcon endpoint agent outward. It added cloud workload protection and identity. Threat intelligence, security information and event management, and data protection followed. Zscaler owns the inline layer connecting users and workloads to applications. The same layer covers branches and artificial intelligence agents. Reuters says the deal would address one of CrowdStrike's biggest gaps in network security.
Hypothesis 1 has the clearest technical support. A combined platform could trace activity from a process through an identity to an application and attempted data transfer. The platform could contain the host and close the path. Deal materials centered on one policy engine, shared telemetry, and coordinated response would strengthen it.
CrowdStrike's November 3 argument and the deal can be read as one strategy. CrowdStrike attacked network-only security, not the network, and a combined platform would hold the endpoint and the gateway together. The dependency is the sharper reading. The first CrowdStrike agent kill switch acts through gateways other companies operate. In the Zscaler fiscal 2026 results release, Chaudhry describes the company as connecting users, workloads, branches, and now agents directly to applications. Owning that gateway would turn a partner dependency into a CrowdStrike product.
The existing CrowdStrike and Zscaler partnership is the counterargument. Customers can already stream logs, share risk, and automate enforcement. CrowdStrike must show ownership creates more than a tighter bundle. Otherwise the company would be paying a public-company premium for capabilities available through an application programming interface and a commercial agreement.
Hypothesis 2, Keeping Zscaler From a Rival
Zscaler is one of the few independent security companies operating a global inline cloud at this scale. If Microsoft or Cisco owned the company, CrowdStrike could lose a neutral partner. The same would hold for Palo Alto Networks or Google. A rival would control the network data entering Falcon and could privilege its own endpoint product. That rival could also slow joint development or make the integration commercially unattractive.
Palo Alto Networks completed its CyberArk acquisition on February 11, 2026, placing identity beside network and cloud security. Cisco completed its roughly US$28 billion Splunk acquisition on March 18, 2024, adding security analytics and observability to its network estate. The scarce asset may matter as much for who cannot buy it as for what CrowdStrike can build with it.
The counterargument to Hypothesis 2 is cost. Defensive acquisitions can destroy value when the buyer pays to prevent a future possibility instead of solving a present customer problem. CrowdStrike could preserve access through contracts, deepen other secure access service edge partnerships, or build more of the network layer itself. A bidding contest would make the insurance still more expensive.
Hypothesis 3, Buying the Growth Target
The Hypothesis 3 arithmetic is difficult to ignore. The combined recurring revenue would land almost on the US$10 billion goal CrowdStrike now expects to reach within fiscal 2030. Zscaler would add profitable cash generation and expand the international business. A stock-heavy transaction would use the buyer's stronger valuation to acquire revenue the market values less, and Zscaler's organic slowdown helps explain the gap.
Hypothesis 3 does not require cynical intent. Public companies buy time, distribution, and future cash flows. CrowdStrike would be purchasing a customer base and a sales organization built over many years. The acquisition could also lower duplicate sales and corporate costs if customers accept one contract.
The danger is paying for the same dollar twice. Joint customers already contribute revenue to both companies. Product consolidation can cause them to remove one overlapping service instead of buying another. Nobody outside the negotiations has the account overlap or shared revenue data needed to calculate the organic growth hiding inside the headline total. Renewal and attach-rate data remain private too.
Hypothesis 4, Giving Zscaler a Chosen Successor
Jay Chaudhry founded AirDefense, CipherTrust, CoreHarbor, and SecureIT before Zscaler. He remains its chief executive and chair. The Zscaler annual report says he and his affiliates beneficially owned about 16.5 percent as of July 31, 2026, and executives, directors, and holders of 5 percent or more together held about 39.7 percent. A transaction cannot proceed as though the founder were an ordinary employee or a passive shareholder.
CrowdStrike and Zscaler also know each other well. Years of joint engineering, selling, and customer support reduce the information gap present in most acquisitions. One theory holds the leaders form a clique closing a deal among friends. The public biographies of Chaudhry, CrowdStrike founder George Kurtz, and CrowdStrike President Michael Sentonas establish no personal friendship or agreement to help one another. Between CrowdStrike and Zscaler, familiarity is the full extent of the evidence behind the clique theory.
