Netskope’s AI Security Push Drives Growth, but Profitability Still Lags

Netskope reached google news after reporting 29% revenue growth, wider adjusted margins, and early demand for its expanding AI security portfolio. The results beat management’s guidance, yet the company still recorded negative free cash flow and a substantial GAAP operating loss.

That contrast matters more than the upbeat headline. Netskope wants investors to see artificial intelligence as both a new security problem and a source of profitable platform expansion. Its quarter supplied evidence for that argument, but not enough to settle it.

The competitive pressure is already visible. Zscaler reported results one day later while promoting its own products for securing AI, agentic operations, and enterprise data. Netskope is therefore racing against an established rival, not waiting for a future market to form.

Netskope Beat Its Targets While Its Core Business Kept Growing

Netskope’s strongest argument is simple: growth accelerated while its adjusted operating loss narrowed sharply.

Netskope reported fiscal second-quarter 2027 revenue of $221 million for the period ending July 31, 2026. Revenue increased 29% from the same quarter one year earlier. The result exceeded the company’s guidance range of $213 million to $215 million.

Annual recurring revenue, or ARR, reached $899 million. ARR estimates the yearly value of active cloud subscription contracts, subject to the company’s renewal assumptions. The metric grew 27% year over year.

Remaining performance obligations grew 36%. That measure captures contracted revenue that Netskope has not yet recognized. It gives investors some visibility into future revenue, although timing and contract structure still affect when those obligations become reported sales.

The company also had 1,686 customers producing more than $100,000 in ARR. That customer group expanded 23% from the previous year. All geographic regions reportedly grew by at least 25%.

These figures support the central message behind the google news coverage. Netskope did not merely attach an AI narrative to a weak quarter. Its subscription platform continued attracting business while the company expanded into additional security categories.

Product adoption also deepened. Netskope said 59% of its customers used at least four products, up from 53% one year earlier. This matters because the company sells more than 25 products through a common platform.

A platform model can improve economics when customers add services without requiring entirely separate infrastructure, sales relationships, or administrative systems. It can also raise switching costs by placing more policies and workflows inside one vendor’s environment.

Net retention presented a less flattering signal. Dollar-based net retention was 114%, down from 118% one year earlier, according to Netskope’s quarterly filing. The metric includes expansion, contraction, and churn among existing subscription customers.

A rate above 100% means the existing customer base still produced more recurring revenue overall. However, the decline indicates that expansion within that base slowed relative to the prior year.

Netskope’s definition also deserves attention. The company assumes active cloud contracts expiring during the next 12 months will renew on existing terms. It may keep an expired contract in ARR while renewal discussions remain active.

That approach is disclosed and common enough to analyze, but ARR is not the same as recognized revenue. Investors should pair it with retention, remaining obligations, cash flow, and reported sales.

The quarter also demonstrated the durability of subscription revenue. Netskope said subscriptions generated about 99% of total revenue during the three and six months ending July 31.

These contracts typically span one to three years, with revenue recognized over the contract term. This accounting model can make reported revenue steadier than new bookings, even when customer purchasing changes.

For enterprise buyers, the immediate takeaway is narrower. Netskope is gaining enough adoption to keep funding product expansion, but customers should not interpret headline growth as independent proof of every product claim.

The important change was not one isolated earnings beat. Netskope combined faster reported growth, more large customers, and deeper product adoption with significantly better adjusted operating leverage.

Why the AI Security Push Is More Than a Google News Headline

Enterprise AI usage is creating a measurable data-control problem, giving security platforms a new stream of traffic, policies, and potential revenue.

Netskope frames AI security as an extension of work it already performs across cloud applications, networks, users, and sensitive data. The company is not starting with a separate appliance built only for generative AI.

Its pitch centers on Netskope One, a unified platform running through the NewEdge private cloud network. The platform applies policies to interactions among employees, applications, AI services, models, agents, devices, and corporate information.

That architecture fits how AI is entering companies. Employees use public chatbots, developers connect models through application programming interfaces, and teams deploy autonomous agents that can call business tools.

Each route creates a different control point. Browser sessions pass through familiar web domains, while automated systems send traffic to APIs. Agents can also invoke tools through the Model Context Protocol, commonly called MCP.

MCP is a standard for connecting AI systems with data sources and software tools. Those connections increase an agent’s usefulness, but they also give compromised or over-permissioned agents more places to retrieve or send information.

Netskope introduced several products for this environment. AI Command Center is designed to show where organizations use AI and which interactions carry risk. AI Gateway inspects traffic moving between users, applications, and models.

Agentic Broker focuses on communications involving AI agents and MCP. AI Guardrails targets risks including prompt injection and jailbreaking, where malicious instructions attempt to alter a model’s intended behavior.

