🏀 Bots pick brackets
- 01Theme 1: AI Capability Is Advancing at a Non-Linear Pace
- 02Theme 2: Anthropic Is Winning the Enterprise AI Market
- 03Theme 3: The U.S. Government Is Formalizing AI as a Geopolitical Export Product
- 04Theme 4: OpenAI Pivoting to Enterprise
- 05Theme 5: The Agentic AI Readiness Gap Is a Real Enterprise Risk
1. Key Themes
Theme 1: AI Capability Is Advancing at a Non-Linear Pace
What was impossible for AI just 12 months ago is now routine. This year, all major chatbots successfully completed a full NCAA bracket — a task they failed at in 2025.
"It's an object lesson for a broader AI reality: What is impossible for AI at one point is often a trivial task just months later."
"The most important lesson is to not assume that because the technology can't do something today, it won't be able to in a few months."
Theme 2: Anthropic Is Winning the Enterprise AI Market
Despite being labeled a supply chain risk by the Pentagon, Anthropic's enterprise business has accelerated — not stalled — signaling that enterprise buyers are making independent judgments about AI vendors, separate from government posture.
"I've seen enough. Anthropic is the new default for businesses." — Ara Kharazian, Lead Economist, Ramp
"Anthropic's enterprise revenue has continued to grow since the designation was applied, according to Pitchbook. Meanwhile, Claude has surged to the top spot on U.S. app downloads."
Theme 3: The U.S. Government Is Formalizing AI as a Geopolitical Export Product
The Trump administration is treating AI infrastructure as a strategic export, bundling chips, models, and tools into deployable packages for allies — a significant shift in how AI is positioned in foreign policy.
"The American AI exports program is meant to bundle the infrastructure, tools and models into ready-to-deploy AI systems for allies and partners, and has been touted as a key part of the White House's AI policy goals."
"Starting April 1, industry-led groups will have 90 days to submit proposals for the White House's ambitious new AI exports program."
Theme 4: OpenAI Pivoting to Enterprise
OpenAI is reportedly shifting its core strategic focus away from consumers and toward business customers — a major realignment of one of the world's most prominent AI companies.
"OpenAI is reportedly making a major strategic shift to focus on business customers, a pivot from its consumer-focused approach." (via Wall Street Journal)
Theme 5: The Agentic AI Readiness Gap Is a Real Enterprise Risk
79% of security leaders are evaluating or deploying AI agents, yet only 13% feel extremely prepared. This gap — particularly around identity and credentials — is already producing measurable security incidents.
"67% of infrastructure and security leaders report heavy use of passwords, API keys and long-lived tokens — linked to a 20-point jump in AI incidents." (Teleport, 2026 Infrastructure Identity Survey)
2. Contrarian Perspectives
Government Pressure on Anthropic Has Backfired, Strengthening the Company
The conventional expectation would be that a Pentagon "supply chain risk" designation would devastate an AI company's enterprise pipeline. The opposite happened — the designation appears to have functioned almost as a trust signal, rallying industry partners and accelerating Anthropic's market position.
"The Trump administration tried to kneecap one of the world's most powerful AI companies. So far, it's just giving it a leg up."
"The designation looked like it could hit Anthropic where it hurts most: enterprise contracts, its core revenue driver. The opposite happened."
Supporting evidence: Claude surged to #1 on U.S. app downloads; multiple Fortune 500 companies confirmed no contract changes; Google's VP of Engineering publicly reaffirmed close partnership.
AI as a Research Assistant Beats AI as a Decision-Maker
The intuitive use case for AI in bracket prediction is to let it make all the picks. The article argues the higher-value application is keeping humans in the decision seat while offloading research.
"Filling out an entire bracket may not be the best way to use AI, anyway... why not use the AI to help with research rather than giving it all the fun and glory of making the picks."
Nvidia's Chip Revenue Projections Signal AI Infrastructure Spend Is Far From Peaking
Amid widespread debate about AI ROI and whether capex is overdone, Nvidia's CEO is projecting at least $1 trillion in revenue from its newest AI chips through 2027 — a figure that implies sustained, massive infrastructure investment ahead.
"Nvidia CEO Jensen Huang said yesterday that he expects the company to reap 'at least' $1 trillion in revenue from its newest AI chips through 2027."
