Nvidia invests $2.1B in IREN; Akamai secures $1.8B AI deal

Nvidia will invest $2.1 billion in IREN Limited to deploy up to five gigawatts of AI infrastructure, sending the bitcoin miner turned data center operator’s stock up nearly 4% and marking the chipmaker’s most aggressive bet yet on co-located computing capacity. Separately, Akamai Technologies surged 20% after a leading U.S. frontier model provider committed $1.8 billion over seven years for its Cloud Infrastructure Services, while DeepSeek is raising more than $7 billion as the startup plots its first serious revenue push. The three moves, announced within hours of each other, underscore a market where capital is flooding into AI compute assets at unprecedented scale and where the gap between winners and losers is widening fast. CoreWeave slid 13% after second-quarter revenue guidance of $2.45 billion to $2.6 billion missed the $2.69 billion consensus, while Upwork tumbled 21% on a restructuring that cuts 24% of its workforce. The message from the week’s deal flow is clear: AI infrastructure is consuming tens of billions in committed capital, and the companies that can secure anchor customers and supply-chain partners are pulling away from those that cannot.

Nvidia’s $2.1 billion bet on co-located capacity

Nvidia’s investment in IREN is not a passive bet. The chipmaker will inject $2.1 billion into the Australian-listed company, which has pivoted from bitcoin mining to AI data centers, to support the deployment of up to five gigawatts of computing capacity. That is roughly the equivalent of five large-scale nuclear power plants’ worth of data center load, and it represents one of the largest single capacity commitments Nvidia has made to a partner outside its own supply chain. The structure gives Nvidia preferential access to GPU server racks deployed at IREN’s sites, effectively locking in long-term hosting capacity for its own systems and for customers that lease Nvidia-powered clusters. For IREN, the deal transforms its balance sheet: the company now carries a blue-chip strategic investor with deep pockets and an unquenchable demand for compute. The nearly 4% stock bump understates the significance. IREN’s market capitalization had already tripled over the past year as it shifted from mining to AI hosting. Nvidia’s $2.1 billion check validates that pivot and provides the capital to build at a scale that few independent operators can match. The partnership also signals that Nvidia is willing to put its own cash behind capacity expansion rather than relying solely on cloud hyperscalers or third-party colocation providers. The deal locks in hosting capacity for Nvidia’s own systems and for customers that lease Nvidia-powered clusters, giving the chipmaker a direct stake in the infrastructure layer.

Akamai’s $1.8 billion anchor tenant deal

Akamai’s $1.8 billion, seven-year commitment from a leading U.S. frontier model provider is the largest single customer deal in the company’s history and explains the 20% surge. The deal covers Akamai’s Cloud Infrastructure Services, a relatively new business line that competes with AWS, Google Cloud, and Microsoft Azure for AI training and inference workloads. For a company that built its reputation on content delivery and edge computing, this commitment validates the thesis that Akamai can win enterprise AI workloads by offering a distributed architecture that reduces latency and egress costs. The revenue impact is material: $1.8 billion spread over seven years averages roughly $257 million annually, which would represent approximately 8% of Akamai’s trailing twelve-month revenue. More important than the absolute number is the signal it sends to other potential customers. Akamai reported a first-quarter adjusted earnings beat and revenue that matched expectations, but the stock had been range-bound as investors waited for proof that the cloud infrastructure push would gain traction. The frontier model provider’s name has not been disclosed, but the size and duration of the commitment suggest a company that expects its compute needs to grow exponentially. Akamai now has an anchor tenant for its cloud infrastructure build-out, which reduces the risk of overbuilding and provides a referenceable customer for the next wave of AI workloads. The deal transforms Akamai’s revenue profile by adding a predictable, long-term stream that investors had not previously modeled into the stock.

The competitive reshuffle: winners and losers in AI infrastructure

The week’s deal flow draws a sharp line between companies that have secured AI infrastructure commitments and those that have not. IREN and Akamai are clear winners: both saw their stocks rise on deals that provide long-term revenue visibility and strategic validation. Gen Digital also popped 10% after current-quarter and full-year revenue guidance exceeded expectations, though its AI exposure is more indirect. The cybersecurity company benefits from increased cloud adoption and data center build-out. On the losing side, CoreWeave slid 13% after guiding second-quarter revenue to $2.45 billion–$2.6 billion, below the $2.69 billion consensus. CoreWeave had been one of the highest-flying AI infrastructure plays, but the miss shows that even with Nvidia’s backing and a massive GPU fleet, the company faces execution risk as it scales. Upwork tumbled 21% after announcing a restructuring that cuts 24% of its workforce, a sign that the freelance platform is struggling to monetize AI-related demand. Trade Desk fell 6% after an earnings miss, underscoring that not every tech company benefits from the AI boom. The divergence reflects a market that is rewarding companies with direct, contracted exposure to AI compute demand while punishing those that lack clear AI revenue streams or face execution challenges. The Verda-Compal partnership, announced via PRNewswire, adds another dimension: European AI cloud provider Verda (formerly DataCrunch) will receive next-generation GPU server systems from Compal Electronics to accelerate infrastructure build-out across Europe and APAC, signaling that the competitive landscape is becoming global.

