1. NVIDIA Invests Up to $2.1B in IREN, Pursues 5GW AI Factory Build-Out — The Dawn of GPU Ecosystem Vertical Integration
• Core Source
“NVIDIA and IREN plan to support the deployment of up to 5GW of NVIDIA DSX-based AI infrastructure across IREN’s global data center pipeline over the long term”
“IREN granted NVIDIA a 5-year right to purchase up to 30 million common shares at an exercise price of $70 per share, giving NVIDIA the right to invest up to $2.1 billion upon fulfillment of certain conditions and regulatory approvals”
“AI Cloud Supply Agreement: a 5-year AI cloud contract worth $3.4 billion was signed to support NVIDIA’s internal workloads.”
“The rights vest incrementally each time GPUs are actually installed in data centers, and fully vest when a total of 600,000 GPUs have been deployed.”
“Current contracted ARR totals $3.1 billion, comprising Microsoft ($1.9 billion), NVIDIA ($700 million), and Prince George ($500 million).”
• Expected Impact
NVIDIA’s decision to invest up to $2.1 billion in IREN and sign a $3.4 billion cloud contract signals that NVIDIA is transitioning from a GPU design-and-manufacturing company into a vertically integrated model that directly controls data center power, land, and operational infrastructure.
The investment structure is particularly noteworthy. NVIDIA’s equity rights vest incrementally as GPUs are actually deployed — a structure that is not a passive financial bet, but rather an institutional expression of NVIDIA’s intent to directly manage the entire GPU supply, deployment, and monetization cycle in tandem with its partner’s infrastructure. NVIDIA has secured investment rights in a similar fashion with Corning’s fiber optic business, and is expanding investments across the AI ecosystem including OpenAI and Marvell.
On the demand side, IREN has already secured $3.1 billion in contracted ARR from Microsoft, NVIDIA, and Prince George, demonstrating that a structure in which hyperscalers proactively deploy capital into infrastructure companies to secure GPU access is becoming the new normal across the AI market. IREN’s trajectory — from Bitcoin mining infrastructure to a strategic partner of NVIDIA anchored by a 5GW power portfolio — also illustrates that companies holding power and land are emerging as the new core asset holders in the AI supply chain, which could catalyze a structural re-rating of the data center energy infrastructure sector broadly.
2. CoreWeave 1Q26 Revenue +112%, Backlog Surpasses $99.4B — The Explosion of AI Cloud Infrastructure Demand and Oligopolization of Long-Term Contract Suppliers
• Core Source
“Q1 revenue was $2.1 billion, up 112% year-over-year and 32% quarter-over-quarter.”
“Revenue backlog expanded to $99.4 billion, up approximately 50% quarter-over-quarter and approximately 4x year-over-year.”
“Active power surpassed 1GW, and contracted power expanded to over 3.5GW through the addition of approximately 400MW during the quarter. The company maintains its target of securing 8GW or more by 2030.”
“New additions include Anthropic, expanded contracts with Meta, and Jane Street. 10 customers now each have contracts exceeding $1 billion.”
“Management noted that the majority of 2026 production capacity has already been sold.”
• Expected Impact
CoreWeave’s Q1 results carry structural implications that matter more than the numbers alone. Revenue growing 112% year-over-year while a $99.4 billion backlog looms ahead means that multiple years of future revenue are already locked in through contracts — evidence that the current growth rate reflects a structural demand regime, not a short-term cycle.
The nature of customer concentration is worth examining. The fact that 10 customers including Anthropic, Meta, and Jane Street each hold contracts exceeding $1 billion indicates that the AI cloud infrastructure market is restructuring around long-term exclusive contract relationships between a small number of validated suppliers and large customers. This creates extremely high barriers to entry for late movers, as infrastructure takes years to build and major customers are already locking in their providers.
On the supply side, management’s statement that 2026 production capacity is already sold out is the key signal. This means GPU and power supply chain bottlenecks are materializing in real time, and a “first-mover advantage” dynamic is in effect — companies that secured infrastructure first hold a structural edge. Despite a staggering FY2026 CapEx guidance of $31–35 billion, completing an $8.5 billion investment-grade term loan at sub-6% interest rates signals that the market is assigning a significant premium to this structural positioning.
3. Coherent Posts Record Backlog, Announces 4x InP Capacity Expansion — Rising as the New Dominant Force in AI Optical Infrastructure Supply Chain
• Core Source
“Order backlog reached an all-time high”
“A new, aggressive expansion plan was announced to quadruple InP capacity from current levels by end of 2027.”
“OCS TAM raised to $4B, double the prior estimate; CPO TAM cited at $15B, reflecting growing confidence in larger revenue opportunities.”
“The ramp-up pace of 1.6T optical modules was described as an ‘astonishing pace'”
“Net debt ratio dropped sharply from 1.7x last quarter to 0.5x this quarter, supported by asset divestitures and NVIDIA’s $2B strategic investment.”
• Expected Impact
The most important signals from Coherent’s earnings are the simultaneous announcement of a TAM upgrade and capacity expansion. Raising the OCS TAM to $4 billion — double the prior figure — and citing a CPO TAM of $15 billion means a direct supplier is confirming that the optical infrastructure market itself is structurally expanding alongside AI infrastructure growth.
On the supply side, the pace of 6-inch InP wafer conversion is the key competitive variable. Coherent is outpacing rival Lumentum in its transition from 3-inch to 6-inch InP wafers. By Q3, 6-inch yields had already surpassed 3-inch yields — a milestone that was still in its early stages just one quarter prior. The 6-inch fab footprint is now expanding to Switzerland, joining existing sites in Texas and Sweden. Just as silicon semiconductors standardized on 12-inch wafers, the InP transition to 6-inch structurally lowers unit costs and widens the cost gap with competitors on a sustained basis.
