top of page

How Blackstone Is Using AI to Evaluate Deals

Blackstone’s CTO talks about how the company uses AI. This isn’t for the faint of heart.

For most companies in general, and in CRE in particular, making cutting edge use of artificial intelligence is still a distance off.

But Blackstone seems deep in, according to an interview that John Stecher, chief technology officer, gave to PE Hub. From what he said, and what he didn’t, there’s a lot to learn.

For example, he noted that what the company is using goes far beyond generative AI like ChatGPT. With the current hype, it’s easy to forget that the early days of artificial intelligence go back to the 1950s. There are many variations of the technology, including older approaches like rule-based and expert systems but also deep-learning systems that can keep improving their performance as they get more data to incorporate.

But sometimes the older and more limited types of AI are the best ones for a particular application. Other times, not. It is critical that a CRE company looking to incorporate AI has enough internal expertise or access to outsiders who can provide it as consultants to property match the tech to the application. Often, AI might not even be in the running.

On a large scale, making it work could be expensive. Stecher mentioned that Blackstone used one big resource it had access to, which was QTS Data Centers, a data center provider that the company acquired in 2021 for about $10 billion, including debt.

Stecher said that Blackstone thinks generative AI could improve dealmaking by letting analysts more quickly and easily go through large amounts of data and then cut the time needed to evaluate deals. However, there was an interesting jump from question to answer. PE Hub specifically mentioned ChatGPT while Stecher didn’t mention the specific product.

That was probably because ChatGPT doesn’t necessarily have access to all the data that a company might want to use. Generative AI would have to be trained on the type of data being analyzed. If a company has enough data in its possession to do the training, that would be possible. But if not, where does it come from? There may be little chance a software tool uses the same types of data you need. Provide your data to a third party and you need enough legal protection to be sure the information remains safe and available only for your use. For its investment process, Blackstone is building the product in-house. That is likely an expensive and complicated task.

Nov 7, 2023


Erik Sherman

bottom of page