Edge AI inference is moving beyond hyperscale clouds, and Cloudflare is becoming an early leader

Overview

This report is based on Morgan Stanley's first survey of edge AI partners/resellers and evaluates the trend of AI inference expanding from hyperscale cloud providers such as AWS and Azure to specialized edge infrastructure platforms. The report argues that training and large-scale model development will remain centered in centralized clouds, but enterprises are increasingly focused on latency, cost, privacy, and operational resilience in inference deployment, so edge AI will form a hybrid architecture with central clouds.

Core views

The core view is that edge AI is still in its early stages, but enterprise adoption will accelerate over the next few years; Cloudflare is the public company best positioned to benefit from edge AI, gaining share through its global network, Workers AI, Workers, R2, developer ecosystem, and performance/latency advantages; Akamai is becoming a credible second platform in distributed inference and AI security; Vercel benefits from AI-native application development and front-end AI deployment; hyperscale clouds will still dominate centralized AI workloads, but some incremental inference deployments are shifting toward specialized edge platforms.

Analysis framework

The report combines partner/reseller surveys, feedback on company product lines, share change indicators, use-case distribution, developer platform sales sources, competitive differentiation, and a long-term valuation framework. The investment view on Cloudflare also incorporates Analyst Day disclosures on Developer Platform ARR growth, AI agent request growth, global PoP presence, and long-term revenue/margin targets.

Methodology notes

• Channel researchPartner/reseller edge AI survey

  Measure enterprise edge AI adoption, vendor share changes, and product demand intensity through feedback from about 40 partners/resellers.

  This method is suitable for capturing procurement intentions, channel momentum, and competitive ranking in an early-stage market, but conclusions are still affected by sample coverage, customer mix, and short-term feedback volatility.

• Competitive landscape analysisPlatform share and differentiation assessment

  Compare the relative positioning of Cloudflare, Akamai, Vercel, Fastly, and hyperscale clouds in edge inference, developer experience, performance/latency, and security capabilities.

  The report focuses on identifying who is gaining incremental inference workloads, rather than judging whether the training market will be replaced by edge platforms.

• Valuation methodologyDiscounted EV/FCF multiple

  Cloudflare's target price is based on CY2034 FCF per share forecasts, an EV/FCF multiple, and an 11% discount rate.

  The base-case target price of $322 corresponds to 51x EV/CY34 FCF/shr of $12.13 discounted at 11%; the bull-case is $428, and the bear-case is $185.

Asset mapping & comparison

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

• Cloudflare Inc (NET.N)

  Core beneficiary

  Strengths

  Leading share gains in edge inference, with 63% of respondents selecting it as the leading edge compute/edge inference platform; strong growth in Workers AI, Workers, and R2; global network, performance/latency, and an integrated developer platform create differentiation.

  Weaknesses

  Enterprise edge AI is still in an early stage, the developer platform's revenue contribution still has a long ramp ahead, and recent share price performance has already widened its valuation premium versus peers.

  Comparison

  Relative to Akamai and Vercel, Cloudflare is viewed by the report as the clearest edge AI leader among public companies; relative to AWS and Azure, it is not the primary central training platform, but it has stronger edge positioning in incremental distributed inference.

  Risks

  Macro environment, pace of AI adoption, AI disruption to software, intensifying competition, weaker-than-expected expansion into large enterprise customers, and greater-than-expected difficulty in long-term margin expansion.

• Akamai Technologies, Inc. (AKAM.O)

  Credible alternative beneficiary

  Strengths

  Recognized by partners for distributed inference, AI security, low-latency distributed workloads, and application/API/bot/DDoS security capabilities; quarterly performance of the Inference Cloud/AI infrastructure product line has been strong.

  Weaknesses

  Customer adoption is at an earlier stage than Cloudflare, with current usage still more concentrated in cost optimization, cloud repatriation, and development/testing environments, while production AI inference remains a smaller share.

  Comparison

  The report positions Akamai as a credible second platform behind Cloudflare, rather than the clear current leader.

