AI adoption benefits cybersecurity and networking ecosystems, with open versus closed weights determining the beneficiaries

Overview

This report discusses the impact of two AI adoption paths—closed-weight models and open-weight models—on the cybersecurity, networking equipment, and optical communications software ecosystem. The core view is that AI adoption itself is positive for networking and security demand because it will bring more traffic, more AI agents, more identities, more data centers, and lower-cost network attacks; however, different model adoption paths will change the benefiting assets and industry value chains.

Core views

The report believes that open-weight model adoption is more favorable to the cybersecurity ecosystem and traditional enterprise suppliers because enterprises need to purchase more orchestration, security, and networking services separately for local, edge, or distributed environments. Closed-weight model adoption is more favorable to optical communications and networking companies with higher cloud exposure because compute and services are more concentrated in frontier model laboratories and cloud platforms. In areas such as optical communications that have parallel U.S.-China supply chains, the closed-weight path may reduce the complexity arising from the global proliferation of open models.

Analysis framework

The report uses a scenario-analysis framework that divides AI adoption into closed-weight and open-weight model paths and compares changes in compute location, service procurement methods, supply-chain complexity, and security needs under each path, thereby mapping the relative benefits across cybersecurity, enterprise networking, cloud networking, and optical communications.

Methodology notes

• Scenario analysisClosed-weight models vs. open-weight model adoption paths

  Projecting the distribution of AI infrastructure and security needs based on the degree of model-weight openness

  The closed-weight scenario assumes that more compute and services are centrally provided by frontier laboratories and cloud platforms; the open-weight scenario assumes that more compute, intelligence, and service procurement occur in enterprise, edge, or local deployment environments.

• Industry viewMorgan Stanley Industry View

  North America Industry View Attractive

  The report discloses an Attractive North America industry view, meaning analysts expect the industry coverage universe to outperform the relevant broad market benchmark over the next 12-18 months.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Cybersecurity ecosystem

  Broadly benefits from AI adoption

  Strengths

  More traffic, identities, data, and AI agents require protection; AI also lowers the cost of attacks, potentially driving security spending.

  Weaknesses

  In a closed-model ecosystem, built-in security services from AI-native vendors may create competitive pressure on traditional security vendors.

  Comparison

  Compared with the closed-weight scenario, the open-weight scenario is more favorable to traditional cybersecurity vendors.

  Risks

  If enterprise security capabilities are provided as built-in features by model laboratories or AI-native platforms, incremental demand for independent security vendors may fall below expectations.

• Optical communications and cloud networking companies (LITE, COHR, ANET)

  Beneficiaries under the closed-weight model scenario

  Strengths

  Under closed-weight models, compute and services are more likely to be concentrated in frontier laboratories and cloud platforms, driving demand for data centers, optical communications, and cloud networking.

  Weaknesses

  Highly dependent on cloud capital expenditures and the pace of centralized deployment by frontier laboratories.

  Comparison

  Compared with the open-weight scenario, the closed-weight scenario is more favorable to the optical communications value chain.

  Risks

  Parallel U.S.-China supply chains, the cloud capital expenditure cycle, and changes in model deployment paths may affect demand realization.

• Traditional enterprise networking vendors (CSCO, FFIV)

  Beneficiaries under the open-weight model scenario

  Strengths

  Open-weight models may drive more edge, local, and enterprise-side compute, requiring enterprises to purchase orchestration, networking, and service capabilities separately.

  Weaknesses

  If closed-weight models dominate, new enterprise-side deployment demand may be lower than concentrated cloud-side demand.

  Comparison

  Compared with the closed-weight scenario, the open-weight scenario is more favorable to enterprise networking vendors.

  Risks

  The pace of enterprise AI adoption, budget allocation, and cloud alternatives may affect revenue elasticity.

Key data

• Report date2026-08-03The report cover displays August 3, 2026 04:01 AM GMT.
• Coverage themeCybersecurity / Networking | North AmericaThe report focuses on the impact of AI adoption on cybersecurity, networking equipment, and related software in North America.
• Industry viewNorth America Industry View AttractiveThe report cover discloses an Attractive North America industry view.
• Beneficiary direction under the open-weight scenarioTraditional enterprise suppliers, such as CSCO and FFIVUnder open-weight models, more compute may occur at the edge, requiring enterprises to purchase more orchestration and services.
• Beneficiary direction under the closed-weight scenarioOptical communications and networking companies with high cloud exposure, such as LITE, COHR, and ANETUnder closed-weight models, more compute and services may be provided by frontier laboratories and cloud platforms.
• Rating framework time horizon12-18 monthsMorgan Stanley discloses that its equity ratings and industry views are generally based on relative performance over the next 12-18 months.

Impact & implications

The investment implication is that AI adoption should not be viewed solely as a uniformly positive theme; the degree of openness in the model ecosystem must be distinguished. If open-weight models proliferate, traditional enterprise networking, security, identity security, and orchestration service companies may show greater relative elasticity; if closed-weight models dominate, demand may be more concentrated in cloud data centers, the optical communications value chain, and cloud-related networking companies.

Risks

• In a closed-weight model ecosystem, AI-native vendors may provide their own cybersecurity services, reducing the share held by traditional security vendors.
• AI enhances attackers' capabilities and lowers attack costs; while this may drive security spending, it also increases enterprise risk exposure.
• Industries such as optical communications have parallel U.S.-China supply chains, and greater globalization of open-weight models may increase supply-chain complexity.
• Specific stock ratings and prices may change; the report advises reviewing the latest research on each company.

What to watch

• The actual proportion of enterprise adoption accounted for by open-weight versus closed-weight models.
• Whether AI workloads run centrally in the cloud or spread to edge, local, and enterprise environments.
• The sustainability of cloud capital expenditures, data-center construction, and optical communications demand.
• The competitive landscape between traditional cybersecurity vendors and AI-native platforms in security services.
• Subsequent orders, guidance, and rating changes for the companies mentioned, including CSCO, FFIV, LITE, COHR, and ANET.