United Internet 2025 revenue rises but falls short of forecasts

United Internet reported fiscal 2025 revenue of €6.10B, missing analyst expectations of €6.24B but up 1.9% year-on-year. Adjusted EPS rose (driven by lower tax), EBITDA was €1.28B and the company added 700,000 fee-based customer contracts. For fiscal 2026 it guided sales to €6.25B, EBITDA to approximately €1.45B and projected cash capex of €600-650M as it invests in expanding 1&1 mobile and fiber-optic networks.

Analysis

The headline demand from Tesla/SpaceX for NVidia-class accelerators is a structural de-risk for NVDA’s ASP and cycle visibility: orders from industrial and edge compute buyers lengthen product lifecycles and create a second bucket of demand less correlated with hyperscaler refresh cadence. That reduces the amplitude of NVDA’s revenue cycles over 6-18 months and increases tolerance for fabs to prioritize H100/A100-class supply at premium prices, supporting both margin and multiple expansion in the near term.

Server OEMs (Supermicro/SMCI analogs) are the natural second-order beneficiaries: persistent, diversified bookings convert into higher billings for dense GPU chassis, optics, and power subsystems over the next 3–12 months. However, this simultaneously intensifies competition for long-lead components (optical modules, PSUs), driving upstream lead times and giving suppliers pricing power — a squeeze that can flip from tailwind to headwind if hyperscalers aggressively cut orders or start vertically integrating.

Key risks are concentrated: (1) OEM vertical integration (e.g., Tesla developing in-house silicon) over 12–36 months that re-routes a share of demand away from NVDA; (2) a rapid easing of supply-side tightness that collapses NVDA ASPs and removes the premium; and (3) cyclic weakness in adjacent end-markets (advertising/mobile users) that drains discretionary capex from smaller buyers. Watch NVDA earnings cadence, SMCI booking trends, and telco/fiber capex announcements as 30–90 day catalysts that will validate whether this demand pool is durable or one-off timing of shipments.

Key Decisions for Investors

• Long NVDA via a 3–6 month call spread (buy-to-open near-the-money, sell-to-open 10–15% OTM) to capture continued enterprise/industrial orders; target +25–40% upside on spread if NVDA prints continued enterprise traction, max loss = premium paid, stop-loss: close on NVDA guide miss or >30% IV contraction post-earnings.
• Long SMCI equity or 9–12 month LEAP calls to play OEM order flow and chassis demand; position size = tactical (2–4% portfolio) with a 25% trailing stop, take-profits at +40–60% as bookings convert; hedge by selling short-dated calls to finance premium if implied vol is rich.
• Pair trade: long NVDA / short APP (3–6 months) to express hardware-infrastructure outperformance vs adtech cyclicality — target divergence: NVDA +30% / APP -15%; risk controls: exit pair on macro ad rebound indicators (CPIs, platform ad RPMs) or if NVDA shows material inventory destocking.
• Event play: small TSLA allocation via calendar call spread around next production/earnings print (buy 6–9 month calls, sell 1–2 month calls) to monetize potential beat on autonomy/compute integration while funding decay; cap position at 1–2% portfolio given execution and vertical-integration risk.