Yazılar

Global companies pour billions into AI infrastructure with mega-deals

/in /tarafından

A wave of multi-billion dollar investments is reshaping the AI landscape as chipmakers, cloud providers, and tech giants race to secure computing power for next-generation artificial intelligence. The surge follows OpenAI’s launch of ChatGPT in 2022, which sparked unprecedented demand for GPUs, cloud infrastructure, and data centers.

Key deals fueling the AI boom:

  • Nvidia & OpenAI – Nvidia to invest up to $100B in OpenAI and supply advanced AI chips, cementing its dominance in the AI ecosystem.

  • Nvidia & Intel – Nvidia invests $5B for a ~4% stake in Intel.

  • Oracle & Meta – In talks on a $20B cloud deal to boost Meta’s AI compute.

  • Oracle & OpenAI – Landmark deal worth $300B over five years for OpenAI to buy Oracle cloud capacity.

  • CoreWeave & Nvidia – $6.3B order ensuring Nvidia-backed startup CoreWeave absorbs unused cloud demand.

  • Nebius Group & Microsoft – $17.4B, five-year GPU deal to bolster Microsoft’s infrastructure.

  • Meta & Google – Six-year, $10B cloud agreement signed in August.

  • Intel & SoftBank – SoftBank injects $2B into Intel, becoming a top-10 shareholder.

  • Tesla & Samsung – $16.5B chip supply deal for Tesla’s next-gen AI6 chip, produced in Texas.

  • Meta & Scale AI – Meta takes 49% stake ($14.3B) in Scale AI, elevating CEO Alexandr Wang’s role in Meta’s AI strategy.

  • Google & Windsurf – $2.4B licensing deal for AI code generation tech.

  • CoreWeave & OpenAI – $11.9B, five-year contract signed before CoreWeave’s IPO.

  • Stargate Datacenter Project – Joint venture by SoftBank, OpenAI, Oracle, backed by U.S. President Donald Trump, with up to $500B in AI infrastructure funding.

  • Amazon & Anthropic – Amazon doubles down with a total $4B investment in Anthropic, developer of the Claude chatbot.

Why it matters:

  • Capital intensity: AI development is now measured in hundreds of billions, with infrastructure demands rivaling traditional energy projects.

  • Strategic alliances: Tech giants are securing long-term chip and cloud capacity to avoid bottlenecks.

  • Geopolitical edge: Governments, particularly the U.S., are encouraging private-public mega-projects like Stargate to keep ahead in the AI race.

The investment frenzy highlights a simple truth: the future of AI hinges not just on algorithms, but on who controls the world’s computing power.

U.S. Considers Annual Chip Supply Approvals for Samsung and SK Hynix China Plants

/in /tarafından

The United States is weighing a proposal to require Samsung Electronics and SK Hynix to seek annual approvals for shipping chipmaking equipment and supplies to their China-based factories, Bloomberg reported Monday, citing people familiar with the matter.

The plan, presented by the U.S. Commerce Department to Korean officials last week, would replace the current validated end user (VEU) designations that granted the chipmakers indefinite export authorizations. Those designations are set to expire at the end of 2025.

Under the draft proposal, Samsung and SK Hynix would need yearly approval for specific quantities of restricted tools and materials, adding regulatory steps but ensuring their Chinese fabs can keep operating. The companies are among the largest foreign chipmakers with plants in China, supplying memory chips vital to global electronics.

Reactions in Seoul were mixed—officials expressed relief that a framework for continued operations remains, but concern over the added bureaucratic burden and potential supply chain uncertainties.

The move comes against the backdrop of intensifying U.S.-China semiconductor tensions. Since 2022, Washington has imposed sweeping export controls to curb Beijing’s chip and AI capabilities. The Biden administration had granted waivers to Samsung, SK Hynix, and TSMC to soften the blow to allied companies, but the Trump administration has pushed for tighter oversight.

The situation is further complicated by political strain: Washington revoked prior waivers days after former South Korean President Lee Jae Myung—who advocated a more balanced U.S.-China stance—signed a defense and investment deal with Trump. Recent U.S. immigration raids on Korean firms’ American subsidiaries have also fueled friction.

Trump to Hit Semiconductor Imports with Tariffs Unless Firms Build in U.S.

/in /tarafından

President Donald Trump announced Thursday that his administration will impose tariffs on semiconductor imports from companies that do not move production to the United States. Speaking ahead of a dinner with top tech CEOs, Trump said the tariffs would be “fairly substantial” but would not apply to companies already investing in U.S. manufacturing.

Trump framed the move as part of his broader strategy of using tariffs to pressure foreign companies and governments to shift production and jobs into the U.S. “If they are not coming in, there is a tariff,” he said. He singled out Apple CEO Tim Cook, noting that Apple’s $600 billion commitment to domestic investment puts it “in pretty good shape.”

The policy comes as global chipmakers respond to U.S. pressure. Taiwan’s TSMC, South Korea’s Samsung, and SK Hynix have all announced major U.S. semiconductor plant investments. Trump had previously floated a 100% tariff on imported chips but said exemptions would apply for companies producing or planning facilities inside the country.

The announcement underscores Trump’s second-term emphasis on tariffs as a cornerstone of economic and foreign policy, a tool he has wielded to renegotiate trade terms and gain leverage in geopolitical disputes. However, legal challenges loom: lower courts have invalidated parts of his earlier tariff regime, and the administration has asked the Supreme Court to uphold the sweeping emergency powers used to justify them.