🇸🇪 Svenska · 🇬🇧 English · 🇪🇸 Español
AI-generated market briefing for 18 October 2026. The information reflects conditions at publication.
1. AI giants’ earnings must justify $1.2 trillion of investment
Meta, Lam Research and Texas Instruments lead next week’s technology reports after the sharp pullback in AI stocks. Five major AI hyperscalers are expected to lift combined capital spending to about $1.2 trillion next year, while third-quarter S&P 500 earnings are forecast to rise more than 30%.
Why it matters: Investors now need profits and cash flow to grow faster than data-centre spending, or AI valuations may face further pressure.
Source: Reuters, 16 October 2026
2. Trump pauses Russia sanctions for 60 days
The United States is delaying new sanctions and tariffs on Russian exports while assessing their impact on energy and consumer prices. The decision comes with oil prices elevated and India still heavily dependent on Russian crude.
Why it matters: The pause may restrain oil prices in the short term, but it also increases geopolitical uncertainty and reduces pressure on Russia.
Source: Reuters, 18 October 2026
3. Lagarde: bond selloff will slow growth and inflation
ECB President Christine Lagarde says rising bond yields have tightened financial conditions more than the central bank assumed in September. The ECB may still raise rates to a restrictive 2.75% in December if energy inflation spreads more broadly.
Why it matters: Markets may do part of the central bank’s work, but more expensive financing hurts European property, consumption and highly indebted states.
Source: Reuters, 17 October 2026
4. Sweden gets Europe’s first commercial AMD AI supercomputer
France’s Scaleway will build its first AI supercomputer outside France in Sweden using AMD Instinct MI355X chips and the open ROCm software stack. The system will be available to European startups and companies, challenging Nvidia’s more closed CUDA ecosystem.
Why it matters: The investment strengthens Sweden’s AI infrastructure and creates a European alternative to Nvidia’s dominant technology stack.
Source: Reuters, 13 October 2026
5. DeepSeek opens search APIs and updates its fast AI model
DeepSeek has launched V4.1‑Flash‑1017 with stronger project development, shorter answers and faster execution, while opening its Search and Fetch APIs. The model supports up to one million tokens of context, offers open weights and costs up to $0.30 per million input tokens and $1.20 per million output tokens.
Why it matters: Cheap search and coding capability increases price pressure on Western model providers and specialist developer tools.
