SANTA CLARA, CALIFORNIA —
AMD Advancing AI 2026 is the most formally structured declaration of AMD’s counteroffensive against proprietary AI infrastructure that Lisa Su Strategy has produced to date, a flagship summit scheduled for July 23, 2026, at the San Francisco Moscone Center Tech venue that will deliver AMD Advancing AI open source ecosystem blueprints for building, deploying, and scaling AI powered by AMD from silicon to software. The event positions Open Source AI development, ROCm Software maturity, and end-to-end Corporate Hardware integration as the three pillars of AMD’s answer to NVIDIA’s closed CUDA ecosystem, a strategic architecture that Lisa Su has been assembling through acquisitions, developer investment, and hyperscaler partnerships across the past three years, and will formally present to the global AI developer community in July.
What AMD Advancing AI 2026 at Moscone Center Will Deliver
AMD announced that Advancing AI 2026, its flagship global AI event, will be held both in person and livestreamed from the San Francisco Moscone Center on July 23, 2026, with the event providing the AI open ecosystem with blueprints for building, deploying, and scaling AI powered by AMD, as AMD leaders join Chair and CEO Dr. Lisa Su alongside AI ecosystem partners, customers, and developers to share how the company’s end to end AI solutions from silicon to software are reshaping the AI and high performance computing landscape.
The Moscone Center Tech venue selection is deliberate, the same location where the industry’s most consequential developer conferences have historically defined platform directions. In addition to announcements, Advancing AI 2026 will host talks, networking, and hands-on events that give attendees the chance to engage directly with AMD’s AI researchers and engineers. AMD’s next-generation 2nm EPYC Venice Zen 6 CPUs are expected to launch alongside the Instinct MI400 in 2026. The MI400 architecture, built on TSMC’s 2nm process and designed to compete directly with NVIDIA’s next-generation Rubin platform, represents the Corporate Hardware foundation upon which the AMD Advancing AI open-source ecosystem blueprints that July’s summit will distribute are intended to run.
Why ROCm Software Is the Strategic Center of Gravity
Rather than build a walled garden where the technology is controlled by a single entity, AMD has gone down the open source route with ROCm, which integrates with other open source projects such as vLLM to allow for faster innovation, and AMD Vice President of AI Software Anush Elangovan stated directly that the company could try to build something closed source but would not get the velocity of an open ecosystem.
The ROCm Software maturity trajectory that Lisa Su’s strategy has funded over the past two years is the most operationally significant development in the Open-Source AI infrastructure landscape outside of model development itself. ROCm 7 introduces full support for lower precision data formats such as FP4 and FP8, enabling developers to run modern AI models significantly faster without sacrificing accuracy, with AMD promising up to a 3.5 times improvement in inference performance compared with previous generations, while also expanding accessibility under the ROCm Everywhere initiative with support broadened to include Windows-based systems and consumer-grade Radeon graphics cards.
The ROCm Everywhere initiative is the Lisa Su Strategy move that most directly addresses the developer adoption gap that has historically constrained ROCm Software growth. A developer who can write and test AI code on a consumer Radeon gaming GPU before deploying to an Instinct data center accelerator faces no platform transition cost between development and production, the same framework APIs, the same model compatibility layer, and the same debugging toolchain operate identically across both environments. AMD has strengthened its integration with the broader ecosystem by offering day-zero support for popular tools such as PyTorch and vLLM, enabling developers to work immediately with new hardware releases.
Open-Source AI Validation Through Hyperscaler Adoption
The AMD Advancing AI open-source ecosystem blueprints that the July summit will formalize are not theoretical proposals awaiting market validation; they are documentation of deployment patterns that hyperscalers have already adopted at production scale. Meta discussed how it is already deploying MI300X GPUs for inference and plans to utilize MI350X for training workloads, citing AMD’s total cost of ownership advantages and high memory capacity as key differentiators particularly for models with 100 billion or more parameters, and Meta representatives noted that ROCm is finally ready for prime time production, while Microsoft through its Azure division confirmed it utilizes AMD GPUs for both the inference and training of OpenAI models.
The confirmations by Meta and Microsoft have strategic implications beyond the monetary value of the purchases each company made. The world’s largest social media AI platform, Meta, and the enterprise-level cloud company that provides the backend for the OpenAI production workloads, both confirm ROCm Software is ready for production use for inference and training tasks of frontier models, changing how the developer community thinks about the risks associated with moving to the type of Open Source AI parts infrastructure based on AMD. Developers will no longer question whether ROCm Software can accommodate the workload; they will ask whether their existing investments in tooling to support CUDA-based applications are justified by the cost of switching to an AMD open ecosystem.
What Lisa Su’s Strategy Means for Corporate Hardware Buyers
AMD SVP and GM of Adaptive and Embedded Computing Salil Raje stated plainly that AMD was not traditionally a software company but has turned its attention to software with ROCm and an explicit push to be more open source friendly so the developer community can work with it at scale, representing a subtle but massive shift from the old AMD whose playbook leaned on a sound chip and hoped the ecosystem would follow, with AI having taught the entire industry that the ecosystem must be engineered, funded, and obsessively supported.
For Corporate Hardware procurement teams evaluating AI infrastructure investments ahead of AMD’s July summit, the AMD Advancing AI event represents the most important opportunity to assess whether the AMD Advancing AI open source ecosystem blueprints Dr. Lisa Su presents in San Francisco align with the deployment requirements and budget constraints that Open Source AI infrastructure must satisfy to displace incumbent NVIDIA configurations across the enterprise tier.
Conclusion
AMD Advancing AI 2026 at the Moscone Center Tech venue on July 23, 2026, formalizes the Lisa Su Strategy, which has been assembled through ROCm investments, hyperscaler validation, and open-ecosystem partnerships over three years of deliberate infrastructure execution. ROCm Software version 7 with FP4 and FP8 precision support, day zero PyTorch and vLLM compatibility, and consumer GPU accessibility establishes the Open-Source AI developer foundation that AMD Advancing AI open-source ecosystem blueprints will document and distribute to the global developer community attending in person and via livestream. Corporate Hardware buyers who treat the July summit as a passive product announcement will miss its structural significance. It is the moment AMD formally presents the complete alternative to the proprietary AI infrastructure that Meta, Microsoft, and OpenAI have already chosen, and the enterprise tier is now positioned to adopt it at scale.
