Q1 2026 AI Trends Review: A High-Fidelity Look at the State of Artificial Intelligence

The organizations that are really struggling here often lack CEOs who have presented a clear vision for the future of work in their organization and what is required and expected of their employees in that future of work.

Welcome to episode 208 of the Artificial Intelligence Show. I'm your host, Paul Roetzer, founder and CEO of SmarterX and Marketing AI Institute. I'm joined by my co-host and SmarterX Chief Content Officer, Mike Kaput. We have a special edition of the weekly podcast this week. I am currently out of the office spending time with my family, so rather than skipping a week, we decided to do a Q1 trends review.

Across the approximately 12 episodes we did in Q1, we covered roughly 150 different topics. Mike has curated those topics into 10 key trends that we're going to recap today. This is a retrospective of what has happened over the previous three months, and as Mike can attest, it is a lot. These quarterly trends are a great way to catch up on the rapid evolution of the industry.

10. The Model Release Frenzy

Mike Kaput: Q1 2026 might be in the running for one of the more compressed periods of frontier model releases so far in AI. The title of "state of the art" changed hands multiple times within weeks, and basically every major lab shipped something significant.

Anthropic released Claude Opus 4.6 in February. Their own reports and benchmarks revealed that it has saturated most automated evaluations to the point where the company plans to discontinue them. Opus 4.6 was followed weeks later by Claude Sonnet 4.6, which approached Opus-class capabilities and took the lead on the GPQA benchmark.

OpenAI countered with several releases, including GPT 5.3 Codex, a coding-focused model that logged 500,000 app downloads in its first week. In March, GPT 5.4 arrived with "Pro" and "Thinking" versions outperforming human professionals on economic benchmarks and setting a new record on the frontier math benchmark. OpenAI also shipped mini and nano variants of 5.4 later in the quarter.

Google released Gemini 3 Deep Think, which hit state of the art on the ARC-AGI 2 benchmark, followed quickly by Gemini 3.1 Pro. Additionally, xAI dropped Grok 4.2 in the same window. It is not only not slowing down; it might be speeding up.

Paul Roetzer: It sure seems like it. I think back a year or two ago when I wished ChatGPT and Gemini would just do the model picking for you. But I have found that the more we use these, the more important it becomes to know which model you are using. I actually like the ability to choose. When I’m doing a high-value strategic project or a no-code app-building project, I will test it in five or six different models.

This is pushing us toward the idea of having your own "evals" to evaluate these models. In the industry, they test for IQ, math, and biology. But as an individual or business leader, you need to think about the evals you can put in place to know which model is best for your specific use case. If you are in an AI-native company, you can use anything. Mike and I use Gemini, Claude, and ChatGPT daily. The challenge of which model is right becomes harder, and we’re going to talk a lot more in Q2 about helping organizations build these evals so they are understandable to a marketer, a salesperson, or an operations person.

Mike Kaput: I’ve heard anecdotally from people who don't follow this as closely as we do, and they are freaking out in a good way. They’ll text me saying, "Wait, Claude can do what?" If you haven't taken the most recent models for a spin outside of your daily driver, I’d highly recommend it.

Paul Roetzer: If you were just using one model and were unaware for three months that Claude Opus 4.6 or Sonnet 4 existed, you’ve missed a leap forward in capabilities. Sometimes it’s incremental, but sometimes we go through these three-month periods where model capabilities are dramatically better.

9. Big AI Becomes Big Lobbying

Mike Kaput: AI has been a first-tier political issue in Q1, and the story starting to capture the shift as we head toward US midterms is the sheer scale of money flowing into AI-focused political operations. Three pro-AI political groups are collectively spending nearly $300 million on US midterm ads, all pushing deregulation and an acceleration agenda.

The largest new entrant is Innovation Council Action, which has the blessing of David Sacks and plans to spend over $100 million. This group is led by a former White House Deputy Chief of Staff and has compiled a scorecard assessing how supportive lawmakers are of an AI agenda. Separately, "Leading the Future" has raised $50 million from donors including OpenAI President Greg Brockman, Palantir co-founder Joe Lonsdale, and Marc Andreessen. Brockman alone has contributed $50 million to this super PAC plus $25 million to a specific administration super PAC, making him one of the largest individual donors. Meta has also launched its own pro-AI super PAC effort expected to spend around $65 million on state-level races.

On the other side, Senator Bernie Sanders and Representative Alexandria Ocasio-Cortez introduced the AI Data Center Moratorium Act to pause all new data center construction nationwide until Congress passes federal AI legislation with protections for workers and the environment.