CrowdStrike dealmaking does run through people who know each other. Elia Zaitsev, the CrowdStrike chief technology officer who left in August 2026, is launching the venture firm Cognition with Gur Talpaz and Tayler Sipperly, Axios reported. Both worked in CrowdStrike corporate development before joining the early-stage fund Brightmind, where they entered SGNL, the identity security startup, at US$105 million and exited at US$740 million. That exit was the CrowdStrike purchase of SGNL. Nothing in the record shows anything improper, and none of it involves Zscaler.
A founder choosing a trusted steward for the next phase remains plausible. Other explanations fit the present record better. Zscaler is growing, Chaudhry has announced no departure, and his ownership gives him reason to demand full value. Terms keeping him in a senior role would support succession. A contested process or rapid exit would weaken it.
Zscaler Alternatives Are Smaller
Kurtz has already argued against the kind of purchase CrowdStrike is now weighing. Asked about Palo Alto Networks buying CyberArk, he told Fortune in August 2025 he saw "a 20-year-old technology company with a lot of integration risk." He added, "When I was at McAfee, we acquired 21 companies, and never quite got them integrated." Zscaler was incorporated in September 2007, making it 19 years old.
A buyer wanting network control has three other places to look. Netskope runs the closest copy of the Zscaler cloud proxy model and reported US$899 million in annual recurring revenue on July 31, 2026. Cato Networks raised money at a US$4.8 billion valuation in June 2025 and reported an annual revenue run rate of US$250 million at the start of that year. Skyhigh Security, the former McAfee cloud business, belongs to the private equity firm Symphony Technology Group (STG).
None of those three matches Zscaler for scale or federal reach, and none solves the revenue problem. Adding Netskope to CrowdStrike's US$5.84 billion would reach about US$6.74 billion, barely above the US$6.6 billion the company already guides to on its own by January 2027. Adding Netskope, Cato, and Abnormal together would still leave CrowdStrike more than US$2.5 billion short of US$10 billion. Only Zscaler carries the combined total close to the fiscal 2030 goal at once.
A buyer wanting breadth would look at email, a layer Falcon does not cover. The CrowdStrike Falcon Fund invested in Abnormal Security on March 29, 2023, and connected the two platforms. The company renamed itself Abnormal AI on April 16, 2025. The fund joined the August 2024 Abnormal round at a US$5.1 billion valuation, when Abnormal had just passed US$200 million in recurring revenue. CrowdStrike has turned a fund stake into an acquisition before, buying the rest of Seraphic after the Falcon Funds held 9.4 percent.
A rival could also move on Zscaler itself. Google closed its Wiz purchase on March 11, 2026, and has shown it will pay for security at that scale. Cloudflare, a CrowdStrike partner since at least March 2022, already shows the network integration available by contract. If CrowdStrike walks away, Zscaler stays on the market for someone else.
Cynical Readings Have Numbers Behind Them
The most cynical reading begins with what CrowdStrike can receive on closing day. It would buy Zscaler's US$3.77 billion in recurring revenue and relationships across about 11,000 customers. It would also buy Zscaler's independence from every CrowdStrike rival. Management cannot build any of those three assets quickly. Product integration would follow. The revenue, customers, and competitive denial would arrive with the transaction.
That sequence can turn financial engineering into a platform story. CrowdStrike can use more highly valued shares to acquire revenue carrying a lower market valuation. The combined recurring revenue would land just short of the long-standing CrowdStrike revenue goal before any new product sale. CrowdStrike has carried that goal for years, most recently targeting 2031. On September 2, CrowdStrike said it would get there within fiscal 2030.
Executive incentives reward growth, and larger companies give their leaders more authority. None of those facts establishes self-dealing. They create a transaction whose financial achievement precedes its technical one.
CrowdStrike customers could pay for the same concentration twice. One contract may simplify procurement and lower the initial price. Replacing any component becomes harder once endpoint and identity sit inside the same commercial relationship as cloud and network policy. CrowdStrike could favor its own integrations or narrow support for competing products. It could also use bundle discounts independent vendors cannot match. The same control making the platform coherent can make the customer captive.
