Read OpenAI's latest internal memo about beating the competition — including Anthropic
OpenAI's chief revenue officer sent a four-page internal memo to staff on Sunday warning that competition has never been fiercer, naming Anthropic as a key threat. The company's strategy centers on locking users into multiple products at once and aggressively growing its enterpri...
OpenAI's top revenue executive just put the company's competitive anxiety in writing, and The Verge got the memo. Hayden Field, senior AI reporter at The Verge, broke the story on April 13, 2026, reporting that chief revenue officer Denise Dresser circulated a four-page memo to OpenAI employees on Sunday outlining a stark assessment of where the company stands and where it needs to go. The core message: building moats, capturing enterprise customers, and stopping users from bouncing to whichever AI model is trending this week.
Why This Matters
A CRO sending a company-wide memo admitting the market is the most competitive she has ever seen is not routine corporate communication, it is a signal that the pressure from Anthropic is landing harder than OpenAI's public posture suggests. The company is simultaneously telling investors it will have 30 gigawatts of compute capacity by 2030 while telling employees to focus on retention, which means OpenAI knows raw model quality is no longer enough to keep customers loyal. Enterprise lock-in has historically been how tech giants defend dominant positions, think Microsoft 365 or Salesforce's CRM ecosystem, and OpenAI is clearly reading from the same playbook. If Anthropic keeps shipping capable models, and it is, OpenAI's consumer lead starts looking thinner by the quarter.
Daily briefing from 50+ sources. Free, 5-minute read.
The Full Story
Denise Dresser, who recently took on expanded responsibilities following former COO Brad Lightcap's leave of absence, sent the memo on Sunday, April 13, 2026. According to the document viewed by The Verge, Dresser did not sugarcoat the situation. "The market is as competitive as I have ever seen it," she wrote, a line that stands out precisely because it came from an executive whose job is to project confidence to customers and partners.
The memo's central argument is that OpenAI has a retention problem rooted in how users currently behave. People are picking AI tools based on which model performs best for a specific task on a specific day. That kind of casual switching, driven by weekly benchmark results and viral Reddit posts, makes it nearly impossible to build durable revenue. Dresser's proposed fix is to get customers using more than one OpenAI product at the same time, so that switching costs rise and competitors face a higher bar to peel customers away.
That strategy is not new in tech, but it is new for OpenAI. The company built its reputation on a single product, ChatGPT, that became the fastest consumer application to reach 100 million users. But consumer adoption does not automatically translate into business sustainability, especially when competitors like Anthropic are actively signing enterprise partnerships and improving their own model, Claude, at a rapid clip. Dresser's memo signals that OpenAI's leadership understands the consumer crown and the enterprise throne are two different prizes.
Alongside the employee memo, OpenAI sent a separate document to investors making a pointed competitive comparison. OpenAI stated it plans to have 30 gigawatts of compute capacity by 2030. It then characterized Anthropic as likely reaching only 7 to 8 gigawatts of compute by the end of 2027, describing Anthropic as "operating on a meaningfully smaller curve." That framing is deliberate, positioning raw infrastructure investment as the decisive variable in long-term AI competition.
The timing matters too. Anthropic had recently announced a powerful new model being rolled out to select companies as part of a new cybersecurity initiative. OpenAI's dual memo strategy, one message to staff about urgency and one to investors about scale advantages, looks like a coordinated response to that specific momentum shift.
Key Details
- Denise Dresser sent the four-page memo to OpenAI employees on Sunday, April 13, 2026.
- Dresser recently absorbed duties previously held by former COO Brad Lightcap, who took a leave of absence.
- OpenAI's investor memo claims the company is targeting 30 gigawatts of compute capacity by 2030.
- OpenAI estimates Anthropic will reach approximately 7 to 8 gigawatts of compute by end of 2027.
- The investor memo describes Anthropic as "operating on a meaningfully smaller curve" on infrastructure.
- The employee memo was 4 pages and focused on enterprise growth and customer retention through product bundling.
- Anthropic, founded by Dario and Daniela Amodei along with other former OpenAI researchers, has been expanding Claude into enterprise cybersecurity partnerships.
What's Next
OpenAI's enterprise push means watch for new product bundles, tighter integrations between ChatGPT, its API offerings, and tools like Sora or Operator, rolled out to business customers in the second half of 2026. Anthropic will almost certainly respond by accelerating its own enterprise sales motion, particularly in regulated industries like finance and healthcare where Claude's safety-focused positioning resonates. The compute capacity race becomes the real story to track by 2027, when Anthropic's projected infrastructure investments come into sharper focus against OpenAI's stated 30-gigawatt target.
How This Compares
Anthropic's recent cybersecurity model announcement is the direct catalyst here, and it follows a pattern. Every time Anthropic ships something credible into a new vertical, OpenAI responds with messaging that emphasizes scale and infrastructure advantages rather than model benchmarks. That is a notable shift from 2023 and 2024, when OpenAI could simply point to GPT-4 as the obvious best-in-class option. Now the competition is close enough that internal memos are acknowledging vulnerability explicitly.
Compare this to Google's approach with Gemini and Google Workspace. Google moved early to bundle its AI models into products that hundreds of millions of enterprise users already pay for, which is exactly the moat strategy Dresser is now advocating. OpenAI is trying to build that bundled ecosystem from scratch, which is harder and slower than extending an existing one. Google had the enterprise relationships and the distribution. OpenAI has the brand and the consumer mindshare, but those are softer advantages.
The investor memo's compute comparison also echoes the messaging strategy Microsoft used against Google in the early cloud wars, claiming infrastructure scale as a proxy for long-term inevitability. It works if the numbers are real and if customers believe compute translates directly into model quality, which is not always true. Open-source models from Meta and others have consistently punched above their compute weight, a variable neither memo appears to address. Follow the latest AI news to see how that dynamic plays out over the next 12 months.
FAQ
Q: What is OpenAI's moat strategy and why does it matter? A: A moat strategy means making it harder for customers to leave by getting them to rely on multiple connected products at once. For OpenAI, that means pushing businesses to use ChatGPT, its API, and other tools together so switching to Anthropic or another competitor requires untangling several integrations rather than just swapping one app.
Q: How does OpenAI's compute capacity compare to Anthropic's? A: OpenAI told investors it plans to reach 30 gigawatts of compute capacity by 2030. The company estimates Anthropic will have roughly 7 to 8 gigawatts by the end of 2027. More compute generally means a company can train larger, more capable models, though it does not guarantee better products in every use case.
Q: Why is enterprise business more valuable than consumer subscriptions for OpenAI? A: Enterprise customers sign longer contracts, integrate AI tools deeper into their workflows, and are much harder to move off a platform once embedded. A business that builds internal tools on top of OpenAI's API is far less likely to switch providers than an individual user who can download a different chatbot app in 30 seconds.
OpenAI's internal candor about competitive pressure is actually a healthy sign that the company is treating this fight seriously rather than dismissing rivals with press releases. Whether the enterprise bundling strategy executes cleanly or stalls on product integration challenges will determine whether Dresser's memo reads like smart foresight or an early warning sign in a few years. Subscribe to the AI Agents Daily weekly newsletter for daily updates on AI agents, tools, and automation.
Get stories like this daily
Free briefing. Curated from 50+ sources. 5-minute read every morning.