3. Companies Identified
| Company | Description | Why Mentioned | Key Quote |
|---|---|---|---|
| Anthropic | AI safety-focused lab, maker of Claude | Centerpiece of story 2; enterprise revenue growing despite Pentagon designation | "Anthropic is raking in more revenue and attention than it did before President Trump and the Pentagon went after the company." |
| OpenAI | Maker of ChatGPT | Mentioned for bracket performance and reported enterprise pivot | "OpenAI is reportedly making a major strategic shift to focus on business customers, a pivot from its consumer-focused approach." |
| Tech giant, maker of Gemini | Highlighted as a continued Anthropic partner; also co-created NotebookLM/Yahoo Sports bracket tool | "We continue to work really, really closely with them, and we have no plans, in terms of the responsibility I have, to change anything." — Brian Delahunty, VP Engineering, Google | |
| Nvidia | Semiconductor company | Cited for massive chip revenue projections through 2027 | "Jensen Huang said yesterday that he expects the company to reap 'at least' $1 trillion in revenue from its newest AI chips through 2027." |
| Yahoo Sports | Sports media platform | Case study in AI-assisted consumer product; partnered with Google's NotebookLM for March Madness data | "Yahoo Sports has teamed up with Google's NotebookLM, creating separate notebooks filled with data on the men's and women's tournaments." |
| Ramp | Financial operations platform | Cited as a source of enterprise AI adoption data showing Anthropic surpassing OpenAI | "I've seen enough. Anthropic is the new default for businesses." — Ara Kharazian, Lead Economist |
| Teleport | Infrastructure identity/security company | Sponsor; cited for 2026 survey data on agentic AI readiness gap | "79% of security leaders are evaluating or deploying agents, but only 13% feel extremely prepared." |
4. People Identified
| Person | Description | Why Mentioned | Key Quote |
|---|---|---|---|
| Ina Fried | Axios AI+ reporter | Author of the March Madness AI bracket story; conducted hands-on testing of all major chatbots | "The last two years of using AI for March Madness reflects my overall experience with AI." |
| Madison Mills | Axios AI+ reporter | Author of the Anthropic enterprise story | Byline on story 2 |
| Ara Kharazian | Lead Economist, Ramp | Cited for market data showing Anthropic overtaking OpenAI in enterprise first contracts | "I've seen enough. Anthropic is the new default for businesses." |
| Brian Delahunty | VP of Engineering, Google | Publicly reaffirmed Google's continued partnership with Anthropic post-Pentagon designation | "We continue to work really, really closely with them, and we have no plans, in terms of the responsibility I have, to change anything." |
| Jensen Huang | CEO, Nvidia | Cited for projecting $1T+ in AI chip revenue through 2027 | "Expects the company to reap 'at least' $1 trillion in revenue from its newest AI chips through 2027." |
| Maxwell Millington | Axios colleague | Contributed tactical advice on how to best use AI for bracket research vs. decision-making | "Why not use the AI to help with research rather than giving it all the fun and glory of making the picks." |
5. Operating Insights
Use AI as a Research Layer, Not a Decision Layer
The most effective AI workflow isn't full automation — it's AI-augmented human judgment. Delegating research to AI while retaining decision authority produces better outcomes than handing over the full task.
"Filling out an entire bracket may not be the best way to use AI, anyway... why not use the AI to help with research rather than giving it all the fun and glory of making the picks."
Be Cautious About Granting Agentic Permissions — Even When Capability Is There
Even when AI agents are technically capable of taking autonomous actions (e.g., submitting forms, entering credentials), the author deliberately chose not to grant those permissions — a useful operating heuristic for enterprise AI deployment.
"I'm pretty sure at least some of the bots could have entered the brackets themselves, given sufficient time and permissions. But I'm hesitant to hand over credentials or agentic powers to my computer."
Evaluate AI Vendor Risk Independently of Government Signals
Enterprise procurement teams should not reflexively follow government designations when making AI vendor decisions. The Anthropic case shows that Fortune 500 companies are conducting independent assessments and holding course.
"Multiple Fortune 500 enterprises have told Axios they are not currently making any changes to their Anthropic contracts."
6. Overlooked Insights
AI Job Displacement Is Now a Tradeable Prediction Market
The newsletter briefly flags that AI's labor market impact is no longer just a debate topic — individuals can now view and bet on the probability that AI will replace specific jobs. This represents the intersection of two of the article's named "hottest trends in tech": AI and prediction markets.
"There's been endless debate about how AI will impact the labor market. Now, you can see the odds it'll impact your job — and even bet on it."
"Full-Stack" AI Export Bundling Opens a New Government Contracting Category
The U.S. AI exports program's "full-stack" framing — bundling infrastructure, tools, and models into ready-to-deploy packages for allied governments — creates a potentially large, structured procurement pathway that AI infrastructure companies should be tracking closely.
"The American AI exports program is meant to bundle the infrastructure, tools and models into ready-to-deploy AI systems for allies and partners, and has been touted as a key part of the White House's AI policy goals."