The Week in Business, Markets & AI

April 2, 2026 marked the one-year anniversary of President Trump's original "Liberation Day" tariff announcement — and he marked it by unveiling a fresh round of tariffs, including levies up to 100% on name-brand pharmaceuticals. The move immediately triggered the largest single-day global stock market decline since 2020, with the Dow falling nearly 10% and the S&P 500 dropping 10% over two days.

The broader picture is troubling for U.S. asset holders. Since Liberation Day one year ago, international markets — including the Shanghai Composite, South Korea's Kospi, and Japan's Nikkei — have outperformed all three major U.S. indexes. Investors outside the U.S. are deliberately reducing American exposure, a trend analysts are calling "ABUSA" (Anywhere But the U.S.A.). The dollar weakened, gold rose, and even U.S. Treasuries saw stress.

A growing body of research has definitively settled one of the central debates of Trump's trade policy: American companies and consumers are paying the tariffs, not foreign exporters. Goldman Sachs estimates U.S. buyers paid 82% of tariff costs in late 2025, and that figure will remain at 75% through mid-2026. Research from the Federal Reserve Bank of New York and the Kiel Institute reaches similar conclusions.

The economic consequences are showing up in the data. U.S. manufacturing activity contracted for the 10th consecutive month in December. Unemployment has reached a four-year high. Consumer sentiment is at a record low. The CBO has formally projected that tariffs will result in lower real GDP than would have otherwise occurred. Meanwhile, businesses that absorbed tariff costs in 2025 are now passing them to consumers — adding another potential inflationary wave.

The S&P 500's cyclically-adjusted price-to-earnings (CAPE) ratio hit 39.9 in recent months — a valuation last seen during the dot-com crash of 2000. Historical analysis suggests that when the index trades at this level, it tends to fall 4% over the next year and 20% over the next two years. The index's forward P/E of 22.2 is also significantly above its 10-year average of 18.8.

The AI investment boom has been the primary counterargument — the case that elevated valuations are justified by the productivity revolution coming. But with tariffs compressing corporate margins and consumer spending decelerating, that cushion is getting thinner.

March and early April 2026 produced one of the densest periods of AI model releases ever recorded. OpenAI shipped GPT-5.4 on March 5 — including Standard, Thinking, and Pro variants — with the Thinking model scoring 75% on desktop task benchmarks and surpassing human-level performance on economically valuable tasks. Google's Gemini 3.1 Pro now leads 13 of 16 major benchmarks at roughly one-third of the API cost of competing models. Anthropic's Claude Sonnet 4.6 leads real-world work evaluations and powers GitHub Copilot's coding agent.

Most significantly: on March 26, a misconfigured Anthropic data store exposed internal files that revealed the existence of a new model called "Claude Mythos" — internally described as a "step change" beyond the current Opus model. Meanwhile, Meta's Llama 4 Maverick (400 billion parameters, 10 million context window) has made open-source models genuinely competitive with proprietary frontiers for the first time.

On April 2, Google released Gemma 4, its most capable open-model family yet, available in four sizes and built on the same architecture as Gemini 3. The models are specifically designed for stronger reasoning and agentic workflows — meaning AI that can take sequences of actions, not just answer questions. The same week, Microsoft launched new in-house models under the MAI label: MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 — available through Microsoft Foundry.

The first quarter of 2026 saw a record-breaking $267.2 billion in venture funding globally, dominated by AI companies. Hyperscaler capital expenditure is expected to rise 33% in 2026 after a 69% surge in 2025. Morgan Stanley estimates that nearly $3 trillion of AI-related infrastructure investment will flow through the global economy by 2028, with more than 80% of that spending still ahead.

One of the most tangible business impacts of 2026's AI boom is in software development. Generative AI coding tools are enabling development timelines that once took weeks to be measured in hours. Agentic systems are now capable of executing multi-step software engineering tasks with minimal human intervention. GPT-5.4's Thinking variant has officially surpassed human-level performance on desktop task benchmarks — meaning AI agents can now navigate files, browsers, and interfaces autonomously.

The Federal Reserve enters Q2 2026 in a difficult position. The labor market has softened meaningfully — hiring in 2025 slowed more than any year since the Great Recession — and unemployment is at a four-year high, creating pressure to cut rates. But inflation remains above the Fed's 2% target, stuck in the upper 2% range, and tariff pass-through to consumers threatens to push it higher in the months ahead.

The Fed cut rates 25 basis points in December and is widely expected to cut again in early 2026. But the pace of easing will likely disappoint markets that were pricing in aggressive cuts. For small businesses: higher input costs and potentially higher borrowing costs than expected are a challenging combination.