The 13-Minute Tax Fix That Saved My Retirement (from Salem Media Group)
What a week!
Chip records shattered. A Pentagon ultimatum ignored.
An $82 billion deal declined in two hours. The AI boom got louder, the power grabs got uglier, and the stock market had an identity crisis.
All in five days. This is the week 2026 stopped being a prediction and became a reality.
Nvidia Earnings Beat Expectations
Nvidia Earnings Report
Key Points:
Data center revenue of $62.3B came in ahead of the $60.69B estimate, driven by Blackwell chip demand from hyperscalers.
Gaming revenue fell 13% QoQ. Nvidia is prioritizing AI infrastructure over consumer GPUs, and anyone trying to upgrade their rig has noticed.
Combined hyperscaler capex for 2026 is tracking toward $650 billion, a figure that makes Nvidia's $78B Q1 guidance look less like a stretch and more like a floor.
Nvidia reported Q4 FY2026 results on Wednesday evening and blew the doors off expectations. Again. Revenue hit $68.1 billion, up 73% YoY, crushing Wall Street's $66.2B estimate. Data center alone pulled in $62.3 billion, over 91% of total sales. Net income nearly doubled to $43 billion. Then CEO Jensen Huang dropped Q1 guidance of $78 billion, which is basically Nvidia telling every skeptic to sit down. The hyperscalers, Microsoft, Google, Meta, and Amazon, now account for over 50% of data center revenue and are buying chips faster than Nvidia can make them.
Jensen Huang declared the "agentic AI inflection point has arrived" and confirmed Vera Rubin next-gen samples are already shipping to customers. James Demmert, CIO at Main Street Research, called the print exactly what "extremely hungry" investors needed: confirmation that the AI storyline is intact. Jensen also pushed back on the "Software sell-off" narrative directly, telling investors the markets "got it wrong" on AI's threat to software companies.
What does Nvidia's quarter tell you?
Anthropic Rejects Pentagon Offer
Under Secretary of War Emil Michael (x.com)
Key Points:
The "supply chain risk" designation is the real weapon here — it could sever Anthropic from Boeing, Lockheed, and every other company in the defense industrial base.
A former head of the Pentagon's AI office, retired Air Force Gen. Jack Shanahan, said "painting a bullseye on Anthropic garners spicy headlines, but everyone loses in the end."
Elon Musk's xAI signed on to the Pentagon's "all lawful purposes" standard this week for classified work, positioning Grok as a potential Claude replacement.
This is the biggest story of the week, and possibly the biggest AI policy moment of the year. Anthropic signed a $200 million contract with the Department of War last summer, making Claude the first AI model cleared for use on classified military networks.
This week, that relationship hit a wall. Defense Secretary Pete Hegseth met with CEO Dario Amodei on Tuesday and issued an ultimatum: allow Claude to be used for "all lawful purposes," including potential mass domestic surveillance of Americans and fully autonomous weapons, or lose the contract.
Amodei rejected the Pentagon's "final offer" on Thursday, saying the new contract language contained loopholes that would allow safeguards to be bypassed "at will." Hegseth threatened to invoke the Cold War-era Defense Production Act and label Anthropic a "supply chain risk," a designation normally reserved for foreign adversaries. That would prohibit every company with a defense contract from using Anthropic products, potentially devastating its enterprise growth at a critical moment. The Pentagon's top spokesman called Amodei a "liar" with a "God complex."
Amodei shot back that the Pentagon's two threats are "inherently contradictory: one labels us a security risk; the other labels Claude as essential to national security." As of Friday's 5:01 p.m. deadline, Anthropic has not blinked. Sen. Thom Tillis broke with the administration, saying Anthropic was "trying to do their best to help us from ourselves."
Where do you stand?
OpenAI and Google Workers Push Back
2025 Bloomberg Finance LP
Key Points:
Both Google and OpenAI have already accepted the "all lawful purposes" standard in their existing Pentagon contracts, which is precisely why their employees are now asking hard questions.
Anthropic staffers went public on X in support of Amodei, with one writing "history is unfolding in front of us."
Google recently updated its ethics guidelines to drop its prior pledge not to develop AI for weapons, and OpenAI removed "safety" as a core value from its mission statement, context that makes the employee letter more striking, not less.
The Anthropic-Pentagon standoff didn't stay contained. By Thursday night, a joint open letter titled "We Will Not Be Divided" began circulating across Silicon Valley, signed by more than 220 current and former employees at Google and OpenAI: 176 from Google, 47 from OpenAI. The letter, addressed to leadership at both companies, urged them to refuse the Pentagon's demand to use their AI models for domestic mass surveillance or fully autonomous weapons.