  Risks

  Security/governance/compliance, cost competitiveness versus hyperscale clouds, and whether AI ecosystem catalysts can translate into meaningful workload wins.

• Vercel

  Beneficiary of AI-native application development

  Strengths

  Benefits from AI-native application development, developer productivity, and front-end AI deployment, with its AI SDK expanding among enterprise customers.

  Weaknesses

  The report discusses its growth drivers mainly from channel perceptions and does not provide a public-market rating or target price.

  Comparison

  Relative to Cloudflare and Akamai, Vercel is more focused on the application development and front-end deployment layer, rather than being a broad edge infrastructure platform.

  Risks

  Enterprise AI application development pace, competition in developer platforms, and monetization capability.

Key data

• Cloudflare target price revisionraised from $305 to $322The report reiterates the Overweight rating, and the target price increase reflects higher confidence in edge AI share gains and developer platform momentum.
• Current AI inference platform usageAWS 68%, Microsoft Azure 37%, Cloudflare 29%, Google Cloud 21%Cloudflare ranks behind the hyperscale clouds and ahead of Google Cloud.
• Edge inference share growthCloudflare 55%, Akamai 34%, Vercel 24%, Fastly 13%Respondents believe Cloudflare had the largest share gain in edge inference over the past three months.
• Leading edge compute/edge inference platformCloudflare 63%, Akamai 26%, Vercel 24%, Fastly 13%Cloudflare is clearly ahead in edge execution capabilities.
• Edge AI adoption timeline31% believe meaningful expansion begins in 2026 or earlier, 52% expect meaningful expansion in 2027This shows enterprise deployment has started but remains in an early stage.
• Main adoption scenariosEdge AI inference 65%, edge compute/serverless applications 61%, AI agent workflows 57%Inference is the core near-term enterprise workload opportunity in edge AI.
• Cloudflare developer product growthWorkers AI 65%, Workers 43%, R2 22%Partners view Workers AI as Cloudflare's fastest-growing developer product.
• Cloudflare win factorsPerformance/latency 74%, AI inference capability 61%Performance, latency, and an integrated developer platform are the main competitive advantages.
• Cloudflare workloads migrating from hyperscale clouds52% of respondents estimate 10-30%, 14% estimate 30-50%This indicates Cloudflare is not only winning new AI deployments, but also taking on some migration of existing cloud workloads.
• Akamai demand status40% say early but building, 27% say healthy and improving, 13% say weak demandAkamai is receiving positive but earlier-stage feedback in distributed inference.

Impact & implications

The investment implication is that the next phase of AI infrastructure spending may not be concentrated only in training and central clouds; latency, cost, and data locality requirements for inference deployment will create incremental opportunities for edge platforms. If Cloudflare can continue monetizing Workers AI, Workers, and R2, and convert migrated hyperscale cloud workloads into paid revenue, it could support higher revenue growth and a valuation premium; Akamai looks more like a solid alternative beneficiary under distributed inference and security demand.

Risks

• Enterprise edge AI adoption may scale more slowly than the expected 2027 ramp.
• Hyperscale cloud providers may regain incremental inference workloads through pricing, product integration, and their own edge capabilities.
• After Cloudflare's valuation premium has widened, it is more sensitive to delivering on growth and margins.
• AI disruption to software business models could affect demand for Cloudflare and its peers.
• Akamai needs to prove that its Inference Cloud can move from cost optimization and testing environments into production inference deployments.
• Security, governance, compliance, and cost competitiveness may limit enterprise adoption of edge AI model hosting.

What to watch

• Revenue contribution and paid conversion of Cloudflare Workers AI, Workers, and R2.
• Whether enterprise edge AI inference scales in 2026 to 2027 as channel checks expect.
• Whether the share of Cloudflare developer platform workloads migrating from AWS, GCP, and Azure continues to rise.
• Whether Akamai Inference Cloud wins more production AI inference workloads.
• Whether Cloudflare can achieve its $5bn revenue run-rate, Rule of 50%+, and long-term 30%+ operating margin targets.
• Share changes among hyperscale clouds, Fastly, Vercel, and other edge platforms.