DELL (Trans): AI backlog of 95 bn; no growth guidance for next year for now.

Dolphin Research summary of $Dell Tech(DELL.US) FY27 Q2 earnings call Trans

Earnings review: NVIDIA ships more? Dell: The good times can get even better

I. Core takeaways

1. Guidance raised

a. Full-year revenue guide lifted by $25bn to a $192bn midpoint (+~70% YoY), with non-GAAP EPS of $25.50 (+~150% YoY). Management signals broad-based strength across infrastructure and client segments.

b. By segment: ISG ~+120% YoY, AI server revenue up 3x YoY to $74bn; traditional servers slightly above +100%; storage and CSG both mid-teens. Mix skews to higher-compute configurations and richer memory/storage bills of materials.

c. Q3: revenue midpoint $49bn (+~80% YoY), ISG ~+145% with AI server revenue of $19bn, and CSG ~+15%. EPS is guided to $6.50 (+150%+ YoY), with diluted shares at ~651mn.

2. Margins and opex

a. Full-year opex at ~8% of revenue, the lowest in 42 years; OP up ~120% with OPM expanding by >200bps YoY. Efficiency and scale are the main drivers.

b. Ex-AI server mix, full-year GPM improves YoY, with H2 outlook better vs. 90 days ago; ISG margin in Q3 to rise by slightly over 100bps YoY. Pricing discipline and mix also help.

c. CSG OPM to ease from 7.6% to ~6% in Q3, with opex down low-single-digit QoQ; I&O expected at -$1.4bn to -$1.5bn. Management remains focused on cost control.

3. Shareholder returns: Q2 returned a record $4.3bn, up $2.2bn QoQ, including buybacks of 9.5mn shares at an Avg. price of $401 and a dividend of $0.63 per share. H1 cumulative returns were $6.3bn, also a record, and the company received a positive credit rating action this quarter.

4. Key financials

a. Revenue of $47bn (+58% YoY), GPM 21.1%, OP $5.9bn (+160% YoY) with OPM at 12.6%, and EPS of $7.04 (+203% YoY). Scale and mix supported profitability.

b. ISG revenue $31.8bn (+89% YoY), margin 15% (+620bps YoY); CSG revenue $15.0bn (+20% YoY), margin 7.6%. ISG led growth on AI and modernization demand.

c. Opex at 8.5% of revenue (-250bps YoY), Adj. FCF of $8.1bn, cash & investments of $14.2bn, and leverage at 0.8x. Balance sheet remains strong.

II. Call details

2.1 Key management commentary

1. AI infrastructure

a. Q2 AI orders hit a record $60.9bn, AI server revenue reached $16.4bn, and AI backlog ended at a record $95.0bn. Momentum continues to build across customer types.

b. Over the past 12 months, $131.7bn converted into orders, with the five-quarter pipeline still up QoQ and sized at multiples of the backlog. Visibility remains strong.

c. Demand is broadening to neocloud, sovereign, and enterprise customers. These are Dell’s core battlegrounds.

d. Some projects require up to 50 unique designs to meet workload performance, power, cooling, and data center conditions. Customization is a clear differentiator.

e. Dell became the first to ship rack systems based on the NVIDIA Vera Rubin platform. Early platform leadership matters.

2. Traditional servers and networking

a. Revenue rose +122% YoY, with demand still exceeding supply; most growth stems from existing customers modernizing and consolidating data centers. Refresh cycles favor higher-core CPUs, more DRAM, and richer storage.

b. CPU compute to support AI and agentic workflows is forming incremental demand. This is additive to traditional server needs.

c. Traditional server share gained by over 10ppt across the past two quarters, with more share gains expected this quarter. Execution remains solid.

3. Storage

a. Revenue grew +26% YoY, with Dell IP demand outpacing the market for six consecutive quarters; Dell IP mix in storage continues to rise and margins keep improving. Proprietary offerings carry higher GP.

b. PowerFlex, PowerStore, PowerProtect, and PowerVault all grew, with PowerStore posting double-digit demand growth for nine straight quarters. Breadth across the portfolio remains strong.

c. PowerScale and ObjectScale drove unstructured storage to double-digit growth for three consecutive quarters. AI use cases are accelerating non-structured data needs.

4. CSG

a. Revenue rose +20% YoY, the fastest in five years, with demand up across all regions and verticals. Commercial strength led the results.

b. Commercial grew +22% YoY, marking eight consecutive quarters of revenue growth and ten consecutive quarters of demand growth; large enterprise refreshes drove double-digit growth across regions. Fleet upgrades remain a tailwind.

c. Price-sensitive customers are stretching refresh cycles, increasing the stock of older devices, which enlarges the long-term refresh opportunity. Deferred demand builds latent potential.

d. Consumer rose +7% YoY, with demand up for the fourth consecutive quarter; profitability benefited from pricing discipline and scale. Mix management supports margins.

5. Strategic views

a. Customers no longer see IT as a cost center but as a value driver, and are expanding and reallocating budgets. This underpins multi-year investment cycles.

b. AI needs decoupled architectures to keep data available and fluid across compute, storage, and networking, while accelerating traditional IT investment. Reference designs are evolving to support this flow.

c. On-prem and edge deployments deliver better token economics for suitable workloads and allow customers tighter control over data and IP. Data sovereignty remains a key theme.

2.2 Q&A

Q: Traditional servers +122% and storage +26% — driven by price hikes and pull-ins, or real demand?