The Reuters report may also be part of the negotiation. Two unnamed people disclosed confidential talks while no agreement was guaranteed. A leak can test shareholder reaction or pressure a board. It can establish an expected premium, attract another bidder, or damage the talks. The public record identifies neither the source nor the objective. No one can yet tell which company benefits from the disclosure.
The rescue theory, the idea CrowdStrike would be bailing out a struggling Zscaler, overstates the slowdown. Organic growth of 20 percent and guidance near 17 percent give the theory some footing. The company still generated US$779.1 million in free cash flow, and thousands of customers already use the integration. CrowdStrike would be buying a healthy company with a product the buyer genuinely lacks. The counter-analysis changes what the buyer may value most. It does not erase the strategic fit.
Cyber Security Megadeals Rarely Stay Simple
The security industry has tried to create strategic advantage through adjacency before. Intel agreed on August 19, 2010, to pay about US$7.68 billion for McAfee, arguing hardware and security software belonged together. On September 7, 2016, Intel and TPG valued the business at US$4.2 billion while restoring McAfee as an independent company. The technical logic did not overcome the difficulty of making security central inside a chipmaker.
Symantec agreed on December 16, 2004, to merge with Veritas in an all-stock transaction valued at about US$13.5 billion and designed to span security and data availability. On October 9, 2014, Symantec announced it would split the businesses. The breadth intended to make the company indispensable instead made its strategy harder to execute.
McAfee ran the closest version of this experiment on January 3, 2018, when the endpoint vendor completed its purchase of Skyhigh Networks, a cloud access security broker. The purchase was meant to stretch McAfee from the device into the cloud. In March 2021, McAfee agreed to sell its entire enterprise unit to STG for US$4 billion so the company could focus on consumers alone, chief executive Peter Leav said.
STG merged the endpoint and detection products with FireEye into Trellix. In March 2022, STG spun the cloud products out as Skyhigh Security, a standalone company. Kurtz had left McAfee years before the Skyhigh purchase, but his account of McAfee integrations describes it.
Even adjacent cloud products can collide at the sales layer. Okta told investors its business suffered short-term execution challenges integrating the Auth0 and Okta sales organization. Product code is only one integration. Compensation plans and channel ownership can decide whether customers ever buy the combined architecture. Account control, renewal timing, and sales credit can do the same.
Cisco and Splunk offer the closer modern comparison, but the deal remains too young for a final judgment. Cisco bought telemetry, analytics, and security software to sit beside the network. CrowdStrike would buy the network enforcement layer beside its telemetry. The similarity strengthens the strategic case without proving the outcome.
Zscaler also owns Red Canary, which moves it into managed detection and response near CrowdStrike Falcon Complete. Both companies sell security operations, data protection, artificial intelligence security, and platform contracts. Those products can complement each other, compete for the same budget, or survive under separate brands. The acquisition announcement would need to say which.
Intel, Symantec, and McAfee each bought adjacency and later gave it up. Intel spun McAfee back out six years after agreeing to buy it. Symantec announced its split from Veritas nearly ten years after agreeing to the merger. McAfee sold Skyhigh with its enterprise unit about three years after closing that purchase. A Zscaler purchase would carry the same risk at more than four times the price Intel agreed to pay for McAfee.
Integration Can Lose the People It Buys
Public evidence does not show a morale collapse at CrowdStrike. A survey CrowdStrike released on November 12, 2026, found 97 percent of participating employees were proud to work there, and Fortune placed the company among its 25 best workplaces worldwide. The survey cannot reveal how employees view an acquisition reported eight days later.
CrowdStrike also eliminated about 500 positions, or 5 percent of its workforce, under a plan announced in May 2025 and completed by July 2026, its quarterly filing says. Full-time headcount still rose from 10,698 on January 31, 2026, to 11,706 on July 31, 2026. The company cut roles while continuing to hire, making the effect on broad sentiment impossible to infer from headcount alone.
Zscaler committed on September 1, 2026, to reduce its worldwide headcount by about 3 percent. A merger would place employees from both companies behind overlapping sales and marketing functions. Security operations, managed detection, support, and corporate teams overlap too. Employees have reason to expect some roles may disappear even if executives promise growth.
