Two Go-To-Market motions, one market.

The generative AI market is the most closely watched Go-To-Market experiment in modern B2B history. Two companies — OpenAI and Anthropic, the makers of ChatGPT and Claude — have emerged as the dominant commercial players, and they have taken almost opposite routes to get there. Both have worked. Neither is finished.

For most of the early era, the narrative focused on models, benchmarks, and safety research. By 2026, a different story has taken hold: the motion matters as much as the model. The way each company has gone to market tells B2B operators more about the next five years of enterprise revenue building than any product announcement will.

The two motions are distinct enough to describe in a sentence each.

Or, more provocatively: OpenAI sells after adoption. Anthropic sells before it. Both are legitimate, and both have proven out at scale.

Motion A: product-led growth at unprecedented scale

OpenAI's GTM machine is the most effective top-of-funnel in B2B technology history. Over 900 million weekly active users on ChatGPT. A consumer flywheel that turns free trials into Plus subscribers, Plus subscribers into team seats, and team seats into enterprise contracts. The sales role is to meet demand that the product has already created.

The advantages are real and hard to replicate. Inbound pipeline is virtually free. Brand recognition is category-defining. Every viral moment on social media — every "look what ChatGPT did" clip — is a zero-cost top-of-funnel event that competitors cannot buy their way into.

The tradeoffs are structural. Consumer subscriptions subsidise compute rather than justify it. Converting a free user into a paid seat is a smaller commercial event than signing a seven-figure enterprise contract from day one. That's why OpenAI has spent 2026 deliberately re-weighting towards enterprise — enterprise now accounts for roughly 40% of revenue, up from around 30% twelve months earlier, and the company has publicly said it is tightening its enterprise focus.

Read this correctly: this is not a motion failing. It is a motion maturing. The PLG engine that built the brand is being paired with an enterprise motion to monetise it.

Motion B: enterprise-first, sales-led from day one

Anthropic built its GTM engine the way traditional B2B SaaS companies do — but faster, and at AI-era scale. Target accounts. Paid pilots. Multi-year contracts. Cloud partnerships with AWS and Google Cloud as distribution. No consumer phase to speak of.

The strengths here are the ones every SaaS operator already recognises. High ACV. Signed multi-year commitments. Predictable expansion. Retention economics that compound. And a trust posture — Constitutional AI, published safety research, HIPAA-ready healthcare offerings — that functions as a procurement asset in regulated industries.

The tradeoffs are equally familiar. Enterprise sales cycles are slow. Pipelines take quarters to build, not weeks. There is no viral distribution, no zero-cost funnel. For the first two years of Anthropic's commercial life, that looked like a disadvantage.

By early 2026, it looked like compounding. Anthropic's annualised revenue passed $30 billion in April, with roughly 80% coming from business customers. Over 1,000 enterprises now spend more than $1M/year on Claude. Eight of the Fortune 10 are customers. The Ramp data on first-time enterprise AI spend swung decisively toward Anthropic in Q1 2026 — not because the motion changed, but because it finally reached scale.

Motion mechanics explain how revenue is captured. The business models underneath explain what is being captured in the first place — and the shape of each company's revenue stack reveals just how differently they've constructed their engines.

Notice the architectural difference. OpenAI is building a portfolio — consumer, enterprise, API, and new bets, each a potentially independent revenue line. Anthropic is building a wedge — three deeply-integrated streams, all pointing at the same enterprise buyer. That contrast shows up even more sharply at the product level.

Read these four contrasts together and you stop seeing two AI companies. You start seeing two B2B archetypes — the horizontal platform and the vertical wedge — rendered in an industry where both can reach tens of billions in revenue inside four years. Which is, of course, why the lesson travels so well.

Why both motions have worked

The temptation is to declare a winner. That reading misses the point. The two motions were optimal for different windows of the same market.

OpenAI's PLG motion was the correct motion for 2022–2024 — a discovery phase in which buyers did not know what they wanted, budgets were experimental, and viral consumer adoption was the fastest route to legitimacy. Selling after adoption made sense because adoption was the thing that needed to be proven first.

Anthropic's enterprise motion is the correct motion for 2025–2026 — a deployment phase in which the same buyers now want re-architecture, multi-year commitments, and vendor accountability. Selling before adoption makes sense because strategic deployment cannot be bottom-up; a procurement committee cannot viral-adopt its way into a seven-figure contract.

The most useful read: the market shifted from experimentation to deployment, and the motion that fit the new phase happened to be the one Anthropic was already running. OpenAI is now building that motion in parallel. Anthropic is building consumer surface area. Both will converge toward multi-motion GTM by 2027.

Five lessons for B2B operators

Drawn from both playbooks, not one.

The operator's takeaway

OpenAI and Anthropic are not a case study in one company beating another. They are a case study in how radically different GTM motions can both be correct — at different moments, for different buyers, with different constraints.

The mistake is to look at the 2026 numbers and conclude one motion is "better". The lesson is to understand why each worked when it did, and to apply that diagnostic to your own business. Which phase of the market are you in? Which buyer is actually signing your contracts? Which motion fits the capital you have, not the capital you wish you had?

Get those answers right, and the motion almost builds itself.