A: Primarily real demand from existing enterprise customers modernizing, with 1.2mn units still on 14G or earlier. Modernization drives consolidation, with new servers packing more cores, more DRAM, and more internal storage; 17G consolidation ratios are 6–8:1, and 18G, shipping next month, could reach 12–14:1. Security posture is a forcing function, with resilience needs and post-quantum cryptography pushing legacy replacements; agentic workloads are also adding to traditional server demand.

On storage, the logic is similar: data volumes keep rising and must be stored, encrypted, and protected. In H1, storage grew +17% YoY in aggregate, with Dell IP demand outpacing the market for six straight quarters; all-flash arrays have grown for ten consecutive quarters, PowerScale for five, ObjectScale for four, and Data Domain for three. Management expects H2 growth to match H1 levels — storage mid-teens and traditional servers continuing triple-digit growth.

Q: Can growth next year keep pace with NVIDIA? How do you view the long-term CAGR?

A: No FY guide for next year; H2 growth guided to 68% and a >$1tn long-term TAM estimate. Management noted H2’s 68% mirrors H1’s 71%, and the five-quarter pipeline continues to rise QoQ. Long-term: inference demand has surpassed training, with tokens driven by inference projected to grow 87x to 3,600 quadrillion by 2030; training demand to rise 5x to 850 zettaflops by 2030; enterprise agentic workloads are expected to be the largest single workload by 2028.

By 2030, AI could account for 75% of total data center demand, with an additional 200GW of power added over the period, roughly half landing in neocloud, sovereign, and enterprise — Dell’s core arenas — implying a >$1tn opportunity. This underpins multi-year infrastructure cycles.

Q: Are traditional server customers still mostly enterprises? Within AI servers, how do enterprise vs. hyperscale growth rates compare?

A: Traditional server growth is mainly enterprise-driven, while AI mix is unchanged but enterprise is accelerating. Demand exceeds supply, so actual demand outstrips reported revenue. AI servers are entering these cohorts, including neocloud, high-frequency trading clients, and leading large enterprises deploying AI.

Dell AI Factory now serves over 6,500 customers, with 3,300 added over the past three quarters, while the first 3,200 took eight quarters; the acceleration is partly enterprise-led. Enterprise customer count, repeat buyers, and revenue all grew QoQ and YoY, with enterprise pipeline also up QoQ. Mix remains steady as sovereign and large neocloud wins continue, and enterprises often purchase more storage and networking alongside compute.

Q: What new storage growth vectors emerge in the AI era?

A: Compliance retention, KV cache, and physical AI data are new storage pathways. AI-generated content faces varying compliance and retention requirements and protection strategies by customer type, creating incremental storage demand; KV cache is used to improve inference efficiency, providing another vector. As agentic workloads and physical AI scale, manufacturing, IoT sensors, and robotics will produce large volumes of multimodal unstructured data, with growth accelerating rather than slowing.

The company is building a data semantic layer and a data automation platform to turn data management into feeding capabilities for AI engines. This aims to operationalize data for AI.

Q: How are supply constraints now vs. last quarter?

A: Constraints are similar; DRAM is tightest, followed by NAND. Supply is ‘never enough’; Dell is optimizing allocation, kit matching to maximize factory output, and shaping demand and scheduling; the $25bn full-year guide raise comes directly from better inbound optimization and shipping capacity. Earlier this year, Dell anticipated a softer PC market in H2 and tilted parts toward infrastructure, a move with a delivery lag now playing out, which is part of why the H2 outlook improved.

Shortage list: sporadic CPU and HDD shortages; advanced-node products broadly constrained; mature-node MOSFETs, power ICs, microcontrollers, and driver ICs constrained; ABF substrates and T-glass shortages; optical modules short; AI supply chain running at full tilt on CDUs and power racks. Bottlenecks remain across components.

Q: What is Lightning contributing today? What is the attach rate in cloud AI server deals?

A: No quantified contribution or attach rate; still in beta. Lightning is a parallel file system for native AI use cases, relatively new and in beta at several customers, with run-off comparisons against competitors underway. Management will update once momentum builds.

Q: For traditional servers and storage growth, how much is volume vs. price? Do customers have headroom for higher prices, any resistance?

A: No volume/price split; growth includes inflation components, and management did not comment on price resistance. On servers, higher core counts, more DRAM, and more storage configurations are lifting costs sequentially, with pricing passing through input costs; underlying demand is strong, with agentic AI bringing new use cases and incremental demand, and supply has lagged demand for two consecutive quarters with a healthy pipeline. On storage, declines in partner IP were offset by higher shipments of Dell IP, which carries higher revenue and margins; unstructured products are seeing unprecedented adoption in AI applications.

Dell IP has outpaced the market for six straight quarters and is expected to keep gaining share, with part of the growth also driven by rising material costs. Software-defined storage is performing well and is incremental. Momentum is broad-based.

Q: Within ISG margin expansion, how much is like-for-like uplift from 14G to 17G/18G migration vs. scale efficiencies?

A: No like-for-like breakdown; scale is the primary driver, contributing slightly over 400bps. In Q2, ISG OPM was 15%, with scale contributing over 650bps on a full-year basis; storage’s full-year +15% growth guide implies roughly $2.5bn of incremental storage revenue, a meaningful contributor. H2 ISG margin to rise by over 100bps YoY, with Q4 above Q3, even as AI server revenue grows over 3x YoY to $74bn.

Further uplift comes from product and regional mix, pricing discipline, and execution by supply chain, engineering, and sales. Scale and mix remain key levers.

<End of text>

Risk disclosure and statement:Dolphin Research disclaimer and general disclosure