MicroLink × NVIDIA · Working Session Reference
← Back to Eight Zones
Prepared 4 May 2026 / Version 0.2 Draft / Confidential
Zone 08 · Contributions · Cluster C

Workforce andecosystem

For Jumbi Edulbehram and the NVIDIA Deep Learning Institute

SJSU is already aligned. The December 2024 NVIDIA, City of San José, and SJSU MOU is the foundation. The pitch is not "let us start something new." It is "let us extend the thing you already signed" into a registered AI Infrastructure Apprenticeship.

Owner
Jumbi Edulbehram
VP Global BD · Public Sector
Adjacent
DLI · SJSU
Curriculum · Academic partner
MicroLink lead
Sancha Olivier
Co-Founder · Programme architecture
Foundation
Dec 2024 MOU
NVIDIA + City + SJSU · Already signed
Working session
5 May 2026
30 min · Teams · 13 attendees
08 The Thesis
SJSU is pre-aligned. The December 2024 NVIDIA, City of San José, and SJSU MOU is the foundation, NVIDIA's first direct workforce-pipeline initiative with a city government. 700+ alumni already at NVIDIA. Free DLI workshops for SJSU educators. Inception mentorship integration. The apprenticeship is the apparatus that converts the MOU into certified operators flowing into the NVIDIA partner network. The NCP rollout is workforce-bottlenecked. Equinix and Omdia project a 2.4M-worker shortage. This is the documented pipeline.
01
MOU foundation
Dec 2024
NVIDIA + City of San José + SJSU. NVIDIA's first direct workforce-pipeline initiative with a city government.
Already signed Extension target
02
SJSU alumni at NVIDIA
700+
Already employed at NVIDIA. The pipeline is real, observed, and operating without a structured apprenticeship.
Operating Pre-validated
03
GovAI Coalition
900+
Member agencies. Khaled Tawfik founded it. The convening surface for replicating the apprenticeship across US cities.
Replication path City CIO led
04
Industry shortage
2.4M
Worker shortage projection per Equinix and Omdia for the NCP rollout. The workforce bottleneck is documented.
Bottleneck Equinix · Omdia

The MOU is already signedand SJSU is already aligned#

Zone 08 is not a new initiative. It is the extension of an existing relationship. Every figure that follows assumes the foundation already exists, because it does.

Already signed · December 2024
NVIDIA, City of San José, and SJSUsigned the foundation eighteen months ago.
It is NVIDIA's first direct workforce-pipeline initiative with a city government. 700+ SJSU alumni are already employed at NVIDIA, free DLI workshops already run for SJSU educators, and Inception startup mentorship is already integrated with SJSU programmes. The apprenticeship is the structured apparatus the MOU has been waiting for. The pitch is the extension memo, not the cold introduction.
Pillar 01
700+
SJSU alumni already employed at NVIDIA, validating the unstructured pipeline.
Pillar 02
DLI · free
Workshops for SJSU educators at no cost, already running under the MOU.
Pillar 03
Inception
Startup mentorship integration with SJSU programmes, already in place.

The framing matters because it changes what we are asking Jumbi to underwrite. We are not asking him to start a programme. We are asking him to extend the existing MOU into a registered apprenticeship structure, with NVIDIA Public Sector as the natural sponsor and the Deep Learning Institute as the curriculum spine. Both already touch SJSU. Neither is being asked to build new relationships.

Physical anchorsthat make SJSU operationally credible

The Cluster C research surfaced three physical anchors that make SJSU more than a name on a letter. They are the reason the apprenticeship has somewhere to run.

Anchor · 01
$186Meight stories
SJSU Interdisciplinary Science Building, opened summer 2023. First new science building in 50 years. Houses the interdisciplinary Center for High Performance Computing.
Anchor · 02
MPELDavidson College
Microelectronics Process Engineering Laboratory in the Charles W. Davidson College of Engineering, with Micron grant funds underwriting TAs.
Anchor · 03
5 labsexisting AI
Circuits, Robotics, and AI Lab. Intelligent Conversational Agents and Neural Networks Lab. Center for AI and Cybersecurity. Silicon Valley Big Data and Cybersecurity Center. Advanced Institute for Ethical Technologies.