Paul Roetzer: It’s going to become a major issue in the midterms. I’m not convinced whether AI is a right or left-leaning issue yet. Jobs and energy affect everybody regardless of who you vote for. If you start losing tens of thousands of jobs this year, you are not going to be a fan of AI. If one party is cast as the accelerationist party at all costs, and one of those costs is the jobs of your family and friends, that can sway you politically. I think both sides are going to be very fluid in their messaging until they figure out what actually moves the needle on votes.

8. Anthropic vs. the U.S. Government

Mike Kaput: This is the biggest ongoing story of Q1. It began in February when Secretary of War Pete Hegseth issued an ultimatum demanding Anthropic grant the Pentagon full unrestricted access to its Claude models. Anthropic refused to remove its red lines against using Claude for mass domestic surveillance and fully autonomous weapons.

After some back and forth, Hegseth designated Anthropic a "supply chain risk." That same night, OpenAI announced it had signed an agreement with the Pentagon. In March, the Pentagon formalized the supply chain risk designation, making Anthropic the first American company to receive this. Federal agencies began ending their use of Anthropic products. Ironically, Claude continued powering Palantir’s Maven smart system, which reportedly identified over 1,000 targets in 24 hours during operations.

Anthropic filed two federal lawsuits to block the designation, warning that hundreds of millions in revenue was at risk. Microsoft and dozens of AI researchers filed supporting briefs. This past week, Federal Judge Rita Lin issued a preliminary injunction blocking the designation, writing that an American company cannot be branded an adversary for expressing disagreement with the government.

Paul Roetzer: We’re anxiously awaiting the appeal from the government, which will likely delay things and give people time to negotiate. There haven't been many leaks lately, which usually means negotiations are happening.

Anthropic is an infinitely fascinating company. They are shipping product like nobody I’ve ever seen—something like 50 releases in the first quarter. But there are also spin-offs around security. Just today, it looks like the Claude code base was published online due to a human error. It feels like things are moving too fast to put the genie back in the bottle.

I’m starting to think that human friction might be the saving grace. The models are getting so good so fast, yet doing anything in an enterprise is so slow. That might actually give us time to figure this all out. If every organization moved as fast as these frontier labs, we would be completely unprepared as a society.

7. The Rise of OpenClaw

Mike Kaput: OpenClaw is an open-source AI agent framework that allows autonomous agents to interact, execute complex tasks without human oversight, and even form communities. This burst into public consciousness with the release of a social network called Multibook, built on OpenClaw. It had millions of agents creating their own posts and engaging autonomously.

Andrej Karpathy called it "the most incredible sci-fi takeoff adjacent thing I’ve ever seen." Ethan Mollick mentioned that while it might be overhyped, it provided a visceral sense of what a takeoff scenario might look like. People were running entire businesses and job functions with it. The impact was significant enough that OpenClaw’s creator, Peter Steinberger, joined OpenAI in February to work on personal agents. Jensen Huang called it "the most important software release probably ever," and Meta acquired Multibook in March.

Paul Roetzer: I’ve been watching OpenClaw from the outside. We haven't built with it yet because of the risk and unknowns. But I recently listened to Claire Vo on a podcast, and she shared her story of going from a skeptic to a true believer. She built nine different agents through OpenClaw to run her sales and executive assistant functions.

You start to see the potential once the risk profile comes down. She has an agent named Sam that does all the outreach, daily analysis, and writes emails. Most enterprises won't touch this for a while, but it is a window into the near future. Once it becomes as easy to spin up an OpenClaw agent as it is to start a ChatGPT thread, it changes the dynamics of what work looks like.

6. Enterprise AI Adoption: The People Problem

Mike Kaput: Organizations are failing to generate significant ROI from AI, often not because of technology, but because of people. There are change management gaps, passive adopters, and legal bottlenecks. Our own AI Pulse survey found that 65% of listeners cited fear and resistance as a major challenge. Gallup research showed a widening gap between power users and everyone else, with 20% to 30% of employees actively resisting adoption.

Paul Roetzer: I would build on the "people problem" to say it starts with a leadership problem. Organizations that struggle often lack CEOs who have presented a clear vision for the future of work. If a CEO doesn't realize that reasoning has gotten good and agentic stuff is emerging, how can they present a vision?

Leaders need to say: "We expect you to take advantage of AI. We will provide licenses and training. In return, we expect you to improve your literacy and drive innovation." We see hundreds of brands where the marketing or sales team takes the initiative to get 50 licenses for a company of 70,000 people. When we ask if the CEO has a plan, the answer is almost always "No."

I think CEOs haven't had that "aha" moment where they realize the significance. They treat it as a technology problem and throw it to the CIO. But it’s a business problem. If an AI transformation isn't in the CEO's top three priorities, you have a problem.

5. SaaSpocalypse

Mike Kaput: In early February, $300 billion was erased from software and data stocks in just two days after Anthropic announced legal and sales plugins for Claude. LegalZoom dropped 20%, HubSpot is down 39% year-to-date, and ServiceNow dropped 27%. The market called this "SaaSpocalypse."

