OpenAI And Anthropic Are Scaling Sales Teams Into A Market Where Nobody Has To Actually Sell

OpenAI scaled its enterprise sales team from 10 to 500 people in under two years. Anthropic is building fast behind it, targeting $20 billion to $26 billion in revenue for 2026. Both companies are hiring aggressively into what might be the easiest enterprise sales environment in the history of software. That is not obviously good news.

Ben Horowitz put it plainly in a recent conversation published by Sequoia Capital: "Right now with OpenAI and Anthropic, everybody wants to buy AI. They're already predisposed to buy." His point was not a compliment. It was a warning about what that environment produces in a sales organization.

The Order-Taker Problem

In May 2023, during a quarterly earnings call, Cloudflare CEO Matthew Prince acknowledged what many enterprise software leaders quietly know. His exact words: many on the sales team "succeeded largely by just taking orders" because the product "solved real problems that every big company faces." Deals arrived with minimal effort. Then macro conditions shifted, and Cloudflare cut approximately 100 sales staff who had collectively contributed around 4% of new business.

That is the structural problem with hot markets. When product demand is so intense that buyers come to you, the distribution of actual sales skill across your team becomes impossible to assess. Mediocre reps crush quota. They get promoted into leadership. Nobody knows they cannot sell (or don't have a playbook for the industry) until inbound stops.

A post by sales industry commentator TechSalesGuy circulating widely on X captured it directly: "My brother works at one of the hottest companies in his space. Inbound everywhere. Quota gets dismantled. Him: 'Half the time, I'm just taking orders.' Now think about the 500 reps OpenAI just hired or Anthropic's team scaling fast behind them. Inbound so hot customers are lined up begging to see the latest and greatest. Those reps are going to crush quota then cash in again in 2-3 years at a new company. The job market rewards logos first, then skills."

Why Hard Sells Build Real Salespeople

Horowitz returns repeatedly to PTC, the 1990s CAD/CAM company, as his reference case. The product, in his telling, was not great. It was difficult to install, difficult to demo, and difficult to justify to skeptical procurement teams. That difficulty forced discipline: systematic account mapping, competitive displacement strategies, airtight technical cases built deal by deal.

His benchmark hire for Databricks was Ron Gabrisko, who came from a company selling secure FTP as a public-market product. Horowitz's logic: if you can make quota selling that, you can sell anything. The adversity is the credential.

He applied the same filter at Okta. Two candidates, one enthusiastic, one who said "let me talk to your customers first." Horowitz told the CEO to hire the second. A rep who qualifies the hiring company is demonstrating exactly the instinct that closes hard deals.

What History Suggests About Market Turns

This pattern has precedent. Salesforce's growth stalled in 2001 as the dot-com contraction forced real qualification discipline on reps who had been riding a wave. Facebook's advertising business hit turbulence around 2012 as advertisers demanded measurable ROI rather than reach. AWS faced its first serious competitive pressure around 2015 as Azure and Google Cloud began offering viable enterprise alternatives with aggressive pricing and migration support.

In each case, the companies that managed the transition well had GTM teams that had learned to sell under pressure before the pressure arrived. The ones that struggled were staffed with people who had been optimized for processing warm inbound.

The AI market is not immune to this dynamic. An a16z enterprise survey published in February 2026 found that 78% of enterprise CIOs are already using OpenAI in production, with Anthropic at 44% and rising. Wallet share is consolidating. As the market matures, three viable options will exist across most enterprise categories. Enterprises will care about pricing, support, vendor risk, and integration depth. That is a different sales conversation than the current one.

The VC Exposure

For investors, this is a material question. OpenAI is planning to nearly double its workforce to 8,000 employees by end of 2026, with sales and customer-facing roles a significant component of that expansion. Anthropic is building toward a $20 billion to $26 billion revenue target for 2026, supported by large-scale partnership deals with Deloitte, Cognizant, and Snowflake that effectively outsource the implementation layer.

Both approaches embed cost and organizational complexity that is difficult to unwind. And Reuters reported in March 2026 that the two companies are competing actively for private equity joint ventures, with OpenAI offering guaranteed minimum returns of 17.5% to attract PE partners. The enterprise land-grab logic assumes continued inbound momentum. If that assumption is wrong, the cost structure does not automatically adjust.

There is also a compounding problem. When the same reps who built their careers in an all-inbound environment move into sales leadership, they tend to hire in their own image and build systems optimized for the conditions they know. The talent selection error propagates up the org chart.

How To Hire Around This

Horowitz's framework for building GTM teams that can actually sell is essentially a value-investor approach to human capital. Ignore the logo as good logos do not build exceptional sales skills. Find the person the market has systematically underpriced because they spent their career at companies nobody has heard of.

The signals: someone who had to fight for every deal because the dominant competitor was already embedded in the account. Team members who built pipeline from scratch when inbound was not there, who learned to systematically displace entrenched vendors, not just respond to RFPs that were already won.

Those reps rarely show up with OpenAI or Anthropic on their resume. They show up with companies that made their craft necessary. That is precisely why they are available and precisely why they are worth hiring.

The real test for both companies is not whether they can staff up during a boom but whether the people they are hiring now can hold enterprise accounts and expand them when the market is no longer doing their job for them.

At the moment they are also mostly selling AI novelty and it's potential. Once the market matures and settles, there will be far less demos and far more questions on guiderails, halucinations and unit economics.

When inbound dries up, and it always does, the companies that built real sales organizations and at least somwhat healthy unit economics will have a durable advantage. The ones that staffed order-takers and selling AI's omnipotence with negative margins will discover the problem and face it at scale.