Existing NVIDIA touchpointsthat the MOU formalised

Already running under the December 2024 MOU
Free DLI workshops Delivered to SJSU educators at no cost, building faculty capacity for further DLI integration.
700+ alumni at NVIDIA Already employed across hardware, software, and operations functions. The pipeline is observed, not theoretical.
Inception startup mentorship Integrated with SJSU programmes, connecting student founders to NVIDIA's startup ecosystem at zero acquisition cost.
→
Extend the existing MOU into a joint AI Infrastructure Apprenticeship, not pitch a new relationship
The MOU is the most under-used asset in the Cluster C corpus. Every subsequent ask in this brief is built on the assumption that the foundation already exists. Jumbi is not being asked to build a programme. He is being asked to underwrite the next clause of an agreement that NVIDIA, the City, and SJSU already signed.

The DLI is the curriculumframe the apprenticeship plugs into#

The Deep Learning Institute already publishes the credential ladder we need. The contribution is wiring the apprenticeship onto NCA-AII as the entry gate and NCP-AIO as the exit credential.

The DLI's three programme components map almost too cleanly onto the apprenticeship structure. The University Ambassador Program lets SJSU faculty deliver DLI workshops at no cost after certification, which means the apprenticeship's classroom delivery is taught by SJSU faculty using NVIDIA-authored material. The Teaching Kit Program, co-developed with NYU's Yann LeCun and Dartmouth's Sam Raymond, covers accelerated computing, data science, deep learning, robotics, and the Generative AI Multimodal Associate Certification. The certification ladder is NCA-AII at entry and NCP-AIO at advanced, priced from $125 to $400 with two-year validity and Credly digital badges.

The frame is built. The contribution is the wiring.

The pathfrom undergrad to certified operator

The path through the apprenticeship reuses every existing DLI lever. SJSU adopts the Generative AI Teaching Kit for undergrad coursework, building the upstream awareness pool. NCA-AII becomes the credential gate for entry into the AI Infrastructure Apprenticeship. NCP-AIO becomes the post-apprenticeship cert that signals readiness for placement into the NVIDIA partner network. Each rung is already published; we are not authoring credentials, we are sequencing them.

Step
Credential or stage
What it produces
Owner
01
Teaching KitGenerative AI Multimodal
SJSU adopts the DLI Teaching Kit in undergrad CS and EE programmes. Builds the awareness pool from which apprentices are recruited.
SJSU
02
NCA-AIIEntry credential
NVIDIA-Certified Associate, AI Infrastructure and Operations. Becomes the named entry gate for apprenticeship cohort selection.
DLI
03
2-yr ApprenticeshipSJSU + MicroLink RWF
Two-year structured programme. Classroom at SJSU. Field placement at the MicroLink RWF deployment from Q1 2027 commissioning. Three specialisations.
Joint
04
NCP-AIOExit credential
NVIDIA-Certified Professional, AI Infrastructure and Operations. The advanced credential earned on graduation; signal of placement readiness.
DLI
05
PlacementPartner network
MicroLink deployments first. NVIDIA partner network second. Bloom Energy and FuelCell Energy power-side partners. City and municipal facilities operators.
Network
§
DLI does not have to author anything new
Every credential and every Teaching Kit referenced above already exists in the DLI catalogue. The DLI's contribution is activation (the University Ambassador Program for SJSU faculty), alignment (Teaching Kit selection for the apprenticeship curriculum), and recognition (NCA-AII and NCP-AIO as named gates). Three small commitments. Zero new IP.

The AI Infrastructure Apprenticeshipis the structure that converts the MOU#

The career track does not exist in a published, certifiable form anywhere in the AI infrastructure industry. The TSMC Arizona model is the analogue. The MicroLink-SJSU programme is the apparatus.