Downstream effects on hyperscalers, fabs, and enterprise buyers

The Nvidia-IREN and Akamai deals create second-order effects across the AI supply chain. For hyperscalers like AWS, Google Cloud, and Microsoft Azure, the deals represent competitive pressure from below: IREN and Akamai are offering specialized AI infrastructure that can undercut the big three on price or latency for certain workloads. Nvidia’s direct investment in IREN also raises questions about whether the chipmaker is becoming a competitor to its own cloud customers. If Nvidia can finance and supply five gigawatts of capacity through IREN, it gains leverage in negotiations with AWS and Azure over GPU pricing and allocation. For chip fabs and packaging houses, the deals signal sustained demand for advanced silicon. TSMC, Samsung, and SK Hynix all benefit from the continued build-out, though the capacity commitments are multi-year and will not show up in orders for several quarters. Enterprise buyers face a more fragmented market: they can now choose between hyperscaler clouds, Akamai’s distributed infrastructure, IREN’s co-located clusters, or European providers like Verda. This fragmentation could drive down pricing for AI compute over time, but it also creates complexity in managing multi-cloud AI workloads. Compal Electronics, the Taiwanese manufacturer supplying Verda, is a bellwether for the hardware side of the AI build-out. Alan Chang, Compal’s representative, and Jorge Santos of Verda framed the partnership as a way to accelerate AI infrastructure development across Europe and APAC, with Verda using renewable energy to power its Finnish data centers. That green angle is a differentiator as enterprise buyers face pressure to reduce the carbon footprint of their AI operations.

What the deal flow signals about market direction

The simultaneous announcements from Nvidia, Akamai, and DeepSeek point to a market that is maturing from experimental AI spending to long-term infrastructure commitments. Nvidia’s $2.1 billion investment in IREN is the chipmaker’s most explicit bet that the AI compute build-out will require dedicated, co-located capacity beyond what hyperscalers can provide. Akamai’s $1.8 billion deal shows that non-hyperscaler cloud providers can win large, multi-year AI workloads if they offer the right architecture. DeepSeek’s more than $7 billion raise, the largest by an AI startup in recent quarters, indicates that investors still believe in the frontier model thesis despite concerns about monetization. Together, these moves show that the AI infrastructure build-out is entering a new phase characterized by larger deal sizes, longer commitment horizons, and more diverse provider ecosystems. The market is rewarding companies that can secure anchor customers and strategic investors while punishing those that miss guidance or fail to articulate a clear AI strategy. The Verda-Compal partnership adds a European dimension, hinting that the next wave of AI infrastructure investment will be global rather than concentrated in the U.S. and China. For investors, the key question is whether the current pace of capital deployment is sustainable or whether it will lead to overcapacity and consolidation within 18 to 24 months. The week's activity also shows that the competitive set for AI infrastructure is broadening: Verda and Compal's Europe-APAC partnership, powered by renewable energy, hints at a coming wave of regionally focused providers that can differentiate on geography and sustainability, adding further pressure on U.S. hyperscalers to justify their premium pricing.

The next 12 months will test whether the current wave of AI infrastructure investment produces the returns that backers expect. Nvidia’s $2.1 billion bet on IREN will need to translate into revenue-generating GPU clusters that customers actually lease, not just capacity that sits idle. Akamai’s $1.8 billion commitment will require the company to deliver on its cloud infrastructure promise at a time when hyperscalers are cutting prices to defend market share. DeepSeek’s more than $7 billion raise will force the startup to prove it can convert its technical reputation into recurring revenue from enterprise customers. The winners will be those that execute on build-out timelines, secure follow-on commitments, and manage the capital intensity of their businesses. The losers will be those that overbuild, underdeliver, or fail to differentiate in a market that is becoming crowded with well-funded competitors. The week’s deal flow provides a snapshot of a market in transition, one where the easy money has been raised and the hard work of building profitable AI businesses has begun.