OpenAI is considered a leading company in the field of artificial intelligence (AI). However, it seems that ChatGPT's parent company is lacking in terms of security and customer service. At least this is what the vendors of Amazon's cloud computing subsidiary, Amazon Web Services (AWS), claim. And they do so based on a series of guidelines provided by the company itself to its vendors.
These internal sales guidelines, which Business Insider has seen , are what AWS salespeople must follow when dealing with customers who have purchased or are about to purchase OpenAI products.
According to these internal sales directives, sellers at Amazon 's cloud computing subsidiary are expected to highlight OpenAI's lack of access to third-party AI models and the weakness of the Sam Altman-led company's corporate contracts. The goal is to emphasize AWS's robustness in running the cloud infrastructure behind the most popular AI services.
“For generative AI workloads, AWS will primarily compete with Microsoft’s Azure OpenAI service, OpenAI (directly), and Google Cloud’s Generative AI at wordpress web design agency Vertex AI,” read, for example, the documents, which were brought to light by Business Insider . They urge vendors to move beyond the current hype around AI chatbots and instead focus on the foundational models that power them and the cloud infrastructure needed to “securely build, integrate, deploy, and manage their own generative AI applications using first-party data.”
The guidelines, which cover the period from late 2023 to spring 2024, suggest that Amazon is keen to neutralise the growth of its rivals in the field of AI, in particular OpenAI, as far as possible . The US multinational has, after all, been investing in this technology for many years, which did not become mainstream until ChatGPT came on the scene .
Yet, despite being a pioneering company in the field of AI, over the past two years Amazon has witnessed how OpenAI, as well as Microsoft and Google, have stolen almost all the spotlight when it comes to trendy technology.
While Amazon's AI models are probably not the most popular, the company led by Andy Jassy is also benefiting enormously from the "boom" caused by this technology, which ultimately needs to be supported by exceptionally robust cloud infrastructures.
AWS is looking to get OpenAI customers to migrate to its AI-focused services
Speaking to Business Insider, an Amazon spokesperson stressed that much of AWS's growth is, in fact, anchored in its AI services, which could generate "billions of dollars" in revenue by 2024.
“It’s still early days for generative AI, and with so many companies offering a variety of services, we’re working to equip our sales colleagues with the information they need to help customers understand why AWS is the best, easiest, and highest-performing place to build generative AI applications . Parsing the language as anything other than that or mischaracterizing our leadership position is misguided speculation,” the Amazon spokesperson stressed.
By highlighting OpenAI’s weaknesses, AWS clearly wants to get those customers who got started in the AI world with ChatGPT’s parent company to migrate to the Amazon subsidiary’s services . AWS is particularly keen for OpenAI customers to use Bedrock , its own AI tool based on partnerships with providers such as Anthropic, Meta and Cohere. And Amazon’s internal documents state, not in vain, that Anthropic’s Claude model (a startup in which the multinational led by Andy Jassy has invested 4 billion dollars in recent months) would have surpassed OpenAI’s GPT models in terms of “intelligence, precision, speed and cost.”
AWS’s efforts to try to attract OpenAI customers to its own AI services are already bearing fruit . Four AWS customers (HUDstats, Arcanum AI, Forcura, and Experiae) that initially used OpenAI products have now switched to AI services from Amazon’s cloud computing division.
AWS is not the only one targeting OpenAI. Its internal sales guidelines also call into question the capabilities of Microsoft and Google in the field of AI. Google is specifically criticised, for example, for its limitations in terms of the number of foundation models offered. And Microsoft is also underscoring the limited availability of its Maia AI chip.