Colette Kress Takes Over as Nvidia Interim CEO After Huang’s Reported Exit

Nvidia CFO Colette Kress has assumed the role of interim chief executive following Jensen Huang’s November 12 resignation, as reported by CryptoPotato. This transition places her at the helm of Nvidia’s data center strategy right ahead of the company’s scheduled November 17 quarterly earnings report.

The corporation’s infrastructure ecosystem spans AI chips, enterprise software, server partners, and systems utilized by cloud providers. This leadership shakeup occurs as operators pour billions of dollars into expansion, even while identifying power and capacity bottlenecks as primary constraints on growth.

Nvidia’s Leadership Transition Comes Before Quarterly Results

What changed at the top of Nvidia?

According to CryptoPotato, Huang stepped down from his positions as CEO and board member, with Kress stepping in on an interim basis. Having spent more than a decade at Nvidia, Kress enters the role while the board of directors weighs both internal and external applicants for the permanent position.

Read More: Jensen Huang Exits Nvidia After 33 Years; Shares Fall Over 10% Premarket

This succession places a seasoned finance leader at the helm of an enterprise that was forged around Huang’s uniquely centralized management style. Core strengths like AI-chip demand, the CUDA ecosystem, and deep industry relationships remain critical assets throughout the transition.

How did the transition unfold?

BeInCrypto notes that board members initiated an inquiry on November 8 regarding whether a foreign intelligence operation had targeted Huang. His resignation followed four days later on November 12, and Elon Musk publicly voiced his support for a reinstatement petition on November 13.

The board retained Wachtell Lipton to handle the inquiry, with former acting US Attorney Joon H. Kim leading the effort. With earnings slated for November 17, the company faces an upcoming disclosure closely trailing its leadership turbulence.

Who Controls Nvidia’s Next CEO Appointment?

What authority does the board have?

Governance outlines show that executive appointments are controlled by a majority of seated directors. Following Huang’s departure, this threshold sits at six out of the 10 remaining board members. Shareholders do not possess a direct vote regarding the CEO.

While Musk holds no seat on Nvidia’s board, his public endorsement grants the reinstatement movement high-profile backing, though the ultimate hiring decision rests strictly with the directors managing the ongoing investigation and search for a permanent successor.

How much influence does Huang retain?

Nvidia’s 2026 proxy indicates that Huang controls roughly 3.58% of the voting power, whereas BlackRock and Vanguard each own stakes exceeding 7%. Meanwhile, an internal corporate post calling for Huang’s return attracted upward of 1,600 comments, highlighting the strong employee reaction to his exit.

Kress’s interim appointment infuses the immediate transition with vital institutional knowledge. Because the board is evaluating candidates both from inside and outside Nvidia, the long-term leadership path remains fluid as the firm approaches its quarterly financial disclosure.

Nvidia Data Center Products Extend Beyond AI Chips

What did Nvidia introduce for enterprises in 2021?

During Computex 2021, Nvidia launched its cloud-based Base Command platform alongside AI-optimized x86 servers via its certified systems program. TechRadar coverage of the rollout identified NetApp as a Base Command collaborator alongside server builders Asus, Dell Technologies, Gigabyte, Hewlett Packard Enterprise, and Lenovo.

The debut tied together hardware and enterprise software: these servers earned certification for VMware vSphere’s AI Enterprise, Omniverse Enterprise, and Red Hat OpenShift. Support for Cloudera data engineering and machine-learning tools was also highlighted.

How did Base Command connect teams to larger systems?

Base Command granted enterprise groups cloud access to Nvidia DGX SuperPOD computing infrastructure leveraging NetApp data management utilities. Its framework was built to accommodate multiple researchers, data scientists, and cross-functional teams running AI initiatives concurrently.

At launch in 2021, the early-access subscription was priced at $90,000 per month. This historical offering demonstrates how Nvidia historically bundled massive computing architectures with development environments and partner software to broaden its scope beyond standalone processors.

Power and Cooling Shape Nvidia Infrastructure Expansion

Where are operators encountering constraints?

DataCenterFeed market reporting points to power availability and capacity limitations as growth bottlenecks highlighted by Equinix and Digital Realty. Equinix logged $533 million in annualized gross bookings, whereas Digital Realty posted $55 million in bookings at 100% share alongside a $2.5 billion backlog.

Oracle’s infrastructure division provides another pathway for delivering Nvidia systems to end users. Oracle rolled out general availability for Vera Rubin NVL72 racks and Vera CPUs via OCI, while also teasing a multi-gigawatt Supercluster roadmap.

How are server suppliers adapting their hardware?

Supermicro’s factory-integrated 48U rack houses 256 Nvidia Vera CPUs, 22,528 Arm cores, and up to 400 TB of memory. Utilizing direct liquid cooling, it is built to execute agentic-AI sandboxes side-by-side with GPU workloads, pushing liquid-cooling demands directly into CPU rack architectures.

Crusoe takes a modular strategy with its sixth-generation Spark module, packaging up to 3.5 MW of compute power inside a factory-built enclosure. The setup features closed-loop liquid cooling requiring zero water and maintains full compatibility with Nvidia GB300 NVL72 and Vera Rubin NVL72 racks.

AI Infrastructure Spending Reaches Tens of Billions

What scale has CoreWeave reached?

CoreWeave reported $3.66 billion in revenue for the third quarter alongside a $132.5 billion backlog and upwards of 1.7 GW of active power capacity. It adjusted its 2026 capital expenditure target to a range of $37 billion to $39 billion, underscoring the massive scale of infrastructure and hardware investments driving AI computing.

The company maintains it will steer clear of parent-level unsecured debt through 2027 under current market yields. Instead, it plans to rely on contract-backed borrowing and customer prepayments, closely tying its funding strategy to secured enterprise contracts.

