Deep|Delta Electronics: 2H26 Re-acceleration on Deck, VPD Still Overlooked

Overview

We initiate coverage of Delta Electronics (2308.TW), the broadest play on AI power spanning the grid-to-core value chain. We expect 2H26 revenue to re-accelerate on stronger AI PSU demand and explosive VPD module growth, with GPM expanding from 26Q4 on a better mix, economies of scale, and price pass-through, and OPM improving even faster. The market, in our view, has overlooked Delta’s DC/DC business: we estimate it jumps to ~8.3%/~12.2% of total revenue in 2026/2027 on the VPD ramp, where Delta holds >70% share at US$400–600 of content per TPU. At a ~50% GPM, the mix shift should more than offset HVDC’s margin dilution. Liquid cooling should grow another ~80% in 2027, led by AWS’s IRHx, while HVDC enters volume production in 26Q4; fears of a timeline delay are overdone. With SST and SOFC optionality layering in from 2028, we model revenue growth of 53%/28% in FY27/FY28 and EPS of TWD77.3/102.4. The stock trades at ~17.8x FY28E EPS (based on the Sep 4 close of NT$1,825), below its historical P/E range.

Near-term momentum building — 26Q3 print is the next catalyst:

• 2H26 revenue to re-accelerate, driven by stronger AI PSU demand and explosive VPD module growth.
• We expect 26Q3 GPM to hold stable or increase slightly QoQ, with 26Q4 GPM expanding further on 1) better product mix, 2) economies of scale, and 3) pass-through of price hikes. Mgmt. noted OPM should improve faster than GPM, making the setup into 26Q3 earnings more favorable.
• We see the 26Q3 print as the key near-term catalyst, with margin profile improvement the main focus; August monthly revenue (due in early September) should show whether the MoM acceleration extends beyond July’s record high.

Multiple growth engines kick in through 2H26–2027:

1. DC/DC is the most under-appreciated driver, in our view. The market has overlooked Delta’s DC/DC revenue growth into 2027: we estimate DC/DC revenue jumps to ~8.3%/~12.2% of total revenue in 2026/2027, driven primarily by VPD module (Figure 2) and DC/DC shelf shipments
   1. A power discrete (DrMOS) shortage constrained VPD shipments in 1H26 and weighed on Delta’s DC/DC revenue growth. Our checks suggest the shortage has been largely resolved, with VPD shipments accelerating alongside the TPU v8t/v8i L6 PCBA volume ramp from 7M26. The v8t rack (L11) began shipping in late July; with L6 running 4–5 weeks ahead of L11, Delta’s record-high July revenue suggests VPD revenue is already kicking in, and v8i rack shipments from October add a second leg into 26Q4.
   2. Our checks indicate both Zebrafish and Sunfish will adopt the VPD solution, at US$400–600 of content per TPU depending on each configuration’s power rail design. With TPU volume growing from ~3.4m units in 2026 to 9m+ in 2027 and Delta holding >70% VPD share, the runway is substantial.
   3. Components structurally carry higher margins than system products (liquid cooling sidecars, PSU shelves). At a ~50% GPM, a rising DC/DC mix should more than offset HVDC’s margin dilution effect in 2027.
2. We estimate Delta’s liquid cooling revenue will grow ~80% in 2027, supported by a strong CSP infrastructure pipeline. Delta is currently the main supplier of liquid-to-air solutions to META/MSFT/AWS, and of GOOG’s liquid-to-liquid solution.
   1. META’s 40kW AALC project continues to ship in volume, but our checks point to EOL by mid-27Q1, so the contribution to Delta’s 2027 liquid cooling revenue should be marginal.
   2. MSFT’s Callen liquid-to-air sidecar enters EOL by end-2026 as Project Atlas picks up pace. Atlas’ 256kW heat dissipation limit is well above Callen’s 144kW, and the transition should lift Delta’s blended liquid cooling ASP.
   3. AWS’s IRHx project is the key 2027 driver: we expect Delta to ship 10k+ rack sets in 2027, driving the segment to grow another 80%.
3. HVDC on track for volume production in 26Q4; fears of a timeline delay are overdone.
   1. GOOG will be the first CSP to adopt Delta’s HVDC sidecar. Google’s Edison car (Figure 1) leverages the ±400V architecture and is rated at 900kW. At a 2:1 HVDC sidecar-to-IT-rack ratio, total lifecycle demand reaches 30k rack sets, of which we estimate ~13k racks will ship in 2027; Delta starts shipping the Edison car in small volume in 26Q4.
   2. Meta’s HVDC sidecar also adopts the ±400V architecture, at a slightly lower 800kW rating; we estimate Delta will ship ~3,000 HVDC racks for Meta’s Nvidia platform in 2027.
   3. MSFT’s Mt. Diablo follows the OCP ±400V design, with lifecycle demand of 20k rack sets; we estimate ~8.6k rack shipments in 2027.
   4. AWS is the outlier: the only CSP not moving to the ±400V HVDC architecture on the Vera Rubin platform. Instead, it is pairing its LVDC sidecar with Trainium and GB300/VR200, while developing a 1.5MW-rated 800V HVDC sidecar that departs from Nvidia’s reference design for Vera Rubin Ultra. Delta should benefit from this highly customized product, with small volume shipments by end-2027.

Figure 1. Google’s Edison Car

Figure 2. Zebrafish VPD module

Long-term optionality not yet in the price