Key Highlights
- Shares of Meta Platforms (META) declined 5% following the announcement of a strategic AI infrastructure partnership.
- Firmus Technologies, based in Australia, will provide Meta with GPU computing power across Southeast Asian markets.
- This partnership extends Meta’s current deployment of Nvidia GB300 NVL72 systems at Firmus’s Melbourne facility.
- Meta’s capital expenditures reached $31.1 billion in the second quarter, with free cash flow dropping 91%.
- Analysts maintain a Strong Buy consensus on META stock, projecting approximately 12% potential gains.
Shares of Meta Platforms (META) experienced a 5% decline after the social media giant announced a strategic AI computing collaboration with Firmus Technologies, an Australian infrastructure provider.
The partnership arrangement calls for Firmus to deliver graphics processing unit computing capabilities to Meta. These resources will originate from a series of AI facilities currently under construction throughout Southeast Asia.
The tech giant intends to leverage this computing infrastructure for artificial intelligence research initiatives, model creation, and training operations. Neither party revealed the financial details of the arrangement.
Firmus specializes in constructing and managing AI facilities designed for enterprise-scale computational demands. The company delivers high-performance computing solutions, cloud services, and AI-as-a-service platforms to corporate and governmental clients.
Expansion of Current Collaboration
This latest agreement builds upon the existing relationship between Firmus and Meta in the Australian market. Currently, Meta operates Nvidia GB300 NVL72 computing systems at Firmus’s data facility located in Melbourne.
According to Firmus, this Melbourne installation represents the most extensive deployment of Nvidia Blackwell Ultra infrastructure throughout the Southern Hemisphere. The upcoming Southeast Asian installations will utilize Nvidia’s DSX AI Factory architecture.
This architecture will integrate seamlessly with Firmus’s proprietary HyperCube thermal management solution. Given the substantial heat output from AI processors, effective cooling infrastructure plays a critical role in operational efficiency and performance optimization.
The liquid-cooling technology developed by Firmus is engineered to reduce power consumption and minimize operational expenses. This configuration may enable Meta to expand its AI training capabilities more efficiently.
Significant Capital Investment in AI Infrastructure
This Firmus partnership comes as Meta continues to ramp up its capital allocation toward AI-related infrastructure. During the second quarter, Meta’s capital expenditures totaled $31.1 billion.
Throughout that same timeframe, free cash flow plummeted 91% to just $784 million. The company has indicated potential capital spending of up to $145 billion for the current fiscal year.
From Firmus’s perspective, securing this agreement represents a significant milestone in its Asia-Pacific growth strategy. Partnering with a technology leader of Meta’s stature strengthens its position in the infrastructure sector.
The arrangement also provides Firmus with more stable revenue streams from GPU capacity leasing. This financial stability could prove beneficial as the company prepares for its anticipated $5 billion initial public offering on the Australian exchange.
Should this IPO proceed successfully, it would potentially become the second-largest public listing in Australian history. Firmus’s business model centers on developing and operating AI facilities that enable large-scale computing operations for enterprise customers.
Financial analysts continue to assign a Strong Buy consensus rating to META stock. This rating reflects 39 Buy recommendations and six Hold recommendations issued within the last three months.
The consensus price target among Wall Street analysts stands at $798.74 per share. This projection suggests potential upside of approximately 12% from present trading levels.
Shares of Nvidia (NVDA) saw a 2% increase following this announcement. Nvidia’s semiconductor technology remains fundamental to the AI infrastructure strategies of both organizations.



