Nvidia said it will invest $3 billion in Meta Platforms' recent equity offering as part of an expanded partnership focused on Meta's Muse AI models. The companies will work together to improve inference and training performance for Muse, while Meta plans to expand its use of Nvidia infrastructure in its data centers, including Vera Rubin and Vera CPUs to power AI agents, and deploy Nvidia networking technology. The move follows positive user response to Muse agents, with the Muse agent app becoming one of the most downloaded apps within a week of launch. Separately, Meta upsized its common stock offering to $25 billion from $22.5 billion, and with a depository share offering, the total could reach $37.5 billion, priced at $625 per common share and $50 per depository share. The raise is aimed mainly at funding Meta's AI infrastructure buildout, and Meta had also said it plans a separate $60 billion non-voting share offering in 2027.
Summary written by Upside AI from the original article. Read the source for full details.
Read the original on Yahoo Finance ↗
Keywords
METANVDAartificial intelligenceAI investmentdata centers
Open a keyword to see other news on the same topic.
Related tickers and likely impact
META
▲ Likely positiveMeta Platforms
Relevance: medium
| Market | US stocks |
|---|---|
| Sector | Communication Services |
| Industry | Interactive Media & Services |
| Market cap | $1.7T |
| Close on 10/26/2026 | 630.47 |
| Change that session | -1.89% |
NVDA
▲ Likely positiveNvidia
Relevance: medium
| Market | US stocks |
|---|---|
| Sector | Information Technology |
| Industry | Semiconductors |
| Market cap | $5.1T |
| Close on 10/26/2026 | 215.98 |
| Change that session | +2.38% |
Impact is scored by AI from the headline and may be wrong. It is not investment advice.
Related news
- Meta, SpaceX, Anthropic stoke the AI funding frenzy. What it means for our stocks
Meta plans to raise up to $95 billion in three steps: $22.5 billion of Class A common stock priced at $642.62 per share, $12.5 billion of preferred shares, and $60 billion from an at-the-market offering of Class C shares no sooner than next year. Elon Musk said SpaceX closed a $40 billion debt financing for GPUs, while Anthropic filed an S-1 for its IPO with a possible valuation of up to $2 trillion; the amount to be raised is undisclosed but could be tens of billions of dollars, if not $100 billion. CoreWeave has fallen 60% from its year-to-date high of nearly $139 on May 6 and hit a 52-week low just above $51. Musk said debt markets have nearly shut for companies without substantial revenue independent of AI data centers, and that the 200+ neoclouds face consolidation. The article argues that hyperscalers with other businesses, such as Microsoft, Amazon and Alphabet, could benefit as weaker neocloud providers struggle.
- SpaceX, Meta Pursue $135 Billion for AI Infrastructure as Musk Warns of Neocloud Funding Crunch
- AI chip stocks bounce back as hyperscaler earnings loom, META equity raise in focus
AI chip stocks rebounded on October 19, with the Nasdaq up 1.3%, NVIDIA gaining 1.38% and Meta Platforms slipping 0.55%. Evercore ISI sees the recent pullback as coming ahead of a strong hyperscaler earnings season, and NVIDIA expects top five hyperscaler capex to reach nearly $800 billion in 2026. Meta guided 2026 capex to $130 billion to $145 billion, while second-quarter free cash flow fell to $784 million from $8.55 billion a year earlier. Analysts expect a Meta stock sale, and according to the article, the market is reportedly comfortable with about $30 billion, though Meta has not announced an offering. The article notes H100 pricing has held up, NVIDIA says it has supply for about 70% of demand, and GPU-backed loans are seen as the go-to option over unsecured debt at 15%.
- Zuckerberg spent years building on Apple’s terms. His $100 billion-plus AI bet is about owning everything
Mark Zuckerberg says he approved Meta’s AI spending of as much as $200 billion before there was any proof it would work. He describes the company’s “full-stack” strategy of building its own data centers, chips and infrastructure, shaped partly by past constraints from Apple and Google on mobile platforms. A one-gigawatt AI data center costs $50 billion to $60 billion to build and about $15 billion a year to operate, and Meta is also building a cloud business to sell AI computing power. Meta’s personal AI agent Muse is free, with paid plans at $20 and $100 a month, and Zuckerberg suggested the company could take a small cut of commerce transactions. Advertising made up about 98% of Meta’s revenue last quarter, so it remains unclear whether AI agents that compare prices and buy on users’ behalf will reduce the ad impressions that fund the business.
- Meta Reports $825 Billion in Future Spending Commitments
Meta said it has agreed to spend about $825 billion on future leases and contracts, many tied to AI computing power and infrastructure. Leases not yet on its balance sheet total $355.7 billion, covering data center and energy infrastructure starting through 2036. As of Sept. 30, the company separately had $454.5 billion in non-cancellable contracts, largely for third-party cloud computing, data center energy and Reality Labs hardware. Meta is also obligated to buy up to $14.7 billion of cloud capacity over five years. The total compares with almost $700 billion reported in July, as Meta boosts its capex plans and creates a new class of nonvoting stock.
