Meta’s been trying hard to position itself as an AI leader this year.
Part of this was a hiring/poaching spree of AI experts. And the other part was planned AI infrastructure build-outs that didn’t sound realistic, until OpenAI outdid them with even bigger and even less realistic plans.
Despite throwing a lot of money around, it appears the company is slowly coming to terms with reality. According to Reuters, sources speaking to the Information, which first reported the deal, revealed that Meta had talks with Alphabet’s Google Cloud about the possibility of using its Gemini models to improve the Facebook parent’s ad business.
Meta already has a six-year cloud computing deal with Google, according to Reuters sources. In addition to the deal with Google, valued at more than $10 billion, Meta is in discussion with Oracle for a cloud computing agreement worth approximately $20 billion.
Building data center infrastructure is simply too expensive, and we will see below how Meta solved its supercomputer problem.
Image source: David Paul Morris/Bloomberg via Getty Images
Bank of America expects Meta to beat Wall Street Q3 estimates
Bank of America analysts Justin Post and his team provided an outlook for Meta (META) stock ahead of the Q3 earnings report, scheduled for October 29.
Analysts expect Q3 revenue of $50.0 billion and earnings per share (EPS) of $7.30. These estimates exceed Wall Street’s projections for revenue of $49.5 billion and EPS of $6.69.
The team provided an estimate for Q4 revenue of $58.8 billion and EPS of $8.90, which is again higher than Wall Street’s estimates of $57.3 billion and $8.12, respectively.
Related: Analysts update Oracle stock forecast
Assuming Q3 at the high end of the guidance, the analysts expect Q4 guidance in the range of $55.5 billion to $59 billion (up 15-22% year over year).
Given the AI build-out, they expect continued fiscal year 2025 investments and think Meta could narrow the calendar year 2025 expense range to $115 billion to $117 billion (from $114 billion to $118 billion). They also believe it could raise the low end of the capital expenditure range by $2 billion and guidance to $68 billion to $72 billion.
For 2026, analysts expect revenue of $237 billion and EPS of $32.09, compared to Wall Street estimates of $229 billion and EPS of $30.03.
Post reiterated a buy rating and the target price of $900, based on a 27 multiple of his estimate for GAAP earnings per share for 2026, plus net cash. “On a total company basis, including Metaverse investments, our valuation is at a slight premium to S&P 500, given Meta’s higher growth rate and AI opportunity,” he said. “Historically, Meta has traded at an average premium of 3pts to S&P 500.”
Analysts noted downside risk factors for Meta:
- Decline in user activity from competition
- Privacy or data issues impacting revenue generation
- Potential for Wall Street to assign a negative value to Metaverse (Reality Labs)
- New regulations that impact monetization
Meta secures funds for Hyperion data center
When Meta CEO Mark Zuckerberg in July announced several AI infrastructure projects, including the 5GW supercomputer Hyperion, he said: “We’re also going to invest hundreds of billions of dollars into compute to build superintelligence. We have the capital from our business to do this.”
“Something” must have happened to those hundreds of billions of dollars Zuckerberg mentioned, because Meta instead made a deal with Blue Owl Capital to finance the Hyperion project in a joint venture.
Related: Analysts reset Broadcom stock forecast, revamp industry outlook
Funds managed by Blue Owl Capital will own an 80% interest in the joint venture, and a portion of the capital raised by Blue Owl will be funded through debt issued to PIMCO and select other bond investors via a private securities offering.
Both companies will fund their respective pro rata share of the approximately $27 billion in total development costs for the buildings and long-lived power, cooling, and connectivity infrastructure at the campus. Meta will retain a 20% equity stake.
Meta superintelligence unit takes a hit
When Meta went on a hiring/poaching spree for its superintelligence unit, I wrote:
“I don’t see this strategy working out for Meta. Hiring a bunch of big names who are probably used to being ‘team leaders’ and paying them ridiculous sums that ‘regular’ team members won’t get can’t be good for that team’s morale. An additional problem is that executives and team leaders who are used to being in charge may not be great team players.”
More AI Stocks:
- More police agencies adopt AI, raising efficiency — and concern
- How AI Could Monitor Brain Health and Find Dementia Sooner
- ChatGPT latest user perk sparks controversy
- Bezos’s AI bubble warning backed by big names in banking
According to Axios, Meta is laying off about 600 positions in its superintelligence lab. In an internal company memo, Meta Chief AI Officer Alexandr Wang wrote:
“By reducing the size of our team, fewer conversations will be required to make a decision, and each person will be more load-bearing and have more scope and impact.”
The layoffs target the company’s FAIR AI research, product-related AI, and AI infrastructure units, and will not affect the TBD Lab unit.
Related: Analyst resets Intel stock forecast before earnings
