Nokia restates financials for discontinued operations

Nokia recast prior-period results to reflect its Fixed Wireless Access CPE and Enterprise Campus Edge units as discontinued operations ahead of planned sales. Under the restated figures, Q1 2026 net sales were €4.4B with an operating margin of 1.9%, but Q2 2026 operating margin fell to -1.0% (sales €4.8B), indicating weaker underlying profitability around the reclassification. Segment performance for H1 2026 showed Network Infrastructure operating margin at 7.4% and Mobile Infrastructure at 10.3%.

Analysis

The real read-through is not that AI spending is rising, but that the hyperscalers are increasingly willing to trade near-term free cash flow for compute control. That tends to favor the picks-and-shovels layer first: networking, optical interconnect, power/cooling, and accelerated compute suppliers with constrained lead times can see pricing power before the platform names see full monetization. For GOOGL, the stock should be judged on whether incremental infrastructure spend preserves search share and expands Cloud attach; if it does, the capex hit is a bridge, not a thesis break.

The contrarian risk is that the market treats spend intensity as a quality signal even when returns lag. If the next two quarters show rising capex without a corresponding step-up in Cloud backlog or ad monetization, GOOGL can underperform despite strong headline demand because FCF expectations will reset lower. The key falsifier is not the spend level itself, but whether incremental revenue per dollar of AI infrastructure stops improving.

For NOK, the portfolio cleanup mostly changes optics, not the earnings power of the remaining core. Stripping out low-return hardware can mechanically lift reported margins, but the bigger second-order effect is competitive: shrinking exposure in edge/CPE can leave Ericsson, Cisco, and other enterprise infrastructure vendors fighting over a smaller, more price-sensitive pool. The stock likely needs proof that core network margins stay above the current run-rate after the divestiture; otherwise this is just a multiple-supporting rearrangement, not a rerating catalyst.

Key Decisions for Investors

• Buy GOOGL on any post-print weakness, or express it with 3-6 month call spreads, on the view that AI capex is protecting search and Cloud franchise value; fade only if the next guide shows capex rising faster than Cloud revenue.
• Long ANET / VRT / NVDA basket versus short software-only AI beneficiaries if you want the cleaner second-order trade from the infrastructure buildout; these names should see the first order-book acceleration before monetization shows up in the platform stocks.
• Avoid initiating NOK longs until the asset-sale process is signed and core Network Infrastructure margin proves durable; a tactical bounce is possible, but the risk/reward is poor if the divested businesses were only masking slow organic growth.
• If GOOGL rallies sharply on the spending news, use that strength to sell near-dated upside and re-enter lower; the key watch item is next-quarter FCF margin and Cloud growth, not the capex headline itself.