AI Red Teaming tests privately hosted models for weaknesses before attackers find them. The Data Security Command Center then brings signals from posture management, identity, threat intelligence, and real-time controls into a shared interface.

These products turn the AI story into a broader data-security proposition. A company might start by discovering unauthorized chatbot use, then purchase controls for sanctioned models, agents, APIs, and internal data.

Netskope’s own research illustrates the demand, although readers should remember that vendor-produced threat data also supports the vendor’s market thesis. Its 2026 threat report found that generative AI users tripled in the average organization.

Monthly prompt volume increased sixfold, from 3,000 to 18,000. The top quarter of observed organizations sent more than 70,000 prompts per month, while the top 1% exceeded 1.4 million.

The same research found that 47% of generative AI users accessed personal AI applications. These unmanaged accounts create “shadow AI,” meaning AI activity that sits outside an organization’s approved controls and visibility.

Netskope observed an average of 223 generative AI data-policy violations per organization each month. It also reported that 50% of organizations lacked enforceable data-protection policies for generative AI applications.

Those figures describe a recognizable workplace scenario. An employee might upload source code, a sales transcript, or a customer spreadsheet to obtain a quick summary. The action feels routine, but it can move confidential data into an unapproved service.

Agents raise the stakes because they can repeat actions without a person reviewing each transaction. An agent with access to email, cloud storage, and customer systems can expose more information than a single mistaken prompt.

The company’s later AI usage research found that median weekly AI adoption rose from 34% to 59% of enterprise users. Average weekly prompt volume climbed from 1,498 to 4,731.

That activity creates more billable inspection events for security vendors. It also makes governance harder because organizations must evaluate the user, model, requested data, destination, and intended action.

Netskope says it has developed more than 190 domain-specific models and holds over 50 related patents. These are company claims, not independent measures of product accuracy or defensive effectiveness.

Still, the mechanism behind the strategy is credible. More AI interactions generate more network traffic and more decisions about identity, content, permissions, and data movement.

The open question is which vendor will control those decisions. Netskope has cloud-security infrastructure, but rivals already sell adjacent access, data, endpoint, and network protections to many of the same enterprises.

Better Adjusted Margins Do Not Equal Profitability

The margin improvement is real, but the gap between adjusted performance and reported losses remains too large to ignore.

Netskope’s non-GAAP gross margin reached 77%, up from 75% in the prior-year quarter. Non-GAAP operating margin improved from negative 20% to negative 9%.

That 11-percentage-point change supplied the strongest evidence for management’s scale argument. Revenue grew while the adjusted operating loss fell from $34 million to $19.3 million.

The company also outperformed its adjusted operating-margin guidance, which had called for a loss between 14% and 15% of revenue. Non-GAAP loss per share was three cents, better than the expected six to seven cents.

Non-GAAP measures exclude expenses that management believes obscure ongoing operating trends. Netskope excludes stock-based compensation, related taxes, acquired-intangible amortization, and restructuring expenses from non-GAAP operating results.

Those adjustments help compare recurring operating behavior across periods. They do not make the excluded costs disappear, and they should not replace the generally accepted accounting principles figures.

On a GAAP basis, Netskope lost $89.8 million from operations. That was nearly twice the $46 million GAAP operating loss reported one year earlier.

GAAP operating margin worsened from negative 27% to negative 41%. Stock-based compensation was an important source of the difference between the adjusted and reported pictures.

This produces the quarter’s core tension. Netskope showed better economics after adjustments while moving further from GAAP operating profitability during the same period.

Cash flow adds another constraint. The company used $16.5 million in operating cash, compared with $16.9 million one year earlier. Its operating cash-flow margin improved from negative 10% to negative 7%.

Free cash flow moved in the opposite direction. Netskope reported negative free cash flow of $29.8 million, compared with negative $19.7 million in the prior-year quarter.

Free cash flow subtracts property, equipment, and capitalized internal software purchases from operating cash flow. Netskope’s free-cash-flow margin consequently weakened from negative 12% to negative 14%.

The company expects capital spending to equal roughly 4% to 5% of revenue for the fiscal year. That spending supports NewEdge, the private network underlying Netskope’s cloud-delivered services.

Management argues that earlier network investment is now lowering the cost of serving additional customers. The higher adjusted gross margin offers some support, but continued capital requirements complicate the story.

Netskope also promotes AI Fast Path, which selects optimized network routes to AI services. It monitors latency, packet loss, jitter, token delivery, and other measures affecting model response times.

Company testing found that AI Fast Path reduced latency by as much as 90% to selected AI destinations. The wording “as much as” signals a best observed result, not a universal customer outcome.

Actual improvement will depend on a user’s location, original route, application, model provider, traffic pattern, and existing network design. Buyers should therefore test performance under their own workloads.

