Capital & Industrial Strategy

Anthropic has agreed terms on a $30 billion funding round at a $900 billion valuation, with Financial Times reporting Dragoneer, Greenoaks, Sequoia Capital, and Altimeter Capital as lead investors. This is a confirmed term agreement, not a closed deal — regulatory and administrative steps remain. Simultaneously, OpenAI CFO Sarah Friar stated publicly that the company may raise additional capital even after completing what she characterised as the largest private fundraising round ever, citing a deepening compute crunch as the driver, per Bloomberg.

The strategic logic is straightforward: compute is the binding constraint, and neither lab believes current capital reserves are sufficient to secure the infrastructure position needed to maintain frontier model leadership. Anthropic's $900 billion valuation implies investors are pricing in not just current revenue but a durable infrastructure and enterprise moat. OpenAI's CFO framing this as compute-driven rather than operational shortfalls is notable — it positions further dilution as a strategic offensive move, not a distress signal. The combined fundraising pressure from both labs is actively shaping the price and availability of data centre capacity and GPU supply globally.

Cerebras priced and debuted on Nasdaq, closing its first trading day with a market cap of approximately $95 billion after a 68% first-day gain, per CNBC and Reuters. CEO Andrew Feldman's personal stake is valued at $3.2 billion. The WSJ framed the debut as a clear signal that Wall Street's appetite for AI equity is intact and broadening beyond the Magnificent Seven.

The strategic significance extends beyond the deal itself. Cerebras is a direct competitor to Nvidia in AI inference hardware, using a wafer-scale chip architecture that offers throughput advantages for specific workloads. A $95 billion market cap on debut validates investor willingness to price in a credible alternative silicon narrative — though CNBC's Cramer noted the valuation is already difficult to justify on fundamentals. What this IPO does practically is open the window for a pipeline of AI-infrastructure pure-plays to access public markets, with Chinese AI-adjacent companies separately filing Hong Kong listings, as Shenzhen Adtek Technology's IPO filing demonstrates, per Bloomberg.

Alphabet sold ¥576.5 billion ($3.6 billion) of yen-denominated bonds — the largest ever yen deal by a non-Japanese foreign issuer — as part of a broader pattern of US tech giants tapping foreign debt markets at unprecedented rates, per Bloomberg and Financial Times. Amazon is part of the same trend. The yen market is particularly attractive given the still-wide interest rate differential between Japan and the US, enabling Alphabet to raise at lower effective cost than domestic bonds.

This capital is flowing directly into data centre buildout. The consequence is already visible in energy markets: power prices on the largest US grid (PJM) jumped 76% in Q1 as data centre demand surged, per Bloomberg. The energy cost inflation is beginning to generate political pressure and will eventually constrain the pace of buildout unless utilities accelerate grid investment. For investment strategists, the critical question is whether AI infrastructure capex is front-running actual monetisable demand or whether it reflects rational capacity pre-commitment given long infrastructure lead times.

OpenAI is actively exploring legal action against Apple, alleging the iPhone maker failed to invest sufficiently in their ChatGPT integration and did not deliver the subscriber volumes or product prominence OpenAI expected, per Financial Times, TechCrunch, and Reuters. This is confirmed as an exploration of legal options by a named source — not filed litigation.

The strategic subtext matters more than the legal mechanics. OpenAI's revenue model depends critically on converting distribution reach into paying subscribers. If Apple's integration underperformed expectations, it suggests that embedding AI into existing device ecosystems does not automatically translate to monetisation — and that platform owners retain leverage over how AI features are surfaced and promoted. For OpenAI, which has disclosed that enterprise is 40% of revenue and growing toward 50%, per Bloomberg, the consumer distribution strategy through Apple may be strategically less important than previously assumed. The dispute also illustrates a pattern: OpenAI has now been in conflict with multiple major partners, signalling cultural and contractual friction in its partnership model.

Cisco posted its best single-day stock performance since 2011 after reporting record quarterly revenue and raising guidance, with CEO Chuck Robbins declaring the networking industry is entering a 'supercycle' driven by AI infrastructure buildout, per CNBC and Reuters. Cisco simultaneously cut nearly 4,000 jobs to redeploy capital into AI, a pattern of operational restructuring-for-AI-investment that mirrors moves at Intel, IBM, and others. Foxconn (Hon Hai) reported a 19% profit jump beating forecasts, with AI server assembly the primary driver, per Bloomberg and Reuters. Applied Materials also issued a forecast well above analyst estimates, driven by AI chip and memory demand, per Reuters.

Across networking, server assembly, and semiconductor equipment, the Q1 2026 earnings cycle is delivering consistent upside surprises tied to AI infrastructure. This is not speculative demand — Cisco's hyperscaler orders, Foxconn's server rack production ramp, and Applied Materials' chip equipment sales are all confirmed revenue. The read-through for investors is that the infrastructure layer of the AI stack is still in an accelerating, not plateauing, spend cycle — though the FT flagged analyst questions about whether an AI spending plateau could emerge on a 12-18 month horizon.