Wall Street had been bracing itself for more defiance from Meta CEO Mark Zuckerberg over his continued determination to invest heavily in AI infrastructure, but few could have imagined that what they would be presented with was a 91% drop in cash flow to fund this spending.

In Q2, the firm reported free cash flow of $784 million, down from $8.55 billion a year ago as CapEx soared to $31.08 billion, with more to come - the company now expects total 2026 spend to be as high as $169 billion, up from a range of $115 billion to $135 billion set out at the start of the year.

Zuckerberg remained defiant about the need for such levels of spending:

In terms of the different opportunities and how we think about the compute, overall, a substantial amount of the compute goes towards training models to be a leading lab. And I think that that's an important investment. But then the rest of it goes towards a set of different products and revenue opportunities, which spans from optimizing and improving our core business to building new consumer products that we're releasing soon to the API, to the business agents work, to the developer tools work on the road map that I alluded to and then also the opportunity to sell compute directly where we have quite a number of offers at a meaningful premium over what we paid for the compute.

That last remark appears to refer to the news that Meta is set to rent AI computing power to rival Anthropic in a $10 billion deal. It's relatively small amount, but from a 'political' perspective, it would generate fresh revenue for Meta and - perhaps! - calm nerves from investors around the firm's ongoing aggressive infrastructure spend.

According to CFO Susan Li:

Our AI investments are going to play a significant role in delivering on this vision, including the expansion of LLM-based content understanding to develop a deeper understanding of posts and creators that people value to capture user interest more precisely and respond more quickly to what they care about in the moment and using AI to surface high-quality, fresh and trending content and reduce the share of low-quality content.

Second, we're continuing to improve our data infrastructure to allow our models to train on more data and leverage that data more effectively. We're adding more detail to how we describe content that users have engaged with in the past and enriching past user interaction sequences with more granular content. That allows our models to more precisely learn which engagements are more or less valuable to users. We're continuing to scale up both the length of user interaction sequences we use during training as well as the complexity of our model architectures across Facebook and Instagram to take advantage of the larger data sets. And then we've already made significant strides leveraging LLMs for content understanding.

She added:

Our longer-term capacity strategy aims to give us the flexibility to continue growing compute in 2028 and beyond by laying down data center and network foundations to accommodate future server decisions. The long-lived nature of these assets inherently provides the flexibility that will make it possible to adjust our investment to the pace of AI adoption. In addition, we have been making strategic investments in areas like our internal custom silicon effort, which will provide long-term strategic flexibility and supply chain leverage. This will be helpful in driving better returns on those long-term investments. Finally, we believe that overall industry capacity is going to remain tight for the foreseeable future. As we've said earlier, we strongly believe that the models, consumer experiences and enterprise offerings that we are building will be the best and highest ROI use of our infrastructure.

She added:

We expect that remaining nimble about these opportunities will help us fund our build-out more efficiently while preserving our strategic flexibility to have the compute when we need it and provide us multiple pathways to generate returns on expected on invested capital. In funding these infrastructure investments, the strength of our balance sheet gives us the ability to attract capital from a wide range of markets to supplement the cash flow generated by our business.

Meta currently has 32 data centers around the world in operation or under construction, 28 of them in the US, but this is not enough to keep up with demand.  In a recent development, the firm announced a joint venture with Blackrock, the world’s largest investment firm, to develop and own a $14 billion data-center campus in El Paso, Texas.

Meta last year unveiled plans to invest $1.5 billion to build the El Paso data center and boosted its spending commitment to $10 billion in March. BlackRock will own 80% of the venture, while Meta will own the remaining 20% and lease the entire campus, which will have 1 gigawatt of compute capacity. The companies expect to begin bringing the compute capacity online in 2028.

Zuckerberg said of the deal:

Building the infrastructure for superintelligence is key to making sure the benefits of this technology are distributed to everyone.

Elsewhere Meta earlier thsi month scaled up the size and scope of a Lousiana data center project to deliver five gigawatts of compute capacity, bumping up the cost from the original estimate of $27 billion to more than $50 billion.

We are really a full-stack technology company. We built our own data centers, our own infrastructure, our own chips, our own low-level software. 

I’ve said before that one thing that Zuckerberg does need to be given credit for is his defiance of Wall Street short-termists screeching to be shown the money from AI, not told about the amount that needs to go into investing in the tech’s essential foundations.

So far, he’s holding the line.