OpenAI highlights Amazon partnership, cites Microsoft client access limitations

The newly named revenue head at OpenAI, Denise Dresser, sent an email to employees this Sunday, highlighting the company's partnership with Amazon as vital for boosting its enterprise business, while mentioning the restrictions with its longstanding arrangement with Microsoft.

Dresser's memo arrives following two months since Amazon revealed plans to invest up to $50 billion in OpenAI as part of a strategic partnership.

Since 2019, Microsoft, Amazon's major rival in cloud services, has invested over $13 billion in OpenAI, supporting the firm long before it initiated the generative AI surge with the release of ChatGPT.

Amazon Web Services, leading the cloud infrastructure field, offers companies access to all major AI models, including those from OpenAI, through a platform called Bedrock.

"Our Microsoft collaboration has been central to our achievements. However, it has also restricted our ability to engage with enterprises where they are located — for many, that's Bedrock," Dresser noted in the memo, seen by CNBC.

"Since we disclosed the partnership at February's end, inbound interest from our customers for this service has been absolutely overwhelming."

OpenAI is keen on capturing market share in enterprise sectors, where competitor Anthropic's Claude model has asserted itself as a market leader, with Google Gemini also fiercely competing.

Claude's gaining traction was the most discussed topic at the AI industry gathering HumanX in San Francisco last week, with Arvind Jain, CEO of enterprise AI firm Glean, describing it as "Claude mania."

OpenAI and Anthropic are both striving to convince investors of their expanding market presence as they prepare for potential IPOs later this year.

The enterprise domain is crucial as companies are heavily investing in AI, a movement that's trimmed the valuation of public software companies, increasingly viewed as vulnerable.

OpenAI, on the other hand, was valued at over $850 billion in its recent fundraising round last March, a month after investors valued Anthropic at $380 billion.

Dresser shared with CNBC earlier this month that OpenAI's enterprise segment accounts for 40% of the company's revenue, and it’s "set to reach equality" with its consumer business by year-end.

In the Sunday email, Dresser mentioned the market can occasionally be "chaotic, changeable, and distracting," and encouraged staff to focus on engaging with customers. She pointed out that Anthropic's approach is built on "fear, restriction, and the notion that a select group of elites should govern AI," while OpenAI’s "positive message" will win in the long term.

Dresser indicated Anthropic has also made a "tactical mistake by not securing adequate compute," echoing statements OpenAI made in a separate memo to investors on Thursday. The firm said Anthropic is "operating below a significant threshold," with its own progression "materially ahead and widening." Anthropic revealed a partnership with Google and Broadcom for "multiple gigawatts" of compute earlier this month.

For OpenAI, its relationship with Microsoft remains critical and strategic, but signs of tension have emerged as the partners expand into each other's domains.

By mid-2024, Microsoft listed OpenAI among its competitors in its annual report, a roster including fellow megacap peers like Amazon, Apple, Google, and Meta for years.

OpenAI is increasingly relying on other cloud providers, such as CoreWeave, Google, and Oracle, for capacity, and last year, Microsoft started publicly examining a proprietary AI model that could boost enhancements to its Copilot assistant for consumers.

OpenAI appointed Dresser, previously Slack's CEO and a seasoned Salesforce executive, as the chief revenue officer in December.

She recently broadened her role to incorporate Brad Lightcap’s commercial duties as he shifted from the operating chief to a new position focused on "special projects."