OpenAI appears to be developing something more substantial than its next incremental update. An internal memo, obtained by The Verge, offers a glimpse into a project codenamed 'Spud'. Denise Dresser, Chief Revenue Officer, has already described 'Spud' as a "significant step in building intelligence for the next generation of workflows," which is intended to make all of the company's products "significantly better." The ambition is high: to achieve a level of understanding of intentions and dependencies that yields more predictable outcomes. Amidst growing demand for solutions that can do more than just generate text – that can independently operate tools in real business scenarios – this development seems a stark necessity. OpenAI itself is not complaining about a lack of interest; the company is recording growth in multi-year contracts worth nine figures, with its own production capacity being the only bottleneck.
The critical element here is the transition from a simple product to a fully-fledged platform. According to Dresser, the market has long moved beyond the stage of "just chat." Clients demand autonomous agents capable of integrating into workflows and functioning in the real world. It is for this purpose that OpenAI is creating the 'Frontier' platform, aiming to make it the "primary venue for enterprise agents." The logic is simple and time-tested: the better the models, the more valuable the platform; the deeper the integration, the harder it is to abandon; the more processes flow through the system, the stronger the grip. Thus, OpenAI is clearly targeting operational infrastructure, not just tool provision. The new alliance with Amazon, granting access to markets outside of Microsoft's usual orbit, appears to be a logical extension of this strategy. However, transforming 'Frontier' into the "default platform" is a non-trivial task, especially with Meta and Amazon also building something, albeit without tangible results yet, though their presentations are ready.
The most intriguing part of the memo is a direct attack on competitor Anthropic. OpenAI accuses them of artificially inflating revenues by $8 billion, asserting that Anthropic counts partnership payouts from Amazon and Google on a gross, rather than net, basis. Against the backdrop of declarations about the advent of a "mature phase" for enterprise AI, where not only power but also integration is crucial, such an accusation seems like an attempt not only to discredit a rival but also to position itself as a benchmark of honesty and maturity. However, it cannot be ruled out that this is simply sharp competitive maneuvering or, conversely, an attempt to cover its own tracks by diverting attention from the details of the 'Spud' project itself.
OpenAI is betting on the evolution of AI from naive assistants to autonomous agents, promising us 'Spud' and 'Frontier.' If this plan is realized, the enterprise solutions market could indeed be reshaped. The open confrontation with Anthropic is a clear sign that the AI infrastructure market is heating up. For you, as a CEO, this means it is worth paying close attention to how exactly OpenAI will embody its ambitions. Promises are good, but real working agents capable of delivering measurable benefits are what will truly change the game. For now, we are seeing only declarations and accusations, and the true potential of 'Spud' remains to be evaluated.