Kimi Antonelli claimed United States Grand Prix victory from pole on a one-stop strategy, beating Charles Leclerc by 13.078s with Max Verstappen third. Mercedes’ pit crew held firm against Verstappen’s two-stop undercut. Ferrari had dominated early timing sheets: Leclerc led FP2 and Hamilton topped final practice, so Mercedes’ shock qualifying pace surprised the paddock and fans. Still, Antonelli left Austin without securing the title; Russell’s eighth and Hamilton’s fourth kept the championship fight alive.

I watched the race at home, but this piece focuses on the two preceding days, when HPE hosted media and analysts inside Mercedes-AMG PETRONAS F1’s program at Circuit of the Americas. The race-winning car was far from the machine rolled out Friday morning, and that rapid transformation is where HPE’s infrastructure work fits in.

HPE has served as Mercedes’ IT partner since 2018, supplying compute, storage and networking linking Brackley HQ to trackside garages. The visit included a Thursday panel with Antonio Neri, Toto Wolff and team CIO Michael Taylor, a Friday in the Silver Arrows Lounge above the pit wall, plus a no-photo garage tour.

The Car Is Never Finished

The defining moment came between Friday’s two practice sessions. Mercedes streamed garage briefings to the lounge, showing mechanics stripping both cars fully within two and a half hours of FP1’s chequered flag, even before engineers sent revised setups from the factory.

Russell’s car received changes first: increased rear negative camber, a return to the original steering rack and a new rear anti-roll bar, all targeting the “three-wheeling” issue where the inside rear wheel lifts in corners. Russell had reported front-wing trouble and locked up into Turn 1 during FP1, while Antonelli’s car required far fewer tweaks.

FP2’s plan used fresh medium and soft tyres for both drivers, plus a long medium-tyre run for Antonelli, who had not sampled that compound in FP1. After FP2, both cars were fully disassembled for system checks; no further trackside setup work followed that night. Final tuning calls came from Brackley’s simulators running overnight on UK time, while mechanics observed the FIA curfew.

FP2 ended with Leclerc first, Antonelli just 0.042s behind, Verstappen third and Russell fifth (+0.744s). Antonelli lost valuable session time early to a loose seat mount but called the run a strong step forward. The anti-roll bar alone cannot explain the gap between the two Mercedes, yet Friday showed the team rebuilding its cars twice in a single day to resolve conflicting feedback from its two drivers.

The Upgrade Package and Austin’s Role

Mercedes filed no new components with the FIA for Austin. The combined setup run here paired the Malaysia-introduced floor and rear assembly with the Singapore front wing and endplates, deployed on both cars together for the first time. Singapore had been a split test: Antonelli ran the new kit and Russell the old, with only one new wing built for limited practice running.

The garage team dubbed it the “double drop” wing, built for 2026 regulations. This narrower wing features a two-element flap that opens on straights in X-mode and defaults to Z-mode for cornering; Mercedes’ prior design only dropped one element on straights. Mechanics also noted the new package is simpler to assemble.

Wolff framed the limits of simulation: one prior upgrade projected a 0.35s gain in models but slowed the car by the same margin. Aerodynamic sensors registered the predicted load, yet drivers reported unpredictable handling, brake lock-ups and entry understeer. “The engineer says it’s all there, and the driver says the car is unpredictable,” he said.

Translator, or Therapist

Asked which source to trust, Wolff joked he trusted “the most expensive sensor — the driver in the cockpit.” He described his role as bridging data, engineering priorities and what drivers need to exploit the car.

He recalled Russell describing the car as a “three-legged chair.” Engineers spent months chasing a lifting wheel until Russell clarified it was rapid wheel oscillation, a separate fault requiring a different fix. Seeing the garage implement steering and suspension changes in under two and a half hours, “translator” feels too mild: Wolff mediates two truthful but conflicting viewpoints to build a car both drivers can trust. “Data don’t take decisions. Whoever human is in charge needs to take that decision.”

HPE’s Trackside Operations

HPE’s chief technologist Max Walker recounted the partnership’s start, when trackside compute was little more than consumer PCs. HPE replaced that with dense, consistent Edgeline systems, rolled out Aruba wired and wireless networking across the garage and Brackley campus, and now supports Wi‑Fi 7. The original HPE account team only numbered three.

Telemetry flows from the car over Wi‑Fi during sessions or hardline umbilicals when stationary. Data is filtered trackside for urgent insights before being sent to Brackley’s Race Support Room. FIA personnel caps mean most engineers analyse runs remotely from the factory. Taylor’s priority was clear: reliability outranks raw performance, because engineers depend on complete telemetry from every run.

Two sealed, vented HPE racks sit behind the pit wall, hardened against dust, heat and sand, fitted with filters and auxiliary cooling, housing the Edgeline hardware. Aruba access points line the garage ceiling, with HPE branding visible on local monitors and control systems.

