AI, tariffs and markets reshape global business

US equities opened mixed on Tuesday as President Trump’s new 10% global tariff formally came into effect. The Dow Jones Industrial Average gained about 0.3%, while the S&P 500 and Nasdaq Composite edged lower, with investors still digesting Monday’s sharp sell-off tied to fears of AI disruption and trade tensions.

The new levy has unsettled trading partners, with the EU and Japan protesting that the duties leave them worse off than existing arrangements. Several reports cited by Yahoo Finance said the White House is working on a formal order to lift the rate to 15%, as previously threatened, and preparing national‑security probes into sectoral imports that could pave the way for further tariffs.

The tariff regime is already feeding into corporate decision‑making. A BBC report on a US business facing Chinese tariffs noted that the company has had to hold back productive investment in order to maintain reserves against unpredictable policy shifts. In Congress, Republicans remain split on tariffs, and House Speaker Mike Johnson described the search for “consensus on any path forward” as a challenge.

Trade policy will stay in the spotlight as Trump delivers his State of the Union address later on Tuesday. Investors are watching for additional detail after the Supreme Court struck down an earlier emergency tariff framework, forcing the administration to rely more heavily on conventional trade tools and time‑limited authorisations.

AI infrastructure spending remains a central theme for large technology groups. A Zacks Investment Ideas note highlighted that combined capital expenditure by Microsoft, Alphabet and Amazon on AI‑related infrastructure is expected to reach roughly $515bn this year and about $600bn next year. Zacks argued that concerns over AI spending and funding are “overdone,” saying such outlays could feed through to adjacent sectors including energy and construction.

Alphabet illustrates the tension. Its shares closed at $311.69 on 23 February, down 1.02% on the day but still up more than 75% over 12 months, with a market value near $3.77trn. Fourth‑quarter 2025 revenue rose 18% year on year to $113.83bn, and annual revenue passed $400bn for the first time. Wells Fargo upgraded the stock to Overweight, citing a plan to expand AI infrastructure from 15GW to 35GW of capacity by 2028 and lifting its price target to $387.

At the same time, Alphabet has guided for 2026 capex of $175bn–$185bn, potentially double prior levels, to fund data centers, servers and networking. Some commentators, including investor Michael Burry, have questioned whether hyperscaler spending could pressure free cash flow as depreciation rises. Alphabet’s shares are down around 7% month‑to‑date amid these concerns, even as analysts broadly characterise its forward P/E of roughly 22–24 as reasonable given growth.

Federal Reserve officials are also tracking AI‑related investment. Governor Lisa Cook told an economics conference that soaring AI spending on data centers and chips may be contributing to strong aggregate demand despite higher interest rates, potentially lifting the near‑term neutral rate. She stressed that while AI could boost productivity, the transition may entail job displacement before new roles emerge, complicating monetary‑policy trade‑offs.

Anthropic stepped up competitive pressure in enterprise software on Tuesday with the launch of new Claude AI capabilities aimed at corporate customers. The company introduced department‑specific plugins, tools for customers to build bespoke plugins, and integrations with services such as Google Drive, Gmail, DocuSign and LegalZoom. Anthropic is also offering a private marketplace for firms to host internal plugins.

The company has framed these integrations as enabling enterprises to build “mini apps” that can automate workflows across teams. Wall Street analysts and investors have interpreted the move as another sign that AI providers may either create rival functionality to existing software suites or make it easier for businesses to develop in‑house tools, increasing pricing pressure on incumbent vendors.

Enterprise software shares have been under sustained pressure since Anthropic first previewed its Claude Cowork product at the end of January. According to sector data cited by Yahoo Finance, ServiceNow is down more than 23% since that announcement, Salesforce has fallen 22%, Snowflake 20%, Intuit 33% and Thomson Reuters 31%. IBM dropped 13.5% on Monday after Anthropic published a blog on AI‑assisted COBOL modernisation, although IBM has highlighted its own investments in code‑modernisation tools and argued that mission‑critical workloads still favour mainframe architectures.

A widely read Substack post from Citrini Research also contributed to volatility by sketching a scenario in which AI “vibe‑coding” of internal tools erodes demand for traditional SaaS products and compresses white‑collar employment and wages. The post has intensified debate over whether AI will ultimately be a net creator or destroyer of jobs in knowledge sectors, a question Fed officials have also acknowledged as highly uncertain.

Rising AI workloads are reshaping investment along the hardware and infrastructure stack. In chips, Advanced Micro Devices’ shares rose in early trading after the company reached a multi‑year agreement to supply Meta Platforms with a large volume of GPUs for its AI build‑out. Wells Fargo separately argued that Qualcomm’s ambitions in AI data‑center inference are under‑appreciated, estimating potential data‑center revenue of $5bn–$7bn annually from 2027 if its AI200 and AI250 platforms gain traction.

On the cloud side, Oracle is positioning its data‑fabric platform, built on Oracle Cloud Infrastructure, as an AI‑enabled layer for hybrid and multicloud deployments. Oracle emphasised support for real‑time data integration across OCI, AWS, Microsoft Azure and Google Cloud, as well as governance and lineage features for AI workloads.

Physical‑layer constraints are also drawing investment. Corning has reported strong demand for its high‑density optical fiber, describing AI data‑center networking as a “critical bottleneck” and highlighting products such as ultra‑high‑count cables and thinner fiber designs that allow more connections in existing conduits. The company’s 2025 core sales grew 13% to $16.41bn, with AI‑linked fiber orders helping it reach a 20.2% core operating margin in Q4 2025.

Global infrastructure investors are seeking to consolidate AI‑related assets. Brookfield’s AI Infrastructure Fund has launched Radiant, a platform formed via a merger with Ori Industries, to build “AI factories” based on NVIDIA’s DSX reference design. Radiant is targeting long‑term contracts with sovereigns and large enterprises, and has access to a $100bn Brookfield investment programme to combine powered land, data centers and software orchestration into integrated offerings.

Network and connectivity vendors are also reacting. Hewlett Packard Enterprise, which has integrated Juniper Networks, announced new PTX‑series routers and telco‑optimised servers at Mobile World Congress to support high‑capacity AI traffic, and highlighted financing programmes to help carriers modernise networks. Separately, data‑platform firm Dataknox said it has secured priority access to NVIDIA’s Rubin NVL8 systems, while security provider Evolv detailed deployments of its AI‑based screening systems at venues such as Inter Miami’s new stadium.