Wall Street's major banks are reporting strong earnings, with JPMorgan and Goldman Sachs benefitting from AI infrastructure buildout and geopolitical volatility. JPMorgan posted net income of $16.5 billion, up 13% year over year, whilst Goldman saw investment banking fees jump 48%. The AI boom is driving unprecedented corporate borrowing, with banks profiting from debt underwriting, bond trading and advisory services. Goldman led Oracle's $25 billion bond offering in February, one of the largest corporate sales recently. JPMorgan CEO Jamie Dimon cited "AI-driven capital investment" as a key macroeconomic driver. Meanwhile, war-related volatility is boosting trading desks. JPMorgan's fixed income trading rose 21%, driven by activity in commodities, credit and currencies. Goldman's equities division surged 27%, reflecting increased client hedging activity amid geopolitical uncertainty.

Goldman Sachs has lowered its price target on Amazon to $275 from $280 whilst maintaining a Buy rating ahead of the company's earnings report on 30 April 2026. The revised target still implies upside from the current share price of around $240. Analyst Eric Sheridan highlighted four key areas shaping Amazon's trajectory: AWS cloud revenue growth and AI investment returns, rising energy prices affecting margins, the commercialisation timeline for Amazon Leo, and the fast-growing advertising platform. Amazon's AI push through AWS has reached an annualised revenue run rate exceeding $15 billion, whilst its chip business surpassed $20 billion in revenue with triple-digit growth. However, capital expenditures could approach $200 billion in fiscal 2026, pressuring free cash flow despite strong overall performance showing net sales of $716.9 billion and operating income of $80 billion for the full year.

Goldman Sachs is strengthening its cyber defences using Anthropic's Claude Mythos Preview AI model, according to CEO David Solomon. The bank is collaborating with Anthropic and security vendors to accelerate investment in its security infrastructure. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened an urgent meeting with Wall Street leaders in Washington, urging banks to test the model against their systems. Mythos is designed to identify complex exploit chains—linked software vulnerabilities used in sophisticated cyberattacks that security researchers often miss. The model has discovered thousands of bugs, including one in OpenBSD that remained undetected for 27 years. US officials are pushing critical industries towards machine-scale cyber defence, though the approach has sparked international friction with European regulators and internal US government disagreements.