- Meta Seeks About $95 Billion for AI Buildout as Nvidia, AMD and Broadcom Consider Investing in Their Own Customer
Meta Platforms (NASDAQ:META) plans to raise about $95 billion to expand its artificial intelligence operations. Three of its biggest chip suppliers, Nvidia (NASDAQ:NVDA), AMD (NASDAQ:AMD) and Broadcom, are considering buying shares in their own customer.
Related markets
Latest news
- Unframe Removes Another Barrier Between AI Capabilities and Adoption
Unframe, an applied AI lab, announced its availability on the AWS, Google Cloud and Microsoft marketplaces. Enterprises can buy directly through their existing cloud accounts, applying committed cloud spend toward Unframe and following established approval workflows. Private offers let Unframe set tailored pricing, contract terms and usage parameters for individual customers. Global VP of Revenue Jacquelyn Goldberg said the move removes another barrier between identifying high-value AI use cases and putting them into production. The company says it delivers production-ready AI solutions in days, with measurable business impact in weeks.
- Elon Musk: The only way to solve America's $40,000 billion debt
In a conversation with biographer Walter Isaacson on The Builders podcast, Elon Musk said AI and robots may be the only way to solve America's public debt. After 130 days as a special government employee, he concluded that spending cuts are not a workable fix because of political opposition, and instead pointed to sharp productivity gains from automation. US public debt hit a record $40,000 billion in August 2026. Musk also predicted that in a multi-planet civilization, currency could be measured in physical quantities such as watts and tons rather than traditional money. The article notes that higher productivity does not guarantee lower federal debt, since the government would still need to collect enough revenue through taxes or other sources.
- Vingroup's biggest-ever project, worth $18 billion: over 12,000 workers, VND 10 billion spent on fuel and oil daily
Vingroup's Ha Long Xanh urban complex (Vinhomes Global Gate Ha Long) in Quang Ninh broke ground on December 19, 2025, with total investment of about VND 456,639 billion (nearly $18 billion), the largest project in the company's history. The construction site employed about 12,799 workers in the first quarter of 2026 and 12,155 in the second quarter, with fuel and oil costs averaging around VND 10 billion per day. The project is recruiting more than 6,000 additional workers, including over 4,000 dump truck drivers, and expects to need about 50,000 workers and engineers over the long term. Some first phases target commercial operation from 2028 across more than 6,200 hectares along the Hanoi-Hai Phong-Quang Ninh expressway corridor. Vingroup is also building a 350 km/h high-speed rail line from Hanoi to Ha Long, which the article describes as a source of value synergy for the project.
- From Golden Population to Golden Workforce: The Human Dimension of Vietnam's AI Vision
The article says Vietnam aims to become a high-income country by 2045, with effective AI use, especially developing workforce skills, seen as a key priority. A 2026 study commissioned by Amazon Web Services (AWS) found 26% of Vietnamese businesses have adopted AI, up from 18% a year earlier, but 61% of these are still in the testing stage, while 87% of surveyed employers say reskilling current staff is important to their AI strategy. The national AI human resources development program, approved in August 2026, targets basic AI skills for about 10 million workers by 2030. Dr. Tan Hong Ming of NUS Business School argues that leadership and redesigning work matter as much as technical skills, since training staff on AI tools may not raise productivity if processes and management do not change. He also warns that AI could remove the struggles that build judgment, so NUS Business School is piloting an AI-powered business simulation for MBA and Executive MBA students.
- Does Jensen Huang Know Something Wall Street Doesn't? Nvidia's CEO Sees No Signs of Memory Demand Fading by 2029
Memory-chip stocks Micron, SK Hynix and Samsung have surged this year on strong AI data-center demand, with Micron up about 172%, Samsung 66% and SK Hynix roughly 81%. Nvidia CEO Jensen Huang told CNBC he does not expect memory demand to fade even by 2029, a longer horizon than most analysts have assumed. The recent sell-off followed news that China's CXMT plans to roll out a new architecture packing more memory cells into the same space by year-end, along with a Korean article warning of a flood of new supply. Still, Micron and SK Hynix have signed multi-year deals of three to five years with large customers, and SK Hynix plans to triple its wafer capacity by 2034. Counterpoint Research forecasts DRAM prices will fall 23% in 2028 versus 2027 but remain 230% above 2025 levels. The author suggests investors avoid overweighting this volatile group.
- Solana bug reports now cost 1 SOL as AI floods security teams with spam
Firedancer, the Solana software developed by Jump Crypto, now requires researchers to permanently burn 1 SOL (about $100) for each bug report; the cost is nonrefundable even if a report is invalid or duplicates an earlier finding, and reports submitted without a burn are ineligible for rewards. The change follows a monthlong $1 million audit competition that drew 809 reports from 162 researchers but produced only 127 unique security findings, with nearly half of valid submissions duplicating earlier reports. Jump believes AI helped discover or write nearly every submission, which slowed engineers' validation work. To offset this, Jump is temporarily doubling payouts, and tripling them for Alpenglow, Solana's new voting system, where critical findings can qualify for up to $300,000.
Get the briefing on Telegram or Zalo
Join the Upside Telegram channel or Zalo group for the Vietnam stocks and gold briefing at 8:00 and 19:00 (Vietnam time) on trading days, plus a weekly wrap on Saturday morning.