"The Pentagon is negotiating with Google and OpenAI to try to get them to agree to what Anthropic has refused," it states. "They're trying to divide each company with fear that the other will give in. That strategy only works if none of us know where the others stand." Google's AI chief Jeff Dean publicly expressed solidarity with Anthropic's position, posting that mass surveillance "violates the Fourth Amendment and has a chilling effect on freedom of expression." Dean Ball, a former White House senior tech policy advisor, told Business Insider: "We're absolutely in uncharted territory. Anthropic could be quasi-nationalized, or they could be driven out of business."
Should AI employees have a say in military contracts?
Netflix Declined to Raise Offer for Warner Bros
Key Takeaways:
Netflix announced that it has declined to raise its offer for Warner Bros
Warner Bros has determined that a revised $111 billion all-cash acquisition offer from Paramount
The combined Paramount-WBD entity would face overlapping libraries, redundant networks, and difficult decisions about what gets cut
In one of the most dramatic reversals in Hollywood history, Netflix walked away from its $82.7 billion deal for Warner Bros. Discovery's studio and streaming assets, and did it in under two hours.
Here's the sequence: Paramount Skydance raised its all-cash bid for the entirety of WBD to $31 per share on Thursday, up from $30, with a $7 billion regulatory termination fee attached. Warner's board declared it a "superior offer" and gave Netflix four business days to respond. Netflix responded in two hours, declining to raise its bid. "This transaction was always a 'nice to have' at the right price, not a 'must have' at any price," co-CEOs Ted Sarandos and Greg Peters said. $NFLX jumped 10% after hours. $PSKY rose 5%. $WBD fell 2%.
If the deal closes, a shareholder vote is set for March 20, and regulators still need to sign off, Paramount Skydance will own HBO, CNN, Warner Bros., Harry Potter, Succession, The White Lotus, Top Gun, Titanic, and The Godfather under one roof.
Netflix walks away with a clean balance sheet, no DOJ antitrust exposure, and a$2.8 billion breakup fee. As strategic exits go, that's a very clean one. The political dimension is impossible to ignore: Trump backed Paramount, met with David Ellison at the White House in the days before the outcome, and Paramount settled a $16M lawsuit with Trump mid-process.
Agentic AI Takes Aim at SaaS
Getty Images
Key Takeaways:
The S&P 500 dropped over 13% in a single five-session stretch earlier in February, erasing more than $800 billion in market value before Nvidia's print partially stabilized sentiment.
IBM is down 22% in 2026 as Anthropic's new Cobol development tool directly targets one of IBM's core legacy enterprise strengths.
Analysts are divided on whether this is a structural repricing or a panic trade — the answer likely depends on which SaaS companies can prove they're data moats rather than just license vendors.
A phrase that started circulating on trading desks this month arrived center stage this week: "Software-mageddon." The thesis is that agentic AI: AI that can autonomously navigate complex enterprise software and execute multi-step workflows may make the traditional per-seat SaaS subscription model structurally obsolete. If a single AI agent can handle work previously requiring a dozen licensed users, why is anyone paying Salesforce, ServiceNow, or Cloudflare per employee? The market asked that question out loud. The S&P 500 Software & Services Index has fallen approximately 20% YTD. Salesforce, ServiceNow, Cloudflare, Zscaler, and IBM all fell hard before a partial mid-week recovery driven largely by Nvidia's blowout results. Nvidia's Jensen Huang pushed back directly, telling investors that the software selloff was an overreaction.
Analysts flagged ServiceNow and Salesforce as potential surprise recovery candidates, citing both companies' deep integration in enterprise data workflows, exactly where AI needs clean, unified data to function. Capital is rotating fast from "software AI replaces" to "hardware and infrastructure AI runs on."
Bottom Line
The thread connecting all five stories this week is the same one: power is being renegotiated in real time.
Who controls AI: the companies building it, or the governments deploying it? Who captures the entertainment future: the disciplined streamers or the legacy studios with political tailwinds? Who captures AI's economic upside: the chip makers printing record profits, or the software platforms quietly being displaced by the very technology they helped fuel?
None of these questions got a clean answer this week. All of them got louder. And that matters, because careers, capital, and policy are all moving in response to how these power shifts settle out.
The story of 2026 is being written right now.
Subscribe to ETF Alert for real-time market news.
We track the trends that move billions, before they hit mainstream headlines.