February 23, 2026
If you’re preparing for investment banking recruiting, this is one of those market weeks where the easy answer — “AI is bullish for tech” — isn’t good enough.
AI is still driving enormous spending. Nvidia’s results were massive. Meta is committing billions to AMD chips. But the same AI theme also hit IBM hard, because investors started questioning whether AI tools can compress one of IBM’s legacy profit pools. That tension is exactly the kind of nuance you want in interviews.
Instead of saying “AI is changing markets,” say something more specific: AI is rewarding companies with scarce infrastructure, while pressuring legacy service models that depend on complexity, labor intensity, and slow modernization cycles.
IBM Shows the Other Side of the AI Trade
IBM shares fell nearly 13.2% to close at $233.35 after investors reacted to Anthropic’s Claude Code tool, which can help modernize legacy systems written in COBOL. That matters because COBOL is closely tied to IBM’s long-standing mainframe ecosystem and legacy modernization services.
The stock move was severe. The decline wiped out billions in market value in a single session and pushed IBM shares down more than 24% year to date.
For interview purposes, don’t just describe the stock drop. Explain the mechanism.
IBM has historically benefited from complex legacy systems that require specialized expertise, long implementation timelines, and meaningful service revenue. If AI tools can streamline parts of that modernization work, investors may start to question the durability of those margins. That doesn’t mean IBM disappears. It means the market is repricing the perceived defensibility of a business line.
This is a clean way to talk about disruption in banking language:
- Revenue risk: AI could reduce demand for manual legacy modernization work.
- Margin risk: If customers expect faster, cheaper modernization, pricing power may decline.
- Multiple risk: Investors may assign lower valuations to slower-growth legacy technology exposure.
- Strategic risk: Established technology companies may need to defend core businesses while also investing aggressively in AI capabilities.
That’s much stronger than simply saying “AI hurt IBM.” The interesting point is that AI can be both a growth driver and a deflationary force, depending on where a company sits in the value chain.
Nvidia Is Still Showing Real AI Infrastructure Demand
On the other side of the AI discussion, Nvidia reported fourth-quarter revenue of $68.1 billion and net income of $43 billion. Profit rose 94%, and sales climbed 73%. Data center products, mainly chips and networking equipment used for AI systems, represented more than 90% of total revenue.
Those numbers matter because investors have been debating whether AI enthusiasm is getting too speculative. Nvidia’s results suggest the cycle is still backed by real spending, not just narrative.
Margins were also notable. Gross margins held at 75%, which points to continued pricing power. Management guided to roughly $78 billion in revenue for the current quarter, above expectations.
If you bring this up in an interview, frame it around capital allocation and customer demand. Major AI customers are still spending heavily on infrastructure, even as the market shifts from training large models toward running those models in real-world applications. That shift from training to inference is important because it changes what customers need, how much capacity they buy, and which chip suppliers benefit.
The risk side is still there. Investors are watching competition in custom chip design and uncertainty around China sales. But the current data point is straightforward: Nvidia’s AI business is not behaving like a theme with no earnings support.
Meta’s AMD Deal Is a Supply Chain and Strategic Optionality Story
Meta’s multibillion-dollar chip deal with AMD is another useful interview example because it connects AI capex, supplier diversification, and deal structure.
Meta agreed to acquire customized chips with 6 gigawatts of capacity from AMD. AMD’s CEO said each gigawatt is worth double-digit billions in deal value. AMD shares rose as much as 14% after the announcement, and the company reached a market capitalization of $342 billion.
The structure is especially interesting. AMD issued Meta a performance-based warrant that gives Meta the option to buy up to 160 million AMD shares in tranches at an exercise price of $0.01 as Meta places successive processor orders. The arrangement could eventually give Meta up to a 10% stake in AMD.
That’s not just a purchase agreement. It’s a strategic partnership with equity upside tied to execution.
For banking interviews, this is useful because it sounds like something bankers would actually analyze. You can discuss why a large technology buyer might want closer alignment with a supplier: capacity access, bargaining leverage, diversification away from a dominant vendor, and participation in upside if the supplier benefits from the buyer’s demand.
Meta also said it will nearly double AI infrastructure spending this year to as much as $135 billion. The chips, a custom version of AMD’s MI450, will be used primarily for inference workloads and require 6 gigawatts of power, roughly equivalent to the power needed by 5 million U.S. households for a year.
That detail is worth remembering. AI is not just a software story. It’s also a power, chips, data center, and supply chain story.
Tariffs Are Still Moving Global Equity and Commodity Markets
Trade policy was another major theme. European markets closed higher after the U.S. tariff rate came in lower than feared. Markets had declined after President Trump threatened a blanket 15% tariff on global imports, but when the policy took effect, the rate was set at 10% for 150 days.
That lower-than-expected tariff helped calm investors. The STOXX Europe 600 rose about 0.3%, France’s CAC 40 gained 0.47%, Germany’s DAX rose 0.76%, Italy’s FTSE MIB jumped 1.11%, Spain’s IBEX 35 climbed 1.49%, and the U.K.’s FTSE 100 advanced 1.18%. Autos, which are especially sensitive to trade policy, rose nearly 2%.