The analogueTSMC Arizona Registered Technician Apprenticeship

The closest precedent in US industrial workforce development is the TSMC Arizona Registered Technician Apprenticeship Program, sponsored by the Phoenix Workforce Development Board, the first US workforce board to sponsor a registered apprenticeship in semiconductors. TSMC invested $5M+, with education partners Estrella Mountain Community College, Maricopa Community Colleges, Northern Arizona University, and Rio Salado College. The programme runs 18 to 24 months for Equipment, Process, and Facilities Technicians, with ~130 apprentices in initial cohorts. Arizona semiconductor processing technicians earn roughly $30 per hour, or $62,370 per year per BLS. The wage premium is what makes the apprenticeship recruit-able.

The TSMC programme is the proof that a city, a workforce board, an industry sponsor, and a community-college consortium can co-build a registered apprenticeship in 18 to 24 months. The MicroLink-SJSU structure follows the same shape with the academic anchor upgraded from community college to a four-year HSI adjacent to the industry sponsor.

Programme structureand curriculum architecture

Structure
Two-year programme. SJSU classroom, MicroLink field.
  • Two-year programme based at SJSU with field placements at the MicroLink RWF deployment from Q1 2027
  • Curriculum developed jointly by MicroLink, SJSU, and DLI
  • Specialised module input from FuelCell Energy on grid-independent operations and Vertiv on liquid-cooling operations
  • Cohort structure aligned to MicroLink's deployment cadence and partner-network demand
Certification stack
Four credentials. One ladder.
  • SJSU certificate on programme completion
  • NVIDIA DLI credential: NCA-AII at entry, NCP-AIO at exit
  • AIChE Center for Hydrogen Safety certification for the fuel-cell specialisation
  • Optional OCP or LF Energy industry credential for the open-standards specialisation track

Three specialisationstracked through the same two-year frame

Spec · 01
Liquid-cooling operations
Direct-to-chip and rear-door coolant systems, secondary loops, CDU operation, ASHRAE TC 9.9 W2 chemistry. Vertiv as named module partner.
Spec · 02
Fuel-cell and grid-independent operations
MCFC and SOFC operations, biogas conditioning, hydrogen safety per CFC §5806 and NFPA 2, microgrid black-start. FuelCell Energy as named module partner.
Spec · 03
AI infrastructure operations
Run:ai scheduling, BlueField-3 SuperNIC, DOCA Platform Framework, Mission Control telemetry, multi-tenant isolation patterns. The DLI-spine specialisation.

Placement targetstiered by partner network proximity

The placement ladder is the part of the apprenticeship that turns the credential into a career. MicroLink deployments come first, absorbing the early cohorts as the San José RWF site, the planned Phase 2 cluster, and subsequent industrial-host sites commission. NVIDIA partner network deployments are second, with the apprenticeship credential serving as a structured signal across the NCP and DGX-Ready Colocation populations. Bloom Energy and FuelCell Energy are obvious power-side employers given the fuel-cell specialisation. City and municipal facilities operators close the loop, returning trained operators to the public sector that helped fund the programme.

!
The apprenticeship is the apparatus that converts the MOU into certified operators flowing into the NVIDIA partner network
NVIDIA's global NCP rollout is workforce-bottlenecked. Equinix and Omdia project a 2.4 million worker shortage for the AI infrastructure rollout. The apprenticeship is the documented pipeline that addresses it. Every certified operator is a unit of friction removed from the next NCP candidate's commissioning timeline. This is how Cluster C compounds.

The Inception startup flywheelconnects existing demand to new physical compute#

Inception is the existing flywheel. The municipal GPU tier is the missing physical anchor. San José becomes the first US city to connect them.

Global precedents for publicly-anchored GPU compute are well-established outside the US. The pattern works. What is missing is a US city in the comparison row.