Both companies are shrinking inside an industry already shedding staff. In the ISC2 2025 Cybersecurity Workforce Study, 24 percent of respondents reported cyber security layoffs and 36 percent reported budget cuts. Cloudflare eliminated about 20 percent of its workforce, more than 1,100 people, in May 2026 to become what it called agentic AI-first. SentinelOne trimmed 8 percent the same month, and Rapid7 approved a reduction of about 12 percent in August.
The large platforms are shrinking too, and Cisco announced nearly 4,000 job losses in May 2026, under 5 percent of its workforce. Google Cloud layoffs reportedly reached the Google Threat Intelligence Group in June, though the company disclosed no number. The closest precedent came from the CyberArk deal. Palo Alto Networks notified hundreds of CyberArk employees of layoffs the day after closing its US$25 billion purchase.
That uncertainty can remove the people CrowdStrike needs most. Competitors can recruit the engineers who understand the joint architecture before integration begins. Salespeople can take customer relationships with them. Product leaders can leave rather than compete for one surviving role. Transaction terms may protect selected executives while leaving the employees carrying operational knowledge less certain. The acquisition can lose part of what it purchased before the products come together.
One Company Becomes a Larger Intelligence Target
CrowdStrike and Zscaler are already intelligence targets. CrowdStrike fired an insider in October 2025 after pictures of internal dashboards reached hackers, while Zscaler disclosed on August 30, 2025, that it lost commercial and support information through the Salesloft Drift supply chain incident. Chainalysis lost customer identity data because the people and organizations hunting threat actors are themselves valuable collection targets.
Nation state threat actors have pursued security vendors through product infrastructure, logistics providers, source code, and human access. Sophos documented five years of China-based adversaries working against its firewalls. SentinelOne found ShadowPad inside the company handling hardware logistics for its employees. F5 lost source code and vulnerability information to a nation state threat actor.
The Cybersecurity and Infrastructure Security Agency (CISA) answered on October 15, 2025, with Emergency Directive 26-01, an order federal civilian agencies must obey. CISA required them to inventory every F5 BIG-IP product, check whether its management interface was reachable from the internet, and apply the F5 updates.
A merger process creates a temporary target larger than either production environment. Bankers, lawyers, and consultants receive access to financial models and customer overlap. Board members and integration teams may see architecture, contracts, product weaknesses, and future roadmaps. Their accounts can exist for months and carry privileges designed for speed. A foreign intelligence service does not need to wait for the acquisition to close.
Employee uncertainty creates another opening. Threat actors can impersonate integration staff, recruiters, benefit providers, or executives while roles are changing. The highest-value target may be the person deciding how Falcon and Zscaler will exchange data. Physical recruitment inside OpenAI showed why access held by one employee can be worth more than a perimeter exploit. Customer deployment maps and detection gaps can be more useful than source code.
The threat to a combined CrowdStrike and Zscaler does not end with theft. A combined platform could let an adversary suppress endpoint detection and network enforcement from one trusted position. Fire Ant hid its tunnel by filtering what administrators could see through compromised infrastructure. FalconFlank turned a defensive feature into a privilege-escalation path. The most trusted security control remains the most useful one to turn.
Corporate ownership alone does not create one technical failure domain. Independent release pipelines and separate administrative planes can preserve barriers after the combination. Isolated signing systems and customer-controlled enforcement can do the same. IDC Frontier lost four physically separate availability zones together because dependencies crossed the boundaries customers could see. The danger appears when integration converts one bad signal, credential, or release into authority across both layers.
The 2024 outage remains the warning. On July 19, 2024, a CrowdStrike content update handed a sensor built to read 20 values a file carrying 21 input fields, and the out-of-bounds read crashed Windows. Microsoft estimated 8.5 million Windows devices went down, and Delta Air Lines canceled more than 7,000 flights. CrowdStrike added staged deployment rings and other safeguards after that failure.
Zscaler brings a separate engineering organization and a separate control plane. Preserving those separations would sacrifice some efficiency, but it would keep one content error from becoming both endpoint failure and network denial. The National Security Agency patching its classified network into an outage shows why even urgent security changes need an independent rollback.