There is still a logical connection between routing efficiency and margins. Direct connections and route optimization can reduce transit costs while improving user experience. A shared global network can also spread fixed costs across more products and customers.

However, management has not isolated how much of the gross-margin gain came from AI Fast Path. Scale, hosting contracts, traffic mix, product mix, and earlier infrastructure investment can all affect service costs.

The quarter does not establish that AI products caused the margin improvement. It shows that Netskope expanded adjusted margins while increasing AI development and positioning its network for more AI traffic.

This distinction matters because the Investing.com headline pairs AI security with margin gains. The available figures support correlation and strategic alignment, but not a clean causal calculation.

Investors should also separate early demand from material revenue. Netskope said AI security products had generated interest, proofs of concept, and early customer activity. It did not disclose AI-specific ARR for the quarter.

Without that figure, readers cannot determine how much AI security contributed to the $899 million total. The products might already support meaningful deals, or they might remain a small component of broader platform contracts.

The company’s planned improvement still requires execution. Full-year guidance targets a non-GAAP operating margin near negative 9% and a positive free-cash-flow margin around 2%.

That forecast would mark progress, but it still relies on adjustments and future performance. One strong quarter cannot establish a durable path to GAAP profitability.

Zscaler Makes Netskope’s AI Security Race Harder

Netskope is pursuing a valid market, but Zscaler is selling a similar consolidation story from a much larger recurring-revenue base.

Zscaler reported fiscal fourth-quarter 2026 ARR of $3.771 billion, up 25% year over year. That total was more than four times Netskope’s $899 million ARR.

The comparison is imperfect because the companies reported different fiscal periods and occupy overlapping, not identical, product positions. However, both compete across zero-trust access, security service edge, data protection, and AI security.

Zscaler also described artificial intelligence as a major growth opportunity. Its portfolio includes security for AI, agentic security operations, data security, and controls for AI agents and models.

According to Zscaler’s fiscal results, its fourth-quarter revenue and ARR each increased 25%. The company also emphasized expanding adoption across its broader platform.

That overlap undermines any assumption that Netskope owns a unique AI-security category. Enterprise buyers can evaluate AI controls as additions to vendors they already use for traffic inspection, access, data-loss prevention, or cloud security.

Palo Alto Networks and other large security companies create further competitive pressure. They can bundle new AI defenses into existing security operations, firewall, cloud, and endpoint relationships.

The primary contest is therefore Netskope against Zscaler’s scale and installed base. The wider market provides supporting context, but Zscaler offers the clearest comparison because both companies center cloud-delivered zero-trust networking.

Netskope’s differentiation rests on the connection between its NewEdge network, granular data context, and one policy architecture. It argues that enterprises can inspect AI activity without forcing users through slow, inefficient routes.

Zscaler makes a parallel case around direct zero-trust connections among users, workloads, branches, applications, and agents. Both vendors claim their architecture avoids the weaknesses of traditional network perimeters.

This is why product counts and feature names will not decide the race alone. Buyers will examine policy depth, detection accuracy, latency, global coverage, integrations, administrative effort, and contractual consolidation.

Customer migration costs also matter. A large enterprise rarely replaces its access and data-security architecture simply because another vendor launches a specialized AI product.

Netskope can reduce that obstacle by expanding within its installed base. The rise to 59% of customers using at least four products suggests that cross-selling already contributes to growth.

The falling net-retention rate tempers that signal. Existing customers still expanded spending overall, but they did so more slowly than one year earlier.

Zscaler’s scale presents another challenge. A larger recurring-revenue base can fund research, acquisitions, distribution, and integrations while giving customers confidence about long-term support.

Netskope’s smaller base can provide room for faster percentage growth. It can also let the company focus its architecture on newer cloud and AI traffic patterns without supporting as many legacy assumptions.

Neither advantage guarantees a win. Security procurement depends on performance in real environments, relationships with channel partners, integrations with identity providers, and confidence in incident response.

AI agents add an important testing ground. When an agent invokes tools, the security platform must identify the agent, interpret the requested action, inspect relevant data, and enforce policy quickly.

False positives can interrupt automated work across an organization. False negatives can allow an agent to retrieve excessive data or execute unsafe operations. Vendors must balance control with usable response times.

Netskope says its Agentic Broker provides visibility and policy enforcement for MCP transactions. Customers will need evidence that those controls remain dependable across changing models, servers, tools, and authentication patterns.

The same scrutiny applies to AI Guardrails. Prompt injection defenses operate in a difficult environment because attackers can disguise instructions inside documents, web pages, tool responses, or retrieved content.

No security layer can safely treat every model output as predictable. Buyers need layered controls, limited permissions, transaction logs, human approval gates, and plans for compromised integrations.

