arrow_back Back to Glasshouse Case study · Strategy-firm partner

Case study / B2B GTM Sim / Strategy consulting firm · B2B software vendor · F500 enterprise buyer

Broadcom / VMware, March 2025.

Five candidate framings on the morning of 27 March 2025, for a VMware commercial-software licensing move scheduled to enter the distributor channel the next day. The directive that actually shipped was the one the rehearsal ranked as structurally worst. It was effectively retracted inside 14 days.

Time to effective retraction

14 days

Distributor-channel leak to full walk-back · sharp hostile reaction across F500 customers, enterprise-IT trade press, the partner channel, and European competition bodies

T=0

2025-03-27

09:00 PT · morning of the commercial-software executive go/no-go review

01 / The decision

Five framings on the table. One window to pull.

On the morning of Thursday 27 March 2025, Broadcom's commercial-software executive team sat in front of a drafted directive to raise the VMware minimum-commitment floor from 16 cores per CPU to 72 cores per product on new VMware Cloud Foundation and vSphere Foundation orders, effective 10 April 2025, with no grandfathering and no tier exemption. The Arrow EMEA distributor memo announcing the change was scheduled for the following morning. The 09:00 PT go/no-go review was the last upstream moment at which the move could be softened, phased, tiered, or pulled.

That is the archetype. A senior partner at a B2B strategy consulting firm runs the rehearsal in one of two directions — as a vendor-side stress test for a B2B software client contemplating a licensing-consolidation move, or as a counterparty rehearsal for a Fortune 500 enterprise buyer on the receiving end of one. The room has five candidate framings. The room cannot empirically test any of them against real customers, partners, or regulators. The decision is made by internal judgment, with one shot, and the downside is a 14-day reversal visible in trade press and competition-authority filings.

The directive that shipped was the hardest of the five: enforce the 72-core floor, no grandfathering. The Arrow EMEA distributor memo leaked the next morning. Within the 14-day window that followed, customer and partner reaction across LinkedIn, Reddit, and the enterprise-IT trade press was sharply hostile, and on or around 10–11 April 2025 Broadcom effectively retracted the 72-core floor — while a Broadcom spokesperson simultaneously told the trade press that "Broadcom has never announced a price change." 1

T=0 / The cutoff

Morning of 27 March 2025.

The seed dossier contains only what Broadcom's commercial-software executive team demonstrably had in front of them as of end-of-day Wednesday 26 March 2025 — the 22 January 2024 VMware perpetual-licensing-end blog post, the January–February 2024 Advantage Partner Program contraction (4,500 → 500 partners), the August 2024 AT&T filing and its November 2024 confidential settlement, the 21 March 2025 VMware v Siemens Delaware filing, Q4 FY24 and Q1 FY25 Broadcom earnings commentary, the Arrow EMEA distributor-practice baseline, the CISPE European positioning, the Tesco and Rijkswaterstaat customer-footprint baselines, and the internal decision-basis memo. 19 stakeholders span the customer, supplier, regulator, analyst, press, competitor, and investor archetypes the executive team would have had in the room.

Held out of the rehearsal and never uploaded to Glasshouse: the Arrow distributor-memo leak on 28 March 2025, the 14-day reaction wave, the Broadcom spokesperson denial, the effective retraction on or around 10–11 April 2025, the Rijkswaterstaat Dutch civil-court ruling of late June 2025, the Tesco plc £100 million claim filed in September 2025, the UnitedHealthcare and Fidelity filings, and the 2026 CISPE VCSP-program-closure complaints. Gemma 3's training cutoff is August 2024 and Gemini 3 Flash's is January 2025 — T=0 is seven months and two months post-cutoff respectively. Clean simulation territory for both backends.

02 / The rehearsal

What Glasshouse produced.

The rehearsal constructed a 30-persona stakeholder pool across four platforms (X, LinkedIn, Reddit, Bluesky) and ran the five framings through 12 rounds of simulated interaction. The dominant reaction narrative surfaced in the first 72 hours was specific: “predatory post-M&A lock-in over mission-critical enterprise infrastructure”, attributable to the F500-regulated-retail, tier-1-telco, regulated-healthcare, and European-regulated-infrastructure customer cohorts, amplified by CISPE and the enterprise-IT trade press. 2

A secondary narrative from the surviving Advantage Partner Program cohort and the former-VCSP-casualty cohort read as “burning through the Pinnacle tier” — the partner-channel frame, structurally distinct from the customer-side economics. A third secondary narrative from CISPE and the European public-sector operator cohort surfaced “European regulated-infrastructure is the soft target for competition-law escalation”, with specific reference to Dutch civil-court standing and EU procurement-directive exposure.

