Tensions remain elevated, keeping oil prices under upward pressure. Germany’s Chancellor Friedrich Merz, angered by the prolonged tensions, said that “the US is being humiliated by Iranian leaders.” Crude near $100 per barrel is unnerving even allies, while trade tensions with the US are pushing countries to forge partnerships outside the US.

Meanwhile, this—arguably unnecessary—geopolitical and trade mess is pushing global prices higher and derailing most major central banks from their initial plans to end monetary tightening. Instead, they may be forced to consider tightening again to tackle rising inflation, even as growth prospects worsen.

In Japan, policymakers halved their FY2026 growth forecast from 1% to 0.5%, while revising inflation expectations higher. They kept rates unchanged instead of hiking. The USDJPY is lower this morning, testing the 159 level. Shorting the yen at these levels is not particularly attractive given the limited upside toward 160, but going long yen could become interesting if the US dollar eases—potentially on the back of renewed peace hopes.

Across the Pacific, the Federal Reserve (Fed) starts its two-day policy meeting today and is also expected to keep rates unchanged. In fact, the Fed is no longer expected to cut rates until December, even though Kevin Warsh supports the idea of cuts — a move that would greatly please the White House.

Inflation expectations have risen uncomfortably over the past two months, driven by a notable jump in energy prices. At this stage, no one — including central bankers — can predict what comes next if Middle East tensions continue to disrupt energy flows. What we do know is that the longer the Strait of Hormuz remains under strain, the stronger the impact on markets will be. Oil prices are currently trading roughly 50% above pre-conflict levels, and that is changing the way investors perceive risk.

Two months ago, appetite for Big Tech was weakening in favour of non-tech and non-US markets. That shift was largely driven by massive AI spending plans from Big Tech, combined with uncertainty around the timing of returns on those investments. Investors also grew concerned that additional spending was eroding free cash flow and forcing companies to take on leverage to finance AI buildouts. Oracle’s CDS levels were thrown to every sauce as a reflection of those concerns: too much investment, too much leverage, uncertain returns.

But things have changed since then.

Today, the investment pledges from Big Tech have not changed, but the way investors perceive the risks has. Big Tech has proven resilient during the Middle East crisis and has increasingly been treated as a “safe harbour”. At the same time, Anthropic has helped support weakening sentiment around OpenAI and broader concerns about circular AI deals. These circular structures are now increasingly forming around Anthropic, with major tech companies investing in the firm, which in return commits to purchasing chips and computing power from them.

However, stepping back, even though the Anthropic news gave a sugar rush to AI investors in Q1, the underlying story has not changed. AI is extremely promising, but also highly capital intensive. Big Tech must continue investing to meet demand, but much of that demand is still circulating within the AI ecosystem rather than being driven by external end users. OpenAI and Anthropic sit at the centre of these flows and are under pressure to DELIVER.

Even though Anthropic’s revenue potential has been boosted by its latest model — which is considered too powerful to be widely released and could support higher-value contracts, including government deals — the gap between revenue generation and costs remains large.

How large? Even with recent estimates, Anthropic’s annualised revenue run-rate is believed to be around $20–30bn, while OpenAI is broadly estimated at $25bn+, depending on sources (both are private companies with no official figures). However, both are backed by far larger capital commitments: OpenAI through hundreds of billions of dollars in compute-related partnerships and expected funding capacity, and Anthropic through tens of billions in recent strategic investments from major cloud providers such as Amazon and Google.

Against that backdrop, profitability is still a long way off. Reports suggest both companies are still losing money once the full cost of running large-scale models is included, particularly GPU and cloud computing expenses. As a result, margins remain under pressure, as the cost of scaling AI is rising almost as quickly as demand.

In that sense, the AI structure is building like a house of cards on expectations of success from Anthropic and OpenAI. If anything were to go wrong with these two, AI would not disappear, but markets would face a significant repricing.

The upcoming Big Tech earnings will therefore be closely watched. Last quarter, most Big Tech companies beat earnings and revenue expectations, but that did not prevent investors from rotating out of the sector. In particular, markets reacted negatively to further increases in capital expenditure at a time when revenue growth was not accelerating in line.

Let’s see whether Anthropic can provide relief. Chipmakers and AI enablers continue to post strong results — unsurprisingly, as spending flows through the AI value chain — but it remains unclear whether this is translating into meaningful end-user revenue growth outside the ecosystem. Investors will therefore remain laser-focused on cloud division growth. Any disappointment could bring back a key concern: are these investments becoming obsolete before they start generating returns?

To me, the risks remain tilted to the upside given the speed of AI adoption and the potential in agentic AI and robotics. But valuations are once again elevated.

The Nasdaq 100’s P/E ratio eased following the latest earnings season, but investors were quick to take advantage of the cheaper valuations. As a result, Big Tech is once again priced to perfection into earnings, leaving little margin for disappointment. Strong results could continue to mask underlying economic weakness (trade, political & geopolitical mess) for a while, but any disappointment could derail tech appetite at a time when non-tech names are unable to pick up the slack.