That makes AI security a broader architecture decision, not a single-feature purchase. It also favors vendors able to connect identity, data classification, network context, and application behavior.

Netskope’s quarter suggests it belongs in that competition. It does not prove that its approach outperforms Zscaler or other alternatives across every enterprise deployment.

The market reaction and google news exposure may increase attention, but customer evidence will matter more. Large production deployments must validate the products beyond beta agreements and proofs of concept.

What Google News Readers Should Watch After Q2

Three signals will determine whether Netskope’s AI narrative becomes a durable financial advantage or remains an effective earnings presentation.

The first signal is disclosed adoption of the new AI security portfolio. Netskope should eventually provide a measurable indicator, such as AI-related ARR, paid production customers, expansion rates, or contract contribution.

Management currently describes early traction and customer interest. Those statements are useful, but they do not reveal how much revenue comes directly from Agentic Broker, AI Gateway, Guardrails, or related products.

Evidence of customers moving from proofs of concept into broad production deployments would strengthen the quarter’s thesis. Continued qualitative commentary without measurable adoption would weaken it.

Production use matters because pilot environments often have limited users, tightly controlled data, and lower transaction volumes. Enterprise deployment introduces geographic variation, complex identities, sensitive repositories, and business-critical workflows.

The second signal is the relationship among adjusted margins, GAAP losses, and free cash flow. Netskope forecasts full-year revenue between $888 million and $892 million.

It also expects an adjusted gross margin near 77%, an adjusted operating margin near negative 9%, and a free-cash-flow margin near positive 2%. Reaching those targets would show that the second-quarter improvement was not isolated.

For the third quarter, Netskope expects revenue between $227 million and $229 million. It projects an adjusted operating margin near negative 8% and an adjusted loss per share between three and four cents.

The company also expects third-quarter free cash flow between $10 million and $20 million. A move from negative $29.8 million to that range would provide a clearer test of near-term operating discipline.

Readers should still compare adjusted progress with GAAP results. A narrower adjusted loss matters less if stock compensation or other recurring exclusions keep reported losses unusually high.

Cash conversion deserves equal weight. Positive free cash flow would indicate that subscription growth and improving service economics can fund infrastructure and software investment.

Missing the cash target would not automatically invalidate the strategy. Contract billing schedules and capital spending can create quarterly volatility, but repeated shortfalls would challenge management’s scale narrative.

The third signal is competitive response, especially from Zscaler. Investors should watch AI-security ARR, major platform wins, agent-control releases, and product bundling from the larger rival.

If Zscaler expands its AI portfolio while maintaining strong growth, Netskope will face more pressure to differentiate its network and data controls. Comparable products could also intensify discounting during large enterprise negotiations.

Netskope’s annual retention trend will show whether consolidation offsets that pressure. A stabilization or improvement from 114% would suggest that customers are expanding into its new products.

A further decline would imply that strong new-customer growth is masking slower expansion within existing accounts. That would weaken the claim that AI products are strengthening the platform’s cross-selling engine.

The broader adoption environment remains favorable. Employees are already using public models, developers are connecting APIs, and companies are deploying agents with access to internal systems.

Security teams cannot address every use case with a simple block list. They need visibility into data, identity, destination, model behavior, and tool permissions.

This creates a genuine opportunity for Netskope. It also creates an attractive market for every major cloud-security vendor, making competitive execution more important than total-addressable-market estimates.

Netskope’s investor presentation projected a $336 billion total market by 2029, including $170 billion attributed to AI security. Those are company estimates assembled from multiple market categories, not guaranteed future revenue pools.

Large market estimates can also contain overlapping products and spending. They should describe strategic ambition, not replace evidence about customer budgets or competitive wins.

The most defensible conclusion from Q2 is narrower. Netskope grew faster, beat guidance, expanded adjusted margins, and presented a coherent reason for AI traffic to increase platform demand.

At the same time, free cash flow remained negative, GAAP operating losses widened, and net retention declined. The company has not disclosed enough AI-specific revenue to separate product traction from the wider platform’s momentum.

That balance is more informative than either an enthusiastic earnings headline or a reflexively skeptical response. Netskope’s AI strategy addresses a real problem, and its quarter showed meaningful operational progress.

Now the company must turn early AI interest into reported production adoption. It must also convert adjusted leverage into sustained cash generation without letting its core customer expansion weaken.

For readers arriving through google news, the next earnings report should answer three direct questions. Are AI pilots becoming paid deployments, is free cash flow turning positive, and is retention stabilizing against Zscaler’s pressure?

Those signals will reveal whether Netskope is building a profitable AI-security layer or simply narrating its existing cloud business in new terms. Enterprise buyers should ask the same questions before placing more critical workflows behind one platform.