Frontier models are releasing features that eat into the core features of traditional SaaS companies. Furthermore, the traditional per-seat pricing models break down when one person with AI can do the work of ten. If headcount drops, seat count drops. Companies are struggling to figure out how to price AI that replaces labor rather than just augmenting a workflow.

Paul Roetzer: I haven't seen any answers yet, just more uncertainty. SaaS valuations are set on predictable multiples. When you realize people might not need as many seats or might build their own alternatives, it creates doubt. As a consumer, if I’m paying for software to do a thing, and the AI helps me do it faster, why am I paying extra for the AI? That’s the software company's cost problem, not mine.

I think we’re going to see turnover at the top of many software companies because navigating this will be difficult. As a buyer, I look at our tech stack and ask, "Is HubSpot going to enable this agentic future, or do I need a third party?" The fact that I even have to ask that isn't great for SaaS companies. You need to build an AI-forward team at all levels—marketing, sales, and success—because you’re going to deal with buyers who are more educated than your own staff.

4. Labs Pivot to AI Agents

Mike Kaput: Every major lab pivoted toward agentic capabilities and enterprise deployment in Q1. OpenAI is consolidating ChatGPT, their browser, and Codex into a desktop "super app" and is trying to build an autonomous AI research intern by September. Anthropic launched Claude Co-work, a more agentic system for non-technical workers. Microsoft restructured Copilot under Satya Nadella’s direct oversight. Even Andrej Karpathy released an auto-research agent.

Paul Roetzer: 2025 was the year of agent hype. We dealt with over-promising, but the reality is starting to set in. Agents are level three on OpenAI’s five levels of AI. They are starting to have their moment as they become more autonomous and reliable.

The security aspect is also huge. Anthropic recently had a leak of their Claude Code source code due to human error. It wasn't the model weights, but it highlights the risk. Dario Amodei has said only a few people at Anthropic have access to the weights because that is what foreign adversaries want. Espionage is a very real thing in this industry. These top engineers are high-value human targets.

3. AI-Driven Layoffs Go Mainstream

Mike Kaput: We haven't seen wide-scale AI layoffs yet, but we’ve seen more chatter. Atlassian cut 10% of its workforce and explicitly attributed it to the transition to the AI era. Block cut 4,000 employees and talked about how AI made them more efficient. The Uber CEO recently estimated AI will replace the work of 70% to 80% of humans within the decade. The PwC US CEO told the Financial Times that employees who think they can opt out of AI "are not going to be here that long."

Paul Roetzer: I wish this trend would go away, but I expect it to gain steam. The US hiring rate fell to 3.1% in February, the lowest since 2020. This "hiring recession" means companies are not adding anyone and are pausing hiring through attrition. I’ve yet to meet a CEO who wants to fire 20% of their staff, but there is going to be tremendous financial pressure to capture efficiency gains. We need to be talking more about answers and solutions because the labs building the tech aren't providing them.

2. Move 37 Moments

Mike Kaput: A "Move 37 moment" is when a professional realizes AI can match or exceed their expertise. We’re seeing more of these. Sam Altman noted Codex suggested features superior to his own team's ideas. Goldman Sachs is deploying Claude for trade accounting. A Polish mathematician reported a Move 37 moment after GPT-5.4 helped solve a problem that resisted conventional approaches. A New York Times quiz showed 54% of people preferred AI-written passages over famous authors.

Paul Roetzer: This was the premise of my MAICON keynote. I challenged people to think about the human side of AlphaGo—what happened to Lee Sedol when he realized the machine was better than him? We will all come to experience that moment.

I was at a basketball tournament recently, talking with a friend about what he’s doing with OpenClaw. I looked around the room at hundreds of people who have no idea what is going on. There’s a part of me that is envious of that ignorance. Once you know what these things are capable of, you can't turn it off. You start to look at the future of education and jobs differently.

1. The Vibe Shift

Mike Kaput: The conversation around AGI entered the public discourse this quarter. The single piece of content that captured this was Matt Shumer’s essay, "Something Big Is Happening," which was viewed 85 million times. He wrote about how the current moment feels like February 2020, right before COVID struck—a few people seeing signals that the world is about to change.

Paul Roetzer: You could feel it when we flipped the calendar to 2026. The online dialogue was different. We get hundreds of questions a month from our classes and events, and the questions have moved. People want to talk about building apps with no-code, messing with OpenClaw, and the political implications.

The gap between the "haves" and the "have-nots" is expanding. The people who are out ahead are moving further ahead and getting compounding value, while others are oblivious. I feel a greater sense of urgency every day because I see so many people who aren't aware yet.

Key Takeaways