Against the five framings, the simulation produced a triple Purchase Decision Metric: a 14-day VMware ARR preservation range, a count-band of plausible net-new enterprise-customer litigation filings, and a European regulatory-action delta (none / reputational-only / substantive-escalation). The enforce_72_floor framing landed in a materially worse ARR and litigation and regulatory posture than the three middle framings — the tiered segment floor, the softer 32-core floor, and the phased grandfathered floor. The null-hypothesis defer_16_floor was confidently better only when the customer cohort was fully allocated relative to partner and competitor voice. 3

03 / What actually happened

Rehearsal alongside record.

Glasshouse rehearsal · T=0 2025-03-27

Framings ranked by structural damage

01 Enforce 72-core floor, no grandfathering — ranked highest risk
02 Tiered segment floor by customer size
03 Softer 32-core floor as compromise anchor
04 Phased grandfathered floor with 12-month runway
05 Defer the floor — hold 16 cores, null hypothesis

The dominant narrative emerged in the first 72 hours of the simulated 14-day window, attributable to the F500-retail, tier-1-telco, regulated-healthcare, and European-public-sector cohorts — with CISPE and the enterprise-IT trade press as the amplifiers.

Real-world record · 28 March 2025 to 2026

Documented reaction

01 Arrow EMEA distributor memo leaked 28 March 2025 — picked up by the enterprise-IT trade press within hours
02 Sharp customer and partner reaction on LinkedIn, Reddit, and trade-press comment threads across the first business week
03 Effective retraction on or around 10–11 April 2025 — alongside a Broadcom spokesperson stating “Broadcom has never announced a price change.”
04 Rijkswaterstaat Dutch civil-court ruling (late June 2025), Tesco plc £100 million claim (September 2025), UnitedHealthcare and Fidelity actions
05 CISPE formal EU competition complaints on the VCSP-program closure, filed and amended through 2026

Primary reporting from The Register, Network World, Computer Weekly, ChannelE2E, and court filings in Dutch civil court, the High Court of Justice of England and Wales, and the EU Commission directorate-general for competition.

The 14-day reversal was structurally visible from the transformation history the commercial-software executive team already had in front of them. The rehearsal rebuilds the reaction space from that pattern, not from retrieval of what actually happened next.

04 / Methodology

Four rules. No exceptions.

Rule / 01

T=0 is explicit.

T=0 for this case is 09:00 PT Thursday 27 March 2025 — the last upstream moment at which the commercial-software executive team could have pulled, modified, or softened the directive before it entered the Arrow EMEA distributor channel. The seed contains only documents the team demonstrably had in front of them as of end-of-day 26 March 2025.

Rule / 02

T=0 is post-cutoff.

Gemma 3's training cutoff is August 2024 and Gemini 3 Flash's is January 2025. The 28 March distributor leak, the 14-day reaction wave, the effective retraction, and the 2025–2026 litigation and regulatory tail are all post-cutoff for both backends. Neither model has training exposure to the reaction or to the longer-tail consequences.

Rule / 03

Anchors are real.

The 14-day reversal window, the effective-retraction timing, the Broadcom spokesperson denial, the Rijkswaterstaat Dutch-court ruling, and the Tesco £100 million claim come from primary reporting in The Register, Computer Weekly, Network World, and from court filings in Dutch civil court and the High Court of Justice of England and Wales. Every anchor on this page is linkable and verifiable.

Rule / 04

Never tuned to pass this case.

Glasshouse is tuned on aggregate eval results across many scenarios. This case passes the rubric as a byproduct of aggregate tuning, not because Glasshouse was reverse-engineered around the 27 March decision. The case earned its place in the public gallery by producing the dominant narrative, the partner-channel frame, and the European competition escalation from the pre-T=0 dossier alone.

05 / Related rehearsals