The interview angle here is simple: markets don’t just respond to good or bad policy. They respond to policy versus expectations. A 10% tariff is still a headwind, but it was better than the feared 15% rate, so equities rallied.
Copper also reflected tariff uncertainty. Copper prices fell 0.7% on the London Metal Exchange to $12,868.50 per ton, slipping below $13,000. The move followed the Supreme Court’s decision to strike down the use of emergency powers for reciprocal trade tariffs and the administration’s announcement of a 15% across-the-board global tariff structure.
The nuance is that some pressure on China’s export sector eased, but existing sectoral tariffs on copper products, aluminum, and steel remained in place. Copper had been consolidating near record highs after an all-time peak in January, but high prices were starting to weigh on physical demand in China, the world’s largest copper consumer.
Rates Are Sending Mixed Signals
The Federal Reserve minutes from the January 27–28 meeting showed a more hawkish tone. The FOMC voted 10-2 to keep the benchmark rate between 3.5% and 3.75%, but some officials were considering rate hikes if inflation does not ease.
That’s important because markets had been focused on cuts. Fed officials appeared less worried about labor market weakness after three consecutive rate cuts in late 2025, and more worried that inflation risks remained elevated. The economy added more jobs in January than in any month in over a year, unemployment fell to 4.3%, and core CPI came in line with expectations.
At the same time, rate traders extended expectations for future cuts into 2027. The December 2026 to December 2027 one-year SOFR spread moved to negative 8 basis points, suggesting investors shifted from expecting hikes in 2027 to expecting cuts. Discussion around AI’s labor market impact contributed to that repricing.
So what do you say in an interview? The Fed is not giving a clean one-direction signal. Near-term inflation concerns are keeping policymakers cautious, while longer-term market pricing is starting to reflect potential labor market disruption and lower future rates.
Mortgage rates also fell below 6% for the first time since 2022. The average 30-year fixed mortgage rate dropped to 5.99%, down from 6.89% a year earlier. Refinancing applications were 130% higher than last year, and an estimated 5.5 million additional households now qualified for a mortgage compared to a year ago.
That matters for banks, housing-linked sectors, consumer confidence, and refinancing activity. It’s a good example of how bond market moves feed into real economic behavior.
Emerging Markets and Small Caps Are Part of the Rotation
Emerging market equities benefited after punitive tariff structures were removed and replaced with a 15% across-the-board framework. The Avantis Emerging Markets Equity ETF, a $20.3 billion actively managed emerging markets ETF, saw a single-day inflow of $429.5 million, pushing assets to an all-time high.
Emerging market ETFs have attracted more than $35 billion in net inflows year to date. South Korea saw particularly strong demand, with $694.7 million of inflows for the week ended February 20, helped by strength in Asia’s semiconductor rally.
South Korea’s Kospi index also broke above 6,000 for the first time after rising more than 40% this year. Semiconductor and automotive stocks drove much of the rally. Samsung Electronics and SK Hynix benefited from tight memory supply across DRAM, NAND, and high-bandwidth memory, while automakers such as Hyundai gained as they repositioned toward AI-related investments including robotics and autonomous-driving infrastructure.
The U.S. rotation story is also worth knowing. The Nasdaq was down roughly 0.36% year to date, and the Magnificent 7 had declined about 7%. The Russell 2000, by contrast, was up more than 6% year to date. That suggests investors were moving out of large-cap technology and into defensive and cyclical sectors.
The proposed trade idea was long defensive and cyclical sectors. The argument had two parts: first, consensus GDP growth was projected around 2.5%, with January ISM at 52.6 versus expectations of 48.5; second, the AI cycle may be moving from the build-out phase to the implementation phase, where non-tech companies use AI to improve productivity.
One M&A Example to Keep Ready: Gilead and Arcellx
Gilead agreed to acquire the remaining stake in Arcellx in a deal valued at about $7.8 billion. Gilead already owned 11.5% and agreed to pay $115 per share in cash for the rest, a 79% premium to Arcellx’s prior close of $64.11.
Arcellx shareholders also receive non-transferable contingent value rights worth up to $5 per share, tied to future sales performance of anito-cel, an investigational CAR T-cell therapy for relapsed or refractory multiple myeloma.
This is a great deal to mention because it includes several banker-relevant features: a large premium, an existing partnership, full control of commercialization, and a contingent value right that shares future upside while managing risk. Gilead and Arcellx first partnered in 2022 to co-develop and co-commercialize anito-cel, and the FDA is currently reviewing the therapy with a decision expected later this year.
How I’d Use This in an Interview
If asked what’s happening in markets, don’t try to recite every index. Pick a theme and connect it across sectors.
“The market is getting more selective on AI. Nvidia’s results and Meta’s AMD deal show that AI infrastructure spending is still very real, but IBM’s selloff shows that AI can also pressure legacy technology service models. At the same time, lower mortgage rates and small-cap strength suggest investors are rotating beyond mega-cap tech, especially as growth expectations improve.”
That answer gives you companies, numbers, and a point of view. It also shows you can think like a banker: who benefits, who gets disrupted, and where capital is flowing next.