Factory Workflow and Cost Cap Pressures

Taylor’s small internal IT team leans on HPE for short-term projects it cannot deliver alone, such as the 2018 software-defined data center rollout. Lap-time gains originate at the factory: CFD designs progress to 60% scale wind-tunnel models, delivering over a second of lap improvement across a full season. Mercedes generates up to 50TB of CFD data weekly; HPE’s updated AMD ProLiant HPC hardware cuts three-day simulation jobs down to one day.

The FIA cost cap restricts spending across Mercedes, Ferrari and Red Bull, limiting purchases including AI systems. HPE GreenLake, paired with HPE Financial Services, structures infrastructure costs flexibly as capital or operational expense, giving the team budget headroom within cap rules.

Mercedes will be the first UK recipient of HPE’s ProLiant Gen13 engineering server, built on AMD Venice CPUs, for pre-launch validation of performance, sustainability and workload density. CFD workloads demand high memory bandwidth, and Venice delivers up to 256 cores, 1.6TB/s bandwidth, DDR5 or MRDIMM memory and PCIe Gen6 within a 1U low-power chassis optimized for dense rack deployments. For a team constrained by rack space and power budgets, the goal is more simulation throughput per watt and per unit.

Mercedes also runs HPE Morpheus Enterprise for IT-as-a-service, automating infrastructure requests for engineers and providing cost-efficient virtualization under the cap.

AI Bound by Rules and Budget

Antonio Neri shared HPE’s own AI adoption: 1,500 internal use cases, with AI spend quadrupling over three quarters to reduce overall operational costs. Mercedes’ AI rollout is constrained less by the cost cap than FIA regulations governing where AI may be used in vehicle development.

Taylor said the team now self-hosts open-weight models on HPE hardware, a cost-effective way to serve over 200 developers without hitting restricted budget lines. He envisions lead developers supported by agentic AI assistants, with prompting a core skill for junior staff, reducing reliance on commercial tools. Still, tuning a race car remains a deterministic problem AI cannot solve reliably for years.

Wolff joked his ideal setup would need no human engineers, only “blinking stacks and GPUs.” He set aggressive targets for non-racing departments — legal, comms, compliance and sponsorship — aiming for 30% efficiency gains within six months using agentic tools, even as teams push back that the targets are unachievable.

The HPE Advertisement

HPE’s widely aired “self-driving network” ad features Russell texting Antonelli about the pitch, leaving Antonelli dejected across press conferences, lunch and a late-night message asking about the US race. The joke lands because the partnership’s real value stays invisible: faster CFD cycles and timely telemetry do not draw fan attention in the paddock. An autonomously managed network frees staff to focus on changes that make the car quicker.

Saturday Qualifying

Saturday began with Hamilton topping FP3, ahead of Verstappen and Leclerc; both Mercedes sat fourth and fifth, with Antonelli losing a lap to track-limit violations. Yet two hours later, Antonelli took pole, his seventh of the season.

Wolff had expected the car to rank third to fifth, calling the pole lap beyond expectations. The upgraded car operates within a narrow performance window; miss that peak and lap times drop sharply. Russell’s Q3 suffered two poor laps, while Antonelli credited factory work and clean, penalty-free running for his strong qualifying. Wolff warned that qualifying speed did not guarantee race pace, making a title clinch in Austin unrealistic.

Sunday Race

Antonelli started on mediums against Verstappen’s softs, led into Turn 1 and stuck to a single hard-tyre stop, while rivals ran two stops. When Verstappen pitted for the second time on lap 31, Leclerc briefly led before Antonelli retook the lead at Turn 12. Mercedes’ pitwall pre-modelled the undercut risk and held firm. “Stick to the data, trust your data,” Wolff summarised.

Russell’s race unravelled early: he ran wide at Turn 12, earned a five-second track-limits penalty, emerged behind Piastri after his stop and could not overtake, finishing eighth with what he described as poor race pace.

After Austin, Antonelli holds 365 points, Russell 242 and Hamilton 241. A top-nine finish in Mexico will make Antonelli F1’s youngest champion. Mercedes leads the constructors’ standings comfortably.

Closing

Austin reinforced that Mercedes’ IT challenge is reconciling drivers, engineers and telemetry, all describing the same car from different angles. Teams must align those perspectives between sessions to unlock faster laps. Wolff’s Sunday instruction to trust the data captures half the equation; Russell’s result shows driver feel still matters deeply.

HPE’s role enables that fast-cycle collaboration across continents: Brackley engineers continue working while Austin mechanics rest, ensuring critical race-day data arrives reliably. Early in the partnership, Neri gave Wolff a chip encased in acrylic, with the guiding question: how can HPE build a lawful competitive advantage? On Sunday, that advantage delivered a 13-second winning margin, bringing both championships well within reach.

Beijing Qianxing Jietong Technology Co., Ltd.
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!