What investment plans has Meta outlined?

Meta issued capital expenditure guidance between $135 billion and $145 billion for 2026. It additionally registered roughly $35 billion in underwritten common and mandatory convertible preferred stock offerings while mapping out a projected $60 billion at-the-market equity program, with Nvidia committing to invest $3 billion into the transaction.

Meta announced that the initial stage of its Prometheus supercluster located in New Albany, Ohio, is operational and has trained its Muse models, with the facility scaling toward approximately 1 million GPUs. The firm also explained that software solutions and battery systems assist in flattening power fluctuations by 30% to 40%, cementing power management as a core component of its buildout.

DGX Spark Brings Nvidia Computing to Smaller Sites

What does the compact Nvidia system offer?

A May 2026 report from Trak.in details DGX Spark as a streamlined Grace Blackwell form factor originally introduced under the name Project DIGITS. Nvidia states it delivers up to 1 petaflop of AI computing performance backed by 128GB of unified memory, supporting local execution of models containing up to 200 billion parameters.

The desktop-format machine targets developers, researchers, and enterprise teams. It gives Nvidia’s infrastructure portfolio a localized footprint that complements massive cloud data centers and sprawling tech campuses, enabling on-premise model training and testing.

What is Span proposing for residential computing?

Startup Span is engineering compact XFRA server nodes tailored for residential deployment. The enterprise claims that thousands of these decentralized nodes could collectively deliver processing muscle comparable to a traditional 100-megawatt AI data center.

The concept leverages pre-existing electrical infrastructure. Nvidia executive Marc Spieler noted that utilizing already-powered sites makes practical sense given the severe hurdles involved in acquiring large-scale power allocations. While DGX Spark and XFRA are distinct hardware offerings, both fit into this broader distributed-computing narrative.

Governance and Financing Risks Accompany the Buildout

Why does the succession matter for infrastructure execution?

Huang’s pivotal footprint at Nvidia makes the executive search crucial for an ecosystem touching chip fabricators, cloud giants, and data center operators. Simultaneously, the ongoing board inquiry introduces corporate governance friction while analysts question whether massive AI capital expenditures will yield adequate returns.

CoreWeave’s funding model underscores the sensitivity of borrowing costs and customer commitments to ongoing expansion. On the distributed edge, Span’s capacity equivalencies remain corporate projections dependent on clustering thousands of smaller nodes across residential power grids.

### What does the separate Singapore server case allege?

In an unrelated legal proceeding, The Defiant reports that Singapore has brought 41 charges against Shi Bingzhe and two corporate entities. These money-laundering allegations revolve around the acquisition of 917 servers as part of an investigation tracking shipments routed to Malaysia.

This case layers additional legal scrutiny over Nvidia-related server acquisitions onto the broader industry landscape. These accusations center specifically on the named defendants and server transactions, remaining distinct from the board’s internal investigation concerning Huang.

November Results and 2027 Manufacturing Plans Come Next

Which events are on the immediate calendar?

Nvidia’s November 17 quarterly financial disclosure marks the next milestone for public updates under Kress’s interim leadership. The executive search runs concurrently alongside this reporting timetable, with the board weighing both external and internal candidates.

Meanwhile, the House Foreign Affairs Committee has formally requested Huang’s presence at a December 1 hearing, accompanied by warnings of a possible subpoena. This introduces a separate Washington complication as directors navigate the ongoing leadership transition and investigation.

What infrastructure production is planned for 2027?

Crusoe aims to kick off manufacturing for its sixth-generation Spark units at its Brighton, Colorado, facility in 2027. Their built-in support for Nvidia GB300 NVL72 and Vera Rubin NVL72 racks ties this production timeline directly into the broader supply chain outfitting AI data centers.

These factory-assembled modules integrate compute power and thermal cooling within a single assembly. They remain separate from Nvidia’s desktop-sized DGX Spark machine: Crusoe’s hardware focuses on macro infrastructure rollouts, whereas Nvidia’s compact system is built for localized research and development.

?FAQ

01What specifications does Nvidia DGX Spark offer?

Nvidia lists key specifications including 128GB of unified memory and up to 1 petaflop of AI computing performance. The Grace Blackwell-based architecture runs local models of up to 200 billion parameters and was first unveiled under the banner of Project DIGITS.

02Which server manufacturers appeared in Nvidia’s 2021 certified systems announcement?

The cited hardware partners included Asus, Dell Technologies, Gigabyte, Hewlett Packard Enterprise, and Lenovo. Their AI-optimized x86 servers made up the certified systems lineup unveiled at Computex 2021.

03What was Base Command’s monthly early-access price in 2021?

The early-access tier debuted in 2021 with a price tag of $90,000 per month. It offered cloud-managed connectivity to DGX SuperPOD computing resources integrated with NetApp data governance tools for collaborative AI engineering.

04Who can appoint Nvidia’s chief executive under its described bylaws?

Selecting a chief executive requires approval from a majority of active board members—defined following Huang’s exit as six out of the 10 remaining directors. Shareholder-backed petitions do not grant a direct vote over the CEO slot.

05When are Nvidia’s next quarterly results scheduled?

The upcoming quarterly earnings release is locked in for November 17, 2026. Kress steps into that earnings presentation as interim CEO after transitioning from her CFO duties.

06When has the House committee asked Huang to appear?

The House Foreign Affairs Committee set a requested appearance date for December 1 alongside warnings of a potential subpoena, making it a formal request rather than verified testimony.

07Where will Crusoe manufacture its planned Spark modules?

Crusoe intends to launch production at its Brighton, Colorado, factory in 2027. These sixth-generation modules are engineered to house Nvidia GB300 NVL72 and Vera Rubin NVL72 server racks, delivering up to 3.5 MW of compute per module.