Helsinki · Finland
LUMI AI Factoryfree access at scale
3,000 projects ~7,000 users 140M GPU-hours €612M
Free access since 2022. The largest publicly-funded AI compute facility in Europe. Pre-Singularity reference for what national-scale subsidised AI compute looks like.
Singapore
NSCC ASPIRE 2A+state-anchored compute
SGD 270M
National Supercomputing Centre flagship. State capital, state-aligned research access, programmatic startup integration. The Asia-Pacific reference.
Quebec · Canada
Mila with Hypertecprovincial AI cluster
CAD 250M LaSalle campus
Provincial-scale anchor with academic and startup access. Mila as the curatorial layer; Hypertec as the infrastructure operator. Canadian reference for academic-private hybrid.
Toronto · Canada
SciNet Trilliumfederal academic anchor
CAD 42.5M Federal
Federally funded. University of Toronto-led. Smaller than LUMI or NSCC but the cleanest precedent for federally-funded municipal-adjacent academic compute.
∅
No US city has built dedicated municipal GPU compute for residents and startups
The US row in the comparison is empty. Not because demand is absent, but because the infrastructure pattern has not been demonstrated. San José becomes the first US city to publish a municipal GPU tier under a structured Inception integration. Every subsequent US city referring to the example does so by name.

The proposal5 to 10% of the 1 MW deployment as reserved tier

A tier of capacity, perhaps 5 to 10% of the 1 MW deployment, equivalent to roughly 50 to 100 GPU-hours per Inception startup per month, reserved for City of San José Inception members under the Inception Premier discount structure. The capacity is tiered, not promotional: it is a stable, contractually defined municipal compute pool with eligibility tied to City of San José residency or programme participation.

The Inception Premier discount structure is the existing commercial mechanism we plug into. We are not asking NVIDIA to invent a new pricing tier; we are asking NVIDIA to apply the existing tier to a geographically scoped municipal cohort. That distinction matters because it removes a procurement-cycle objection before it surfaces.

The replication logicthe city becomes the convening node

Once San José is the first US city to publish a municipal GPU tier integrated with Inception, the GovAI Coalition becomes the natural convening surface for replication. Khaled Tawfik, San José's CIO and the Coalition's founder, already operates a 900+ member agency network across US cities and counties. The flywheel exists. We are connecting it to physical compute. The city becomes the convening node for a US municipal Inception programme, replicable across the next thirty cities through the Coalition.

"
The MOU is signed. The alumni are at NVIDIA. The DLI workshops are running. We are extending the existing relationship into the structure it was always going to need.
Zone 08 frame
For Jumbi Edulbehram

The askand what MicroLink contributes#

Three columns of small, specific commitments. Each one is an extension of something that already exists. Bilateral working sessions on a clean 30 to 60-day clock.

Tier · Workforce and ecosystem · MOU extension
Extend the December 2024 MOUinto a registered AI Infrastructure Apprenticeship

SJSU is aligned. The DLI is the curriculum spine. Inception is the startup flywheel. The contribution is binding the three into a single apparatus that produces certified operators flowing into the NVIDIA partner network.

From Jumbi Edulbehram
  • Bilateral session within 30 days on extending the December 2024 MOU
  • Public Sector advocacy for the AI Infrastructure Apprenticeship as an NCP-recognised credential
  • Bridge to Khaled Tawfik, City CIO and GovAI Coalition founder, for replication framing
  • Co-presenter for the apprenticeship launch at SJSU
  • Inception Premier discount structure approval for City Inception members
From the NVIDIA Deep Learning Institute
  • University Ambassador Program activation for SJSU faculty
  • Teaching Kit alignment for the apprenticeship curriculum
  • NCA-AII as the named entry credential
  • NCP-AIO as the named exit credential
  • Co-development of the three specialisation modules: liquid cooling, fuel cell, AI infrastructure
What MicroLink contributes
  • Field placements at the San José RWF deployment from Q1 2027 commissioning
  • Specialised module input from FuelCell Energy and Vertiv as named partners already engaged
  • Inception startup hosting at preferential rates under municipal incentive
  • First US municipal GPU-compute tier for residents and startups
  • Replication template through the GovAI Coalition for the next thirty cities
Bilateral working sessions requested · 30 minutes each
  • Jumbi Edulbehram · extending the December 2024 MOU into the AI Infrastructure Apprenticeship · within 30 days of the LOI
  • NVIDIA DLI lead · curriculum integration, NCA-AII entry credential, NCP-AIO exit credential · within 60 days
  • NVIDIA Inception (David Messina) · Inception Premier discount structure for City Inception members · within 60 days