AI infrastructure

Wall Street saw mixed trading amid optimism about avoiding a government shutdown. The Dow hit a record high, while tech stocks pressured the S&P 500 and Nasdaq. Anthropic announced a $50 billion AI infrastructure investment, including data centers in New York and Texas. JPMorgan Chase forecasts $5 trillion in global AI infrastructure investment by 2030. TD Cowen raised Broadcom’s price target, citing increased AI spending, but acknowledges high expectations are already priced in. The focus shifts to future revenue expectations after the earnings call.

Cisco reported strong Q1 2025 results, exceeding expectations for both revenue ($14.88B) and EPS ($1 adjusted). Overall revenue grew 8% YoY, driven by a 15% surge in the networking segment. AI infrastructure orders reached $1.3 billion, highlighting Cisco’s strategic positioning. While security and collaboration segments saw declines, Cisco anticipates robust Q2 revenue ($15-$15.2B) and raised its full-year outlook, driven by AI and networking demand. Cisco shares are up 25% year-to-date.

Anthropic is investing $50 billion to build AI infrastructure in the U.S., starting with data centers in Texas and New York in partnership with Fluidstack. This includes 800 permanent and 2,000 construction jobs. The move emphasizes U.S.-based AI capabilities amid rising competition. Anthropic aims for profitability by 2028, leveraging custom facilities and partnerships with Amazon and Google for compute resources. Questions remain regarding energy demands, potential AI bubble risks, and government involvement in AI infrastructure funding.

CoreWeave (CRWV) shares plummeted 13% after CEO Mike Intrator cited data center delays impacting its full-year projections. While Intrator portrayed the issue as isolated, CNBC’s Jim Cramer suggested broader problems, potentially tied to Core Scientific (CORZ), whose shares also fell 7%. CoreWeave revised its 2025 revenue forecast downwards despite a strong third-quarter revenue growth of 134%. The market reacted to the ambiguity surrounding the delays’ extent and their potential impact on CoreWeave’s ability to fulfill major AI infrastructure contracts with companies like Meta and OpenAI.

CoreWeave, an AI-focused cloud provider, reported strong Q3 earnings, exceeding revenue expectations with $1.36 billion, a 134% year-over-year increase. While reporting a net loss, it was significantly improved compared to the previous year. The company boasts a $55.6 billion contracted backlog and secured major partnerships with OpenAI and Meta. Since its March IPO, the stock has risen 164%. However, a proposed acquisition of Core Scientific was rejected. Investors are watching for updates on capacity expansion and profitability strategies.

This week’s portfolio adjustments involve a partial sale of Cisco (CSCO) due to security revenue concerns, despite long-term AI optimism. Proceeds are strategically reinvested into Corning (GLW) and Meta Platforms (META). Corning is expected to benefit from growing demand for fiber optic solutions driven by AI data center infrastructure. Meta is seen as undervalued after a recent pullback, presenting an attractive entry point based on its AI potential and monetization capabilities.. These moves aim to balance risk and capitalize on long-term growth trends in AI and data infrastructure.

OpenAI CFO Sarah Friar clarified that the company is not seeking a government “backstop” for its infrastructure plans, despite earlier suggestions. Instead, she emphasized a collaborative approach between the private sector and government to build U.S. technological capacity. This comes as OpenAI undertakes $1.4 trillion in infrastructure agreements to meet AI demand. Friar projects $13B+ revenue this year, but spending commitments have raised concerns about financial sustainability, which CEO Sam Altman dismissed. Friar highlighted ongoing collaboration with the U.S. government, underscoring AI’s role as a national strategic asset.

OpenAI projects over $20 billion in annualized revenue this year, aiming for hundreds of billions by 2030. CEO Sam Altman emphasizes massive infrastructure expansion to meet AI model demand, with reported deals exceeding $1.4 trillion. CFO Sarah Friar’s suggestion of a potential federal “backstop” for semiconductor investments sparked controversy and was later clarified, with OpenAI asserting they do not seek government guarantees for data centers, relying instead on private sector investment. Analysts highlight the risks and competition in OpenAI’s aggressive expansion strategy.

Google has launched its seventh-generation Tensor Processing Unit (TPU), Ironwood, now generally available to cloud customers. This custom-designed AI accelerator aims to rival Nvidia GPUs by offering specialized, high-performance computing power for AI workloads like training LLMs and real-time inference. Ironwood boasts interconnectivity of up to 9,216 TPUs in a single pod. AI startup Anthropic has pledged to utilize up to 1 million Ironwood TPUs. The move underscores Google’s commitment to AI infrastructure amid growing cloud revenue and increased capital expenditure.

Super Micro Computer (SMCI) shares plunged 10% in after-hours trading as fiscal first-quarter results missed analyst expectations. EPS was 35 cents (adj.) versus 40 cents expected, and revenue reached $5.02 billion, below the $6 billion forecast. Revenue declined 15% year-over-year. Net income also more than halved. The company cited “design win upgrades” for the shortfall, shifting revenue to Q2, where they now forecast $10-$11 billion in sales. Despite benefiting from AI growth, SMCI’s growth seems to have slowed amid increasing competition.