Governments Inherit the Concentration
Zscaler says it serves 14 of the 15 United States Cabinet-level agencies. That statement does not mean it carries every agency connection or every packet. It establishes the reach of the customer relationship. A combined vendor could hold endpoint, identity, cloud, and network roles across the same government estate.
The United States would review a transaction of this size under the Hart-Scott-Rodino Act. The 2026 reporting threshold begins at US$133.9 million, far below any plausible price. Antitrust authorities could examine product competition, bundling, interoperability, and whether a combined company could disadvantage rival endpoint or secure access products.
The Committee on Foreign Investment in the United States reviews foreign investment for national security risk. CrowdStrike and Zscaler are both based in the United States, leaving the most obvious national security review mechanism outside its normal purpose. Antitrust review can consider competition. It is not designed primarily to decide whether one vendor should control two layers of government cyber defense.
China creates a different problem. Authorities told domestic companies in January 2026 to stop using products from more than a dozen US and Israeli cyber security vendors, including CrowdStrike, Reuters reported. CrowdStrike said it has no sales, offices, employees, or hosted infrastructure in China.
Zscaler still delivers China Premium Access to the mainland employees of multinational organizations. Zscaler built the service with a partner in China, routing customer traffic through data centers in Beijing or Shanghai. A CrowdStrike acquisition would change the owner of that service. It would not transfer direct ownership of the partner infrastructure.
Zscaler's mainland revenue and CrowdStrike's absence from the market may leave the transaction below China's normal filing thresholds. The State Administration for Market Regulation can still call in a deal below them, and local authorities retain leverage over the partner-delivered service. Pressure on China Premium Access is more plausible than a formal veto on the public evidence.
Allied governments face the same concentration question without China's political conflict. A buyer can demand sovereign hosting, independent encryption keys, separate administrative access, and a contractual right to export logs. Japan's Digital Agency runs its Government Cloud more than eighty percent on Amazon Web Services, and jark.me analysis of the sovereignty market in Japan found residency moves the servers, not the operator. Governments asking where control sits face that same gap with a combined CrowdStrike and Zscaler.
Control Is the Price
The technical fit between Falcon and Zscaler is direct, and a rival owning Zscaler could turn a valuable integration into a dependency. The existing partnership cuts the other way, because CrowdStrike would be paying for control rather than capability it can already rent. Weighing both, I assess with high confidence a principal strategic motive for CrowdStrike is ownership of the network control plane Zscaler now supplies through partnership.
Buying Zscaler would carry the combined annual recurring revenue near the fiscal 2030 goal before overlap, attrition, or accounting adjustments, and no alternative target comes close. CrowdStrike has the market value to pay in shares, an incentive plan built partly on net new recurring revenue, and a flexible-contract model able to absorb another product estate. On that record, I assess with moderate confidence financial acceleration is a coequal motive.
The CyberArk, Splunk, and Wiz transactions show large vendors buying adjacent control planes, leaving Zscaler one of the few independent platforms at its scale. That scarcity points to timing, and I assess with moderate confidence competitive denial influences it. Chaudhry's ownership and history make his preferences decisive, but nothing public shows he wants an exit, so I assess with low confidence founder succession drives the talks.
The first evidence able to separate those judgments will be the transaction itself. A stock-heavy structure would support the currency and growth explanation. Product commitments around one policy engine would support the control-plane explanation. Long-term interoperability guarantees would weaken competitive denial. A continuing operating role for Chaudhry would strengthen succession.
Customers running Falcon and Zscaler can act before the companies do. They can require independent release pipelines and separate signing systems. Each product needs staged enforcement and a tested rollback. Customers can keep security logs outside the company generating them and preserve another path into critical applications. Automation able to isolate thousands of users needs a circuit breaker. At Osaka Metropolitan University, one compromised virtualization layer reached roughly 500 servers at once, and encrypted backups turned recovery into a different problem.
The deal makes business sense because CrowdStrike would buy growth, distribution, and the one security control the company does not own. The risk comes from the same sentence. One company would watch the process launch on the endpoint, score the risk of the identity running it, inspect its traffic on the internet, and decide whether the execution and the connection continue. CrowdStrike would buy the distance between detection and denial.