Questions About OpenAI’s Spending to the Fore

The Wall Street Journal reported a major news item this morning, one that brings into focus questions about massive spending on artificial intelligence.  According to their exclusive report, OpenAI is missing key revenue and user targets, and reporting that sits “CFO and board have questioned the wisdom of massive data-center spending in the face of slowing growth.”  After a huge rally for semiconductor stocks and ahead of major hyperscaler earnings releases tomorrow, this is highly inconvenient, to say the least.  I asked an “expert” about it.

Let me preface the rest of the piece by saying that I never use AI to write on my behalf.  It is an excellent editor, though, and I rely on company-provided Claude to check my writing for grammar, punctuation, spelling, and sloppy cut/pastes before I submit my writing for publication.  It can also be an excellent research assistant, grabbing and analyzing data faster than I otherwise could.  Today, however, you will find that I relied on AI more than usual for my analysis – something that will be obvious as you read on.

My use of AI is admittedly modest, but a report like today’s once again brings forth questions we raised months ago, some of which seemed to be taken seriously by the market earlier this year.  First, do the productivity gains at present and in the near future justify the massive spending on data centers and the hardware to power them?   Second, how is a buzzy startup with little in the way of profits and heavily negative cash flow going to pay for the massive commitments that were announced with great fanfare?  Those questions are the crux of today’s modest angst in tech shares.

Thus, I decided to ask an “expert” about the WSJ report’s accuracy and about how it views OpenAI’s path to profitability and its likelihood of meeting its spending targets. The “expert” was ChatGPT.  

If you ask me about matters pertaining to Interactive Brokers’ financials and growth prospects, I need to respectfully decline.  I can comment freely about market matters, but not about my own company.  This is standard practice at most companies.  ChatGPT seems to have no such qualms, which I discovered after a simple initial query:

Considering that OpenAI is a private company, there is no need for them to confirm or deny their profitability to anyone other than investors and other key counterparts.  Yet, ChatGPT seemed to confirm the public opinion on that matter.  That led me to a follow-up, asking whether the bot could confirm the company’s ability to meet its spending targets.

Dare I say that ChatGPT is less sanguine about its own prospects and its ability to continue to spend as aggressively as many investors seem to hope?  There is clearly a path to profitability, which the answers lay out nicely, but there is no shortage of potential pitfalls that can interfere with the market’s assumptions for progress along that path.  Presumably these are the types of risks that would be outlined in a prospectus if OpenAI does indeed go public, as is widely rumored, but they are quite real, not just boilerplate legalese.

Finally, I asked OpenAI to comment directly about the veracity of the WSJ reporting.  Again, anyone other than a limited number of senior managers at a normal company would be prohibited from commenting about a story of this nature, but OpenAI’s most visible “spokesperson” had plenty to say:

The OpenAI story is meaningful for investors, though it is hardly the end-all, be-all for the AI story.  Sure, the start of the current bull run almost perfectly coincides with the launch of ChatGPT in November 2022, and much of the growth that we have seen over that period is directly related to enthusiasm for, and the massive build-out to support generative artificial intelligence.  But ChatGPT is hardly the only game in town at this point.  Anthropic is now perceived as a worthy, if not better, competitor among private AI providers, and the “four horsemen of AI” who all report earnings after tomorrow’s close – Alphabet (GOOG, GOOGL), Amazon (AMZN), Meta Platforms (META), Microsoft (MSFT) – all have formidable models and seemingly bottomless financial resources. 

Investors will be listening closely tomorrow, not only for their current EPS and future revenue, earnings, and cash flow guidance, but also for their spending plans.  The recent 18-day winning streak in the Philadelphia Semiconductor Index (SOX), which took the index about 40% higher, was predicated on billions of dollars continuing to flow into data centers and the hardware needed to power them.  If there is any whiff of reduced spending from those behemoths, much of that recent rally can come undone quickly; if the spending stream is likely to continue, so should enthusiasm for the affected stocks.  We will also get some clues as to whether the reported problems are general to the industry or specific to OpenAI.  There are plenty of examples of companies that seemed to be early conquerors of the internet that found themselves surpassed as technological progress marched on.  Could OpenAI simply be analogous to AOL or the like?  Possibly, but even if it is a lone also-ran, it would still negatively impact the companies that are expecting its largesse.  Stay tuned tomorrow!