The restoration of financial health is equally significant. The net debt ratio collapsing from 1.7x to 0.5x in a single quarter — underpinned by NVIDIA’s $2 billion strategic investment — means Coherent has been formally recognized as a core partner in the AI optical infrastructure supply chain. This creates a virtuous cycle in which future large-scale capacity investments can be funded at lower capital costs.
4. GlobalFoundries Silicon Photonics Revenue to Double in 2026, $1B Target by 2028, CPO Solution SCALE Launched — The Inflection Point Where CPO Moves from Possibility to Reality
• Core Source
“2026 revenue: approximately 2x growth year-over-year (analyst estimate ~$400M)”
“Target: exceed $1B by end of 2028”
“Design-wins secured with 3 of the top 4 pluggable optical transceiver companies”
“May launch of SCALE, a CPO optical module solution / Industry’s first OCI MSA-compliant platform, specialized for AI scale-up architecture”
“Short-term contract volumes: tight supply/demand → price increases planned for 2H26–2027”
• Expected Impact
GlobalFoundries’ silicon photonics business is projected to double in revenue within 2026 alone, and to exceed $1 billion by 2028. The core driver of this growth trajectory is the accelerating transition within AI data centers from pluggable optics to CPO (Co-Packaged Optics).
CPO integrates optical modules within the same package as switch chips, dramatically reducing power consumption and increasing signal transmission speed compared to conventional pluggable modules. As AI clusters scale to hundreds of thousands of GPUs, the volume of data movement between GPUs is exploding — making CPO not an optional enhancement but an engineering necessity.
SCALE, launched this month, is the industry’s first CPO platform compliant with the OCI MSA (Open Common Interface Multi-Source Agreement) standard, giving GF a first-mover advantage in establishing an open ecosystem standard. Design-wins already secured with 3 of the top 4 pluggable transceiver manufacturers indicate that GF’s silicon photonics technology is consolidating its position as the market standard. The roadmap extending from 1.6T support to 3.2T and beyond implies that this structural benefit will carry through to the next generation of AI accelerators.
5. Rocket Lab 1Q26 Revenue +63.5%, Backlog $2.22B, Book-to-Bill 2.9x — Establishing Oligopolist Position in the Commercial NewSpace Launch Market
• Core Source
“Revenue: $200 million (+63.5% y-y, beat consensus by 5.7%)”
“Backlog: $2.22 billion / Book-to-bill ratio: 2.9x”
“Gross profit: $76.5 million (+117.0% y-y, beat consensus by 0.9%)”
“Electron’s launch price has risen from $5–6 million in the past to approximately $8.5 million currently.”
“The company maintained its target of a test launch within Q4, and announced new launch contracts including 5 Neutron launches.”
• Expected Impact
The most notable figure in Rocket Lab’s Q1 results is the book-to-bill ratio of 2.9x. This means that for every $1 of revenue recognized, $2.90 of new orders are being booked — a clear signal that demand growth is structurally outpacing the company’s ability to expand supply capacity. A $2.22 billion backlog represents approximately 11x current quarterly revenue, effectively locking in years of future revenue visibility.
The shift in pricing power is equally meaningful. Electron’s launch price rising from $5–6 million to approximately $8.5 million reflects a market structure in which rising demand and constrained supply capacity are enabling Rocket Lab to command a premium. Gross margins have also climbed from 33.4% to 43%, confirming that economies of scale are beginning to materialize in earnest.
From a medium-term perspective, the successful development of Neutron — the mid-size launch vehicle — is the pivotal variable. Neutron addresses not just small satellite constellation deployment but the large payload market as well, and 5 Neutron launch contracts have already been signed. Successful commercialization of Neutron would allow Rocket Lab to expand from its current small-launch oligopoly into the medium-launch market, driving both TAM expansion and further ASP (average selling price) appreciation.
6. Datadog 1Q26 Revenue +32%, RPO +51%, AI Lab Customer Wins — Structural Growth in AI Infrastructure Monitoring and Observability Platform Dominance
• Core Source
“Q1 revenue was $1.01 billion, up 32% year-over-year. ARR surpassed $4 billion, and the sequential revenue increase was the strongest Q1 since 2022.”
“RPO (remaining performance obligations) reached $3.48 billion, up 51% year-over-year.”
“56% of all customers are now using 4 or more products, reflecting continued multi-product expansion.”
“New wins at major hyperscale AI training labs, strengthening AI-related growth momentum.”
“Billings were $1.03 billion, up 37% year-over-year.”
• Expected Impact
The most strategically significant development in Datadog’s Q1 results is the newly secured hyperscale AI training lab customers. AI training clusters operate tens of thousands of GPUs simultaneously, generating enormous volumes of telemetry data — an environment that demands far more sophisticated observability solutions than conventional cloud monitoring. Winning this customer segment signals that Datadog is beginning to establish itself as the platform standard in AI-native infrastructure.
The fact that RPO grew 51% year-over-year to $3.48 billion demonstrates that revenue growth is rooted in multi-year contractual commitments rather than transient demand. The metric showing 56% of customers using four or more products is equally important. It means customers have internalized Datadog not as a point solution but as a core platform for infrastructure operations — a dynamic that implies deep structural lock-in, where switching costs are prohibitively high once the platform is embedded.
As AI infrastructure investment scales up, monitoring and observability costs scale proportionally — meaning Datadog is positioned as a direct beneficiary of AI capital expenditure expansion. Its rapidly expanding AI-native product lineup — including GPU monitoring, Bits AI Security Agent, and Experiments GA — suggests the company is pioneering an entirely new TAM: dedicated AI infrastructure monitoring, distinct from traditional cloud observability. This is a compelling medium-term growth driver beyond its